diff --git a/src/.removed_features/domain/Domain.hpp b/src/.removed_features/domain/Domain.hpp
index d100acf2ae1d84fddee66ee8f7fb85e1568eaa44..7d2b657f5f10388e0e95fa1b34c2612db35cde00 100644
--- a/src/.removed_features/domain/Domain.hpp
+++ b/src/.removed_features/domain/Domain.hpp
@@ -1,5 +1,5 @@
/** @file feature_creation/domain/Domain.hpp
- * @brief A representation of the domain for each feature
+ * @brief A representation of the domain for each feature
*
* A dictionary representation of the domain for a features (key = domain, value = power)
*
diff --git a/src/.removed_features/log_fit_c_regressor/ModelLogRegressorFitC.hpp b/src/.removed_features/log_fit_c_regressor/ModelLogRegressorFitC.hpp
index f8023684d1f83ef0223d488653de43856f793265..8a88603d299bfeabb7f508c21775b11c587fb09f 100644
--- a/src/.removed_features/log_fit_c_regressor/ModelLogRegressorFitC.hpp
+++ b/src/.removed_features/log_fit_c_regressor/ModelLogRegressorFitC.hpp
@@ -1,5 +1,5 @@
/** @file descriptor_identifier/ModelLogRegressorFitC/ModelLogRegressorFitC.hpp
- * @brief Object to store the models generated form SISSO
+ * @brief Object to store the models generated form SISSO
*
* Creates a ModelLogRegressorFitC generated from SISSO and the corresponding output file.
* It also has functionality to read in an output file to regenerate the model.
@@ -123,9 +123,9 @@ public:
*/
inline std::vector<std::array<double, 2>> c_coefs(){return _c_coefs;}
/**
- * @brief Copy the error into a new array
+ * @brief Copy the training error into a different vector
*
- * @param res pointer to the beginning of the vector to store the residual
+ * @param res Pointer to the head of the new vector to store the residual
*/
inline void copy_error(double* res){std::copy_n(_train_error.data(), _n_samp_train, res);}
@@ -153,4 +153,4 @@ public:
*/
std::ostream& operator<< (std::ostream& outStream, const ModelLogRegressorFitC& model);
-#endif
\ No newline at end of file
+#endif
diff --git a/src/.removed_features/log_fit_c_regressor/SISSOLogRegressorFitC.hpp b/src/.removed_features/log_fit_c_regressor/SISSOLogRegressorFitC.hpp
index 961eb2809ad6d5b111e00e41a191f6380a63f15b..01eb4c381e9434adc2719588629908144aa2eb81 100644
--- a/src/.removed_features/log_fit_c_regressor/SISSOLogRegressorFitC.hpp
+++ b/src/.removed_features/log_fit_c_regressor/SISSOLogRegressorFitC.hpp
@@ -1,5 +1,5 @@
/** @file descriptor_identifier/SISSOLogRegressorFitC.hpp
- * @brief Perform SISSO on a previously generated Feature Space
+ * @brief Perform SISSO on a previously generated Feature Space
*
* Takes in a feature space and performs SISSO on it while storing the selected features in _models
*
@@ -41,7 +41,7 @@ public:
* @param leave_out_inds List of indexes from the initial data file in the test set
* @param n_dim Maximum dimensionality of the generated models
* @param n_residual Number of residuals to pass to the next SIS operation
- * @param n_models_store Number of features to store in files
+ * @param n_models_store The number of models to output to files
* @param fix_intrecept If true fix intercept to 0
*/
SISSOLogRegressorFitC(std::shared_ptr<FeatureSpace> feat_space, Unit prop_unit, std::vector<double> prop, std::vector<double> prop_test, std::vector<int> task_sizes_train, std::vector<int> task_sizes_test, std::vector<int> leave_out_inds, int n_dim, int n_residual, int n_models_store, bool fix_intercept=false);
@@ -76,7 +76,7 @@ public:
* @param prop The property to fit
* @param n_dim the dimensionality of the model
*/
- void l0_norm(std::vector<double>& prop, int n_dim);
+ void l0_regularization(std::vector<double>& prop, int n_dim);
/**
* @brief Acessor function for models
@@ -97,7 +97,7 @@ public:
* @param leave_out_inds List of indexes from the initial data file in the test set
* @param n_dim Maximum dimensionality of the generated models
* @param n_residual Number of residuals to pass to the next SIS operation
- * @param n_models_store Number of features to store in files
+ * @param n_models_store The number of models to output to files
* @param fix_intrecept If true fix intercept to 0
*/
SISSOLogRegressorFitC(
@@ -127,7 +127,7 @@ public:
* @param leave_out_inds List of indexes from the initial data file in the test set
* @param n_dim Maximum dimensionality of the generated models
* @param n_residual Number of residuals to pass to the next SIS operation
- * @param n_models_store Number of features to store in files
+ * @param n_models_store The number of models to output to files
* @param fix_intrecept If true fix intercept to 0
*/
SISSOLogRegressorFitC(
diff --git a/src/classification/ConvexHull1D.hpp b/src/classification/ConvexHull1D.hpp
index 8125c442508137c41665249106e6a9d934108bbc..d94d542811afaabb26feda34e664ccb95ce34fa2 100644
--- a/src/classification/ConvexHull1D.hpp
+++ b/src/classification/ConvexHull1D.hpp
@@ -11,13 +11,12 @@
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
-/** @file classificationutils.hpp
- * @brief A set of functions to create convex hulls, and compare overlap/distance
+/** @file classification/ConvexHull1D.hpp
+ * @brief A set of functions to create convex hulls, and compare overlap/distance
*
* @author Thomas A. R. Purcell (tpurcell)
* @bug No known bugs.
*/
-
#ifndef CONVEX_UTILS
#define CONVEX_UTILS
@@ -29,13 +28,13 @@
class ConvexHull1D
{
protected :
- std::vector<double> _sorted_value; //!< The value sorted based on the property
+ std::vector<double> _sorted_value; //!< The value of the input vector sorted based on the property vector
std::vector<double> _cls_max; //!< The maximum value of the property in each class
std::vector<double> _cls_min; //!< The minimum value of the property in each class
- std::vector<int> _sorted_prop_inds; //!< The value sorted based on the property
- std::vector<int> _cls_start; //!< The number of values in each class
- std::vector<int> _cls_sz; //!< The number of values in each class
+ std::vector<int> _sorted_prop_inds; //!< The indexes of where each element of an input vector is in _sorted_value
+ std::vector<int> _cls_start; //!< The element where each new class starts
+ std::vector<int> _cls_sz; //!< The number of elements in each class
std::vector<double> _task_scores; //!< The scores for each task
@@ -46,20 +45,20 @@ public:
/**
* @brief Constructor
*
- * @param sizes The size of the tasks
+ * @param sizes The size of each task
* @param prop The pointer to the property vector
*/
ConvexHull1D(const std::vector<int> sizes, const double* prop);
/**
- * @brief Initialize the projection objects
+ * @brief Default constructor
*/
ConvexHull1D();
/**
* @brief Initialize the projection objects
*
- * @param sizes The size of the tasks
+ * @param sizes The size of each task
* @param prop The pointer to the property vector
*/
void initialize_prop(const std::vector<int>& sizes, const double* prop);
@@ -67,7 +66,7 @@ public:
/**
* @brief Calculate the projection scores of a set of features to a vector via Pearson correlation
*
- * @param value The pointer to the value of the data
+ * @param value The pointer to the input data
* @param width The buffer used for calculating the overlap
*
* @returns The projection score for the particular feature
diff --git a/src/classification/LPWrapper.hpp b/src/classification/LPWrapper.hpp
index d517d130b0bd4a7bfae4deeee823f8a573623c5a..5002b33712378f62ad875583165d470e73aec32e 100644
--- a/src/classification/LPWrapper.hpp
+++ b/src/classification/LPWrapper.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file classification/LPWrapper.hpp
- * @brief A class used to wrap Coin-Clp into a single class
+ * @brief A class used to wrap Coin-Clp into a single class
*
* @author Thomas A. R. Purcell (tpurcell)
* @bug No known bugs.
@@ -30,9 +30,9 @@
class LPWrapper
{
protected:
- ClpSimplex _simplex; //!< Model used to find if a point is inside the convex hull
+ ClpSimplex _simplex; //!< LP algorithm used to determine if a point is inside the convex hull
- std::vector<double> _elements; //!< The elements of the A matrix (descriptor matrix) for the linear optimization problem
+ std::vector<double> _elements; //!< The elements of the A matrix (descriptor matrix) for the linear programing problem
std::vector<double> _row_lower; //!< The lower bound for the row constraints (value of the data point to check)
std::vector<double> _row_upper; //!< The upper bound for the row constraints (value of the data point to check)
std::vector<double> _row_lower_test; //!< The lower bound for the row constraints for the test set (value of the data point to check)
@@ -42,20 +42,20 @@ protected:
std::vector<double> _column_upper; //!< The upper bound for the column constraints (all infinity)
std::vector<int> _column_start; //!< Vector len(n_samples) of the starting column for the A matrix
- std::vector<int> _row; //!< Vector saying what row each data point is in
- std::vector<int> _row_test; //!< Vector saying what row each data point is in for the test set
+ std::vector<int> _row; //!< Vector storing what row each data point is in
+ std::vector<int> _row_test; //!< Vector storing what row each data point is in for the test set
- std::vector<int> _n_row_per_class; //!< Number of rows per class
- std::vector<int> _n_row_per_class_test; //!< Number of rows per class in the test set
+ std::vector<int> _n_row_per_class; //!< //!< A vector storing the number of rows in each class/task combination for the training set [task * N_TASK * N_CLASS + class]
+ std::vector<int> _n_row_per_class_test; //!< //!< A vector storing the number of rows in each class/task combination for the test set [task * N_TASK * N_CLASS + class]
- double _tol; //!< tolerance value (\epsilon) for the final row constraint
+ double _tol; //!< tolerance value (\epsilon) for the final row constraint 1 +/- _tol for the upper and lower constraint
int _n_dim; //!< The number of dimensions for the SVM problem
- int _n_samp; //!< The number of samples per feature
- int _n_samp_test; //!< The number of samples per feature in the test set
+ int _n_samp; //!< The number of samples in the training data per feature
+ int _n_samp_test; //!< The number of samples in the test data per feature
int _n_class; //!< The number of classes in the data set
int _n_row; //!< The number of rows in the A matrix
- int _task_num; //!< The task number the LPWrapper is used for
+ int _task_num; //!< The current task number the LPWrapper is used for
int _n_col; //!< The number of columns in the A matrix
int _n_overlap; //!< The number of misclassified data points in the training set
int _n_overlap_test; //!< The number of misclassified data points in the test set
@@ -66,14 +66,14 @@ public:
/**
* @brief The constructor for the LPWrapper
*
- * @param samp_per_class number of samples per class
- * @param task_num The task ID number to perform the calculation
+ * @param samp_per_class A vector storing the number of rows in each class/task combination for the training set [task * N_TASK * N_CLASS + class]
+ * @param task_num The task ID number used to perform the calculation
* @param n_class Number of classes in the dataset
* @param n_dim Number of dimensions of the problem
* @param n_samp Number of samples in the dataset
- * @param n_samp_test Number of samples in the dataset for the test set
- * @param samp_per_class number of samples per class for the test set
* @param tol The tolerance used to have a fuzzy border around the convex hull
+ * @param samp_per_class A vector storing number of test samples per class
+ * @param n_samp_test TheA vector storing tnumber of rows in each class/task combination for the test set [task * N_TASK * N_CLASS + class]
*/
LPWrapper(
std::vector<int> samp_per_class,
@@ -122,75 +122,69 @@ public:
/**
* @brief Copies the data from a set of feature indexes (sorted_dmatrix) into the x_space
*
- * @param cls class the number of the class to copy over
- * @param inds list of indexes to pull data for
+ * @param cls Index defining what class to copy
+ * @param inds A vector storing the _d_mat_ind of the features to pull data from the Descriptor Matrix storage array
*/
void copy_data(const int cls, const std::vector<int> inds);
/**
* @brief Copies the data from a set data pointers
*
- * @param cls class the number of the class to copy over
- * @param val_ptrs The pointers to the feature's data
+ * @param cls Index defining what class to copy
+ * @param val_ptrs A vector storing pointers to the features training data
+ * @param test_val_ptrs A vector storing pointers to the features test data
*/
void copy_data(const int cls, const std::vector<double*> val_ptrs, const std::vector<double*> test_val_ptrs);
/**
- * @brief Copy the data from a set of feature indexes (sorted_dmatrix) into the x_space and train the SVM model
+ * @brief Calculate the number of points in the overlap region of the convex hulls for the training set
*
- * @param inds list of indexes to pull data for
+ * @param inds A vector storing the _d_mat_ind of the features to pull data from the Descriptor Matrix storage array
*/
int get_n_overlap(const std::vector<int> inds);
/**
- * @brief Copy tthe data from a set data pointers and train the SVM model
+ * @brief Calculate the number of points in the overlap region of the convex hulls for the training and test sets and set the error vectors
*
- * @param val_ptrs The pointers to the feature's data
- * @param test_val_ptrs The pointers to the feature's data (Test set)
- * @param error Pointer to the error vector
- * @param test_error Pointer to the test_error
+ * @param val_ptrs A vector storing pointers to the features training data
+ * @param test_val_ptrs A vector storing pointers to the features test data
+ * @param error Pointer to the head of the error vector
+ * @param test_error Pointer to the head of the test error vector
*/
void set_n_overlap(const std::vector<double*> val_ptrs, const std::vector<double*> test_val_ptrs, double* error, double* test_error);
/**
* @brief The number of classes in the training set
- * @return The number of classes in the training set
*/
inline int n_class() const {return _n_class;}
/**
- * @brief The task id number
- * @return The task id number
+ * @brief The index of the current task that is stored in the LPWrapper
*/
inline int task_num() const {return _task_num;}
/**
* @brief The number of dimensions of the Convex Hulls
- * @return The number of dimensions of the Convex Hulls
*/
inline int n_dim() const {return _n_dim;}
/**
* @brief The number of samples in the training set
- * @return The number of samples in the training set
*/
inline int n_samp() const {return _n_samp;}
/**
* @brief The number of samples in the test set
- * @return The number of samples in the test set
*/
inline int n_samp_test() const {return _n_samp_test;}
/**
* @brief Number of misclassified samples in the data set
- * @return Number of misclassified samples in the data set
*/
inline int n_overlap() const {return _n_overlap;}
/**
* @brief Number of misclassified samples in the test set
- * @return Number of misclassified samples in the test set
*/
inline int n_overlap_test() const {return _n_overlap_test;}
};
diff --git a/src/classification/SVMWrapper.cpp b/src/classification/SVMWrapper.cpp
index 36736850277772192cc6b08627a658cd1bb6ade2..4fa2bf162830be4a22e07df08df693835229fc69 100644
--- a/src/classification/SVMWrapper.cpp
+++ b/src/classification/SVMWrapper.cpp
@@ -250,19 +250,19 @@ void SVMWrapper::train(const std::vector<double*> val_ptrs, const bool remap_coe
train(remap_coefs);
}
-std::vector<double> SVMWrapper::predict(const int n_test_samp, const std::vector<double*> val_ptrs)
+std::vector<double> SVMWrapper::predict(const int n_samp_test, const std::vector<double*> val_ptrs)
{
if(val_ptrs.size() > _n_dim)
{
throw std::logic_error("Size of the val_ptrs vector is larger than _n_dim");
}
- std::vector<double> y_est_test(n_test_samp, 0.0);
+ std::vector<double> y_est_test(n_samp_test, 0.0);
- std::vector<svm_node>x_space_test(n_test_samp * (val_ptrs.size() + 1));
- std::vector<svm_node*>x_test(n_test_samp);
+ std::vector<svm_node>x_space_test(n_samp_test * (val_ptrs.size() + 1));
+ std::vector<svm_node*>x_test(n_samp_test);
- for(int ss = 0; ss < n_test_samp; ++ss)
+ for(int ss = 0; ss < n_samp_test; ++ss)
{
for(int dd = 0; dd < val_ptrs.size(); ++dd)
{
diff --git a/src/classification/SVMWrapper.hpp b/src/classification/SVMWrapper.hpp
index f584c91439cc49503259f4567dc4332861a8a546..c8a589551a0c9c7dcfa4ae2c487d7aea34f043ca 100644
--- a/src/classification/SVMWrapper.hpp
+++ b/src/classification/SVMWrapper.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file classification/SVMWrapper.hpp
- * @brief A class used to wrap libsvm in more C++ oriented data structure
+ * @brief A class used to wrap libsvm in more C++ oriented data structure
*
* @author Thomas A. R. Purcell (tpurcell)
* @bug No known bugs.
@@ -47,7 +47,7 @@ protected:
const double _C; //!< The C parameter for the SVM calculation
const int _n_dim; //!< The number of dimensions for the SVM problem
- const int _n_samp; //!< The number of samples per feature
+ const int _n_samp; //!< The number of training samples per feature
const int _n_class; //!< The number of classes in the data set
int _n_misclassified; //!< The number of misclassified data points in the training set
@@ -57,7 +57,7 @@ public:
*
* @param n_class Number of classes in the dataset
* @param n_dim Number of dimensions of the problem
- * @param n_samp Number of samples in the dataset
+ * @param n_samp Number of training samples in the dataset
* @param prop pointer to the start of the property vector
*/
SVMWrapper(const int n_class, const int n_dim, const int n_samp, const double* prop);
@@ -77,7 +77,7 @@ public:
* @param C The C value for the SVM calculation
* @param n_class Number of classes in the dataset
* @param n_dim Number of dimensions of the problem
- * @param n_samp Number of samples in the dataset
+ * @param n_samp Number of training samples in the dataset
* @param prop pointer to the start of the property vector
*/
SVMWrapper(const double C, const int n_class, const int n_dim, const int n_samp, const double* prop);
@@ -128,7 +128,7 @@ public:
/**
* @brief Copies the data from a set of feature indexes (sorted_dmatrix) into the x_space
*
- * @param inds list of indexes to pull data for
+ * @param inds A vector storing the _d_mat_ind of the features to pull data from the Descriptor Matrix storage array
* @param task the task number for the calculation
*/
void copy_data(const std::vector<int> inds, const int task);
@@ -150,16 +150,16 @@ public:
/**
* @brief Copy the data from a set of feature indexes (sorted_dmatrix) into the x_space and train the SVM model
*
- * @param inds list of indexes to pull data for
+ * @param inds A vector storing the _d_mat_ind of the features to pull data from the Descriptor Matrix storage array
* @param task the task number for the calculation
* @param remap_coefs If true remap the final coefficients back to the unscaled feature space
*/
void train(const std::vector<int> inds, const int task, const bool remap_coefs=true);
/**
- * @brief Copy tthe data from a set data pointers and train the SVM model
+ * @brief Copy the data from a set data pointers and train the SVM model
*
- * @param val_ptrs The pointers to the feature's data
+ * @param val_ptrs A vector storing pointers to the features training data
* @param remap_coefs If true remap the final coefficients back to the unscaled feature space
*/
void train(const std::vector<double*> val_ptrs, const bool remap_coefs=true);
@@ -167,58 +167,50 @@ public:
/**
* @brief Predict the class of a set of data from the SVM model
*
- * @param n_test_samp the number of test samples to predict the class off
- * @param val_ptrs vector of the data pointers for the feature's test set
+ * @param n_samp_test the number of test samples to predict the class off
+ * @param val_ptrs A vector storing pointers to the features test data
*
* @return The predicted class of the test samples
*/
- std::vector<double> predict(const int n_test_samp, const std::vector<double*> val_ptrs);
+ std::vector<double> predict(const int n_samp_test, const std::vector<double*> val_ptrs);
/**
* @brief the coefficients of the decision planes for each of the one against one SVM models
- * @return A list of all coefficients for the decision planes
*/
inline std::vector<std::vector<double>> coefs(){return _coefs;}
/**
- * @brief the list of the bias terms in all the Svm models
- * @return The bias term for each SVM generated decision function
+ * @brief the list of the bias terms in all the SVM models
*/
inline std::vector<double> intercept() const {return _intercept;}
/**
* @brief The number of classes in the training set
- * @return The number of classes in the training set
*/
inline int n_class() const {return _n_class;}
/**
* @brief The number of dimensions of the SVM model
- * @return The number of dimensions of the SVM model
*/
inline int n_dim() const {return _n_dim;}
/**
* @brief The number of samples in the training set
- * @return The number of samples in the training set
*/
inline int n_samp() const {return _n_samp;}
/**
* @brief The predicted class for each sample in the data set
- * @return The predicted class for each sample in the data set
*/
inline std::vector<double> y_estimate() const {return _y_est;}
/**
* @brief The actual class for each sample in the data set
- * @return The actual class for each sample in the data set
*/
inline std::vector<double> y_actual() const {return _y;}
/**
* @brief Number of misclassified samples in the data set
- * @return Number of misclassified samples in the data set
*/
inline int n_misclassified() const {return _n_misclassified;}
diff --git a/src/classification/prop_sorted_d_mat.hpp b/src/classification/prop_sorted_d_mat.hpp
index 143abeb2937c0a69229f68318bf6d9e1286d7462..8434fb2f4014c461f767cc7aed6bea366b33c627 100644
--- a/src/classification/prop_sorted_d_mat.hpp
+++ b/src/classification/prop_sorted_d_mat.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file classification/prop_sorted_d_mat.hpp
- * @brief Central storage area for a sorted descriptor matrix based on the task/class of a sample
+ * @brief Central storage area for a sorted descriptor matrix based on the task/class of a sample
* @author Thomas A. R. Purcell (tpurcell)
* @bug No known bugs.
@@ -27,10 +27,10 @@
namespace prop_sorted_d_mat
{
- extern std::vector<double> SORTED_D_MATRIX; //!< The descriptor matrix
+ extern std::vector<double> SORTED_D_MATRIX; //!< The descriptor matrix with the feature's input data sorted by the class index
extern std::vector<int> CLASS_START; //!< The starting point for each class
- extern std::vector<int> N_SAMPLES_PER_CLASS; //!< number of features in each class/task combination [task * N_TASK * N_CLASS + class]
+ extern std::vector<int> N_SAMPLES_PER_CLASS; //!< The number of samples in each class/task combination [task * N_TASK * N_CLASS + class]
extern int N_TASK; //!< Number of tasks for the calculation
extern int N_CLASS; //!< Number of classes per task for the calculation
@@ -44,20 +44,19 @@ namespace prop_sorted_d_mat
* @param n_feat The number of features
* @param n_task The number of tasks
* @param n_class The number of classes
- * @param n_samples_per_class The number of samples in each class
+ * @param n_samples_per_class The number of samples in each class/task combination [task * N_TASK * N_CLASS + class]
*/
void initialize_sroted_d_matrix_arr(int n_feat, int n_task, int n_class, std::vector<int> n_samples_per_class);
/**
- * @brief Resize the descriptor matrix
- * @details Using the total number of features
+ * @brief Resize the descriptor matrix to match the current number of features
*
* @param n_feats Number of features to select
*/
void resize_sroted_d_matrix_arr(int n_feats);
/**
- * @brief Get the size of a particular class
+ * @brief Get the number of samples in a particular task/class combination
*
* @param task The task number
* @param cls the class number
@@ -78,11 +77,13 @@ namespace prop_sorted_d_mat
{
return SORTED_D_MATRIX[feat_ind * N_SAMPLES + samp_ind];
}
+
/**
* @brief Access the sorted descriptor matrix by feature ind
*
- * @param feat_ind feature index
- * @return pointer to the start of the selected feature
+ * @param feat_ind The feature index
+ *
+ * @return Pointer to the start of the selected feature
*/
inline double* access_sorted_d_matrix(int feat_ind)
{
@@ -90,11 +91,11 @@ namespace prop_sorted_d_mat
}
/**
- * @brief Access the sorted descriptor matrix by the class and task number
+ * @brief Get a pointer to the head of the sorted description matrix for a given task/class combination
*
* @param task The task number
* @param cls The class number
- * @return pointer to the start of a given class and task for feature 0
+ * @return Pointer to the start of a given class and task for all features
*/
inline double* access_sorted_d_matrix(int task, int cls)
{
@@ -107,7 +108,7 @@ namespace prop_sorted_d_mat
* @param feat_ind feature index
* @param task The task number
* @param cls The class number
- * @return pointer to the start of the selected feature for a given class and task
+ * @return pointer to the start of the selected feature for a given feature, class, and task
*/
inline double* access_sorted_d_matrix(int feat_ind, int task, int cls)
{
diff --git a/src/descriptor_identifier/Model/Model.cpp b/src/descriptor_identifier/Model/Model.cpp
index 946470fef15510922509246485e74ca6b9c52d93..e3ef4be28b2562c0b00b0136f07090e14526d842 100644
--- a/src/descriptor_identifier/Model/Model.cpp
+++ b/src/descriptor_identifier/Model/Model.cpp
@@ -21,7 +21,7 @@ Model::Model(
const std::vector<int> leave_out_inds
) :
_n_samp_train(feats[0]->n_samp()),
- _n_samp_test(feats[0]->n_test_samp()),
+ _n_samp_test(feats[0]->n_samp_test()),
_n_dim(feats.size()),
_feats(feats),
_leave_out_inds(leave_out_inds),
diff --git a/src/descriptor_identifier/Model/Model.hpp b/src/descriptor_identifier/Model/Model.hpp
index afdc32917f6bef63122da177069e70fd702ab1e3..2a4f1790227467e338a6dae859de85c661197497 100644
--- a/src/descriptor_identifier/Model/Model.hpp
+++ b/src/descriptor_identifier/Model/Model.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file descriptor_identifier/Model/Model.hpp
- * @brief Object to store the models generated form SISSO
+ * @brief Object to store the models generated form SISSO
*
* Creates a Model generated from SISSO and the corresponding output file.
* It also has functionality to read in an output file to regenerate the model.
@@ -45,14 +45,14 @@ class Model
protected:
int _n_samp_train; //!< The number of samples per feature
int _n_samp_test; //!< The number of test samples per feature
- int _n_dim; //!< Dimension of the model
+ int _n_dim; //!< The number of dimensions of the model
std::vector<model_node_ptr> _feats; //!< List of features in the model
std::vector<std::vector<double>> _coefs; //!< Coefficients for the features
std::vector<int> _leave_out_inds; //!< The indexes used as the test set
- std::shared_ptr<LossFunction> _loss;
- std::string _prop_label; //!< label for the model
+ std::shared_ptr<LossFunction> _loss; //!< The LossFunction used to evaluate the model
+ std::string _prop_label; //!< The property label for the model
Unit _prop_unit; //!< The Unit for the property
int _task_eval; //!< Which set of coefficients to use for evaluating the model for a new data set
@@ -111,15 +111,14 @@ public:
Model& operator= (Model&& o) = default;
/**
- * @brief Copy the error into a new array
+ * @brief Copy the training error into a different vector
*
- * @param res pointer to the beginning of the vector to store the residual
+ * @param res Pointer to the head of the new vector to store the residual
*/
virtual void copy_error(double* res) const = 0;
/**
- * @brief Read an output file and extract all relevant information
- * @details Takes in an output file and extracts all data needed to recreate the model
+ * @brief Read an output file and extract all relevant information from it
*
* @param train_filename Name of the training output file
* @param test_filename Name of the test output file
@@ -129,10 +128,10 @@ public:
/**
* @brief Create the LossFunction for the Model when constructed from a file
*
- * @param prop_train Vector storing all data to train the SISSO models with
- * @param prop_test Vector storing all data to test the SISSO models with
- * @param task_sizes_train Number of training samples per task
- * @param task_sizes_test Number of testing samples per task
+ * @param prop_train The property vector for all of the training samples
+ * @param prop_test The property vector for all of the test samples
+ * @param task_sizes_train The number of training samples per task
+ * @param task_sizes_test The number of test samples per task
*/
virtual void make_loss(
std::vector<double>& prop_train,
@@ -142,7 +141,7 @@ public:
) = 0;
/**
- * @brief Based off of the model line in a model file determine if the intercept is fixed
+ * @brief Based off of the model line in a model file determine if the bias term is 0.0
*
* @param model_line the model line from the file
* @return True if the intercept should be fixed
@@ -244,13 +243,13 @@ public:
// DocString: model_n_samp_train
/**
- * @brief Total number of samples being trained on
+ * @brief The number of samples in the training data per feature
*/
inline int n_samp_train() const {return _n_samp_train;}
// DocString: model_n_samp_test
/**
- * @brief Total number of samples being tested
+ * @brief The number of samples in the test data per feature
*/
inline int n_samp_test() const {return _n_samp_test;}
@@ -295,7 +294,7 @@ public:
// DocString: model_fix_intercept
/**
- * @brief Access whether the intercept is fixed at 0 or not
+ * @brief True if the bias term is 0.0
*/
inline bool fix_intercept() const {return _fix_intercept;}
@@ -311,10 +310,15 @@ public:
*/
virtual std::string toString() const = 0;
+ /**
+ * @brief Convert the model to a latexified string
+ */
+ virtual std::string toLatexString() const = 0;
+
/**
* @brief Get the matlab expression for this model
- * @return The matlab exprssion for the model
+ * @return The matlab expression for the model
*/
virtual std::string matlab_expr() const = 0;
@@ -327,52 +331,47 @@ public:
void write_matlab_fxn(std::string fxn_filename);
/**
- * @brief The training data property to be learned
- * @return _prop_train as a numpy ndarray
+ * @brief The property vector for all of the training samples
*/
virtual inline std::vector<double> prop_train() const {return _loss->prop_train();}
/**
- * @brief The test values for the property
- * @return _prop_test as a numpy ndarray
+ * @brief The property vector for all of the test samples
*/
virtual inline std::vector<double> prop_test() const {return _loss->prop_test();}
/**
- * @brief The training data property to be learned
- * @return _prop_train as a numpy ndarray
+ * @brief The estimated property vector for all of the training samples
*/
virtual inline std::vector<double> prop_train_est() const {return _loss->prop_train_est();}
/**
- * @brief The test values for the property
- * @return _prop_test as a numpy ndarray
+ * @brief The estimated property vector for all of the test samples
*/
virtual inline std::vector<double> prop_test_est() const {return _loss->prop_test_est();}
/**
- * @brief Access the size of the training set for each task
+ * @brief The number of training samples per task
*/
inline std::vector<int> task_sizes_train() const {return _loss->task_sizes_train();}
/**
- * @brief Access the size of the training set for each task
+ * @brief The number of test samples per task
*/
inline std::vector<int> task_sizes_test() const {return _loss->task_sizes_test();}
/**
- * @brief The training error
- * @return _train_error as a numpy ndarray
+ * @brief The error vector for the training data as a numpy array
*/
virtual inline std::vector<double> train_error() const {return _loss->error_train();}
/**
- * @brief The test error
- * @return _test_error as a numpy ndarray
+ * @brief The error vector for the test data as a numpy array
*/
virtual inline std::vector<double> test_error() const {return _loss->error_test();}
#ifdef PY_BINDINGS
+
// DocString: model_coefs
/**
* @brief The coefficient array for the model (n_task, n_dim + bias term)
diff --git a/src/descriptor_identifier/Model/ModelClassifier.hpp b/src/descriptor_identifier/Model/ModelClassifier.hpp
index 78aa6fe13b2939b215c18aa966ae983f75306883..6459c29a465636651665e2629f0619f6e7f7c8c9 100644
--- a/src/descriptor_identifier/Model/ModelClassifier.hpp
+++ b/src/descriptor_identifier/Model/ModelClassifier.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file descriptor_identifier/ModelClassifier/ModelClassifier.hpp
- * @brief Object to store the models generated form SISSO
+ * @brief Object to store the models generated form SISSO
*
* Creates a ModelClassifier generated from SISSO and the corresponding output file.
* It also has functionality to read in an output file to regenerate the model.
@@ -37,7 +37,7 @@
*/
class ModelClassifier : public Model
{
- int _n_class; //!< The number of classes
+ int _n_class; //!< The number of classes in the model
int _train_n_convex_overlap; //!< Number of data points in regions where the convex hulls overlap in the training set
int _test_n_convex_overlap; //!< Number of data points in regions where the convex hulls overlap for the test data set
@@ -45,6 +45,7 @@ class ModelClassifier : public Model
int _train_n_svm_misclassified; //!< The number of points misclassified by SVM (training set)
int _test_n_svm_misclassified; //!< The number of points misclassified by SVM (test set)
public:
+
/**
* @brief Construct a ModelClassifier using a loss function and a set of features
*
@@ -66,7 +67,7 @@ public:
/**
* @brief Construct a ModelClassifier from a training output file
*
- * @param train_file (str) Previously generated model file
+ * @param train_file (str) Previously generated model output file for the training data
*/
ModelClassifier(const std::string train_file);
@@ -74,8 +75,8 @@ public:
/**
* @brief Construct a ModelClassifier from a training and testing output file
*
- * @param train_file (str) Previously generated training model output file
- * @param test_file (str) Previously generated testing model output file
+ * @param train_file (str) Previously generated model output file for the training data
+ * @param test_file (str) Previously generated model output file for the test data
*/
ModelClassifier(const std::string train_file, const std::string test_file);
@@ -110,10 +111,10 @@ public:
/**
* @brief Create the LossFunctionConvexHull for the Model when constructed from a file
*
- * @param prop_train Vector storing all data to train the SISSO models with
- * @param prop_test Vector storing all data to test the SISSO models with
- * @param task_sizes_train Number of training samples per task
- * @param task_sizes_test Number of testing samples per task
+ * @param prop_train The property vector for all of the training samples
+ * @param prop_test The property vector for all of the test samples
+ * @param task_sizes_train The number of training samples per task
+ * @param task_sizes_test The number of test samples per task
*/
void make_loss(
std::vector<double>& prop_train,
@@ -123,7 +124,7 @@ public:
);
/**
- * @brief Based off of the model line in a model file determine if the intercept is fixed
+ * @brief Based off of the model line in a model file determine if the bias term is 0.0
*
* @param model_line the model line from the file
* @return True if the intercept should be fixed
@@ -198,9 +199,9 @@ public:
std::string coefs_header() const;
/**
- * @brief Copy the error into a new array
+ * @brief Copy the training error into a different vector
*
- * @param res pointer to the beginning of the vector to store the residual
+ * @param res Pointer to the head of the new vector to store the residual
*/
inline void copy_error(double* res) const {std::copy_n(_loss->error_train().data(), _n_samp_train, res);}
diff --git a/src/descriptor_identifier/Model/ModelLogRegressor.hpp b/src/descriptor_identifier/Model/ModelLogRegressor.hpp
index b94332a9283b274288e1f789eb5797a8fb87988b..b57965afe654e6723174e4d5b6b9c12f8156b52c 100644
--- a/src/descriptor_identifier/Model/ModelLogRegressor.hpp
+++ b/src/descriptor_identifier/Model/ModelLogRegressor.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file descriptor_identifier/ModelLogRegressor/ModelLogRegressor.hpp
- * @brief Object to store the models generated form SISSO
+ * @brief Object to store the models generated form SISSO
*
* Creates a ModelLogRegressor generated from SISSO and the corresponding output file.
* It also has functionality to read in an output file to regenerate the model.
@@ -64,19 +64,17 @@ public:
// DocString: model_log_reg_init_train
/**
* @brief Construct a ModelLogRegressor from a training output file
- * @details Reads in all of the data from the output file and recreates the model object
*
- * @param train_file (str) Previously generated model file
+ * @param train_file (str) Previously generated model output file for the training data
*/
ModelLogRegressor(std::string train_file);
// DocString: model_log_reg_init_test_train
/**
* @brief Construct a ModelLogRegressor from a training and testing output file
- * @details Reads in all of the data from the output files and recreates the model object
*
- * @param train_file (str) Previously generated training model output file
- * @param test_file (str) Previously generated testing model output file
+ * @param train_file (str) Previously generated model output file for the training data
+ * @param test_file (str) Previously generated model output file for the test data
*/
ModelLogRegressor(std::string train_file, std::string test_file);
@@ -108,15 +106,13 @@ public:
*/
ModelLogRegressor& operator= (ModelLogRegressor&& o) = default;
-
-
/**
* @brief Create the LossFunctionLogPearsonRMSE for the Model when constructed from a file
*
- * @param prop_train Vector storing all data to train the SISSO models with
- * @param prop_test Vector storing all data to test the SISSO models with
- * @param task_sizes_train Number of training samples per task
- * @param task_sizes_test Number of testing samples per task
+ * @param prop_train The property vector for all of the training samples
+ * @param prop_test The property vector for all of the test samples
+ * @param task_sizes_train The number of training samples per task
+ * @param task_sizes_test The number of test samples per task
*/
void make_loss(
std::vector<double>& prop_train,
@@ -126,7 +122,7 @@ public:
);
/**
- * @brief Based off of the model line in a model file determine if the intercept is fixed
+ * @brief Based off of the model line in a model file determine if the bias term is 0.0
*
* @param model_line the model line from the file
* @return True if the intercept should be fixed
@@ -171,14 +167,14 @@ public:
/**
* @brief Get the matlab expression for this model
- * @return The matlab exprssion for the model
+ * @return The matlab expression for the model
*/
std::string matlab_expr() const;
/**
- * @brief Copy the error into a new array
+ * @brief Copy the training error into a different vector
*
- * @param res pointer to the beginning of the vector to store the residual
+ * @param res Pointer to the head of the new vector to store the residual
*/
inline void copy_error(double* res) const {std::copy_n(_loss->error_train().data(), _n_samp_train, res);}
diff --git a/src/descriptor_identifier/Model/ModelRegressor.hpp b/src/descriptor_identifier/Model/ModelRegressor.hpp
index 2c90c4619da366d5cc51f28773d22a19e213b908..93713be12e2210201ff4e8b4c282e20ad2360ebc 100644
--- a/src/descriptor_identifier/Model/ModelRegressor.hpp
+++ b/src/descriptor_identifier/Model/ModelRegressor.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file descriptor_identifier/ModelRegressor/ModelRegressor.hpp
- * @brief Object to store the models generated form SISSO
+ * @brief Object to store the models generated form SISSO
*
* Creates a ModelRegressor generated from SISSO and the corresponding output file.
* It also has functionality to read in an output file to regenerate the model.
@@ -63,19 +63,17 @@ public:
// DocString: model_reg_init_train
/**
* @brief Construct a ModelRegressor from a training output file
- * @details Reads in all of the data from the output file and recreates the model object
*
- * @param train_file (str) Previously generated model file
+ * @param train_file (str) Previously generated model output file for the training data
*/
ModelRegressor(const std::string train_file);
// DocString: model_reg_init_test_train
/**
* @brief Construct a ModelRegressor from a training and testing output file
- * @details Reads in all of the data from the output files and recreates the model object
*
- * @param train_file (str) Previously generated training model output file
- * @param test_file (str) Previously generated testing model output file
+ * @param train_file (str) Previously generated model output file for the training data
+ * @param test_file (str) Previously generated model output file for the test data
*/
ModelRegressor(const std::string train_file, const std::string test_file);
@@ -96,10 +94,10 @@ public:
/**
* @brief Create the LossFunctionPearsonRMSE for the Model when constructed from a file
*
- * @param prop_train Vector storing all data to train the SISSO models with
- * @param prop_test Vector storing all data to test the SISSO models with
- * @param task_sizes_train Number of training samples per task
- * @param task_sizes_test Number of testing samples per task
+ * @param prop_train The property vector for all of the training samples
+ * @param prop_test The property vector for all of the test samples
+ * @param task_sizes_train The number of training samples per task
+ * @param task_sizes_test The number of test samples per task
*/
virtual void make_loss(
std::vector<double>& prop_train,
@@ -139,7 +137,7 @@ public:
ModelRegressor& operator= (ModelRegressor&& o) = default;
/**
- * @brief Based off of the model line in a model file determine if the intercept is fixed
+ * @brief Based off of the model line in a model file determine if the bias term is 0.0
*
* @param model_line the model line from the file
* @return True if the intercept should be fixed
@@ -184,7 +182,7 @@ public:
/**
* @brief Get the matlab expression for this model
- * @return The matlab exprssion for the model
+ * @return The matlab expression for the model
*/
virtual std::string matlab_expr() const;
@@ -201,9 +199,9 @@ public:
std::string coefs_header() const;
/**
- * @brief Copy the error into a new array
+ * @brief Copy the training error into a different vector
*
- * @param res pointer to the beginning of the vector to store the residual
+ * @param res Pointer to the head of the new vector to store the residual
*/
virtual inline void copy_error(double* res) const
{
diff --git a/src/descriptor_identifier/SISSO_DI/SISSOClassifier.cpp b/src/descriptor_identifier/SISSO_DI/SISSOClassifier.cpp
index d68959c376916727de078ba8ffab5915e177d9df..437e27ca8829ea11d75c34732dddc79b654b57c2 100644
--- a/src/descriptor_identifier/SISSO_DI/SISSOClassifier.cpp
+++ b/src/descriptor_identifier/SISSO_DI/SISSOClassifier.cpp
@@ -124,7 +124,7 @@ void SISSOClassifier::transfer_d_mat_to_sorted() const
}
}
-void SISSOClassifier::l0_norm(const int n_dim)
+void SISSOClassifier::l0_regularization(const int n_dim)
{
const int n_get_models = std::max(_n_residual, _n_models_store);
std::vector<double> coefs(n_dim + 1, 0.0);
@@ -317,7 +317,7 @@ void SISSOClassifier::fit()
}
start = omp_get_wtime();
- l0_norm(dd);
+ l0_regularization(dd);
_mpi_comm->barrier();
duration = omp_get_wtime() - start;
diff --git a/src/descriptor_identifier/SISSO_DI/SISSOClassifier.hpp b/src/descriptor_identifier/SISSO_DI/SISSOClassifier.hpp
index 2d505b0babcf5f98400cb1e438b7523e3f21e6fe..0ba5dc5dc0d3795b431a48274c642e97090e7368 100644
--- a/src/descriptor_identifier/SISSO_DI/SISSOClassifier.hpp
+++ b/src/descriptor_identifier/SISSO_DI/SISSOClassifier.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file descriptor_identifier/SISSOClassifier.hpp
- * @brief Perform SISSO on a previously generated Feature Space
+ * @brief Perform SISSO on a previously generated Feature Space
*
* Takes in a feature space and performs SISSO on it while storing the selected features in _models
*
@@ -51,21 +51,20 @@ protected:
int _n_class; //!< The number of classes in the calculation
public:
- /**
/**
* @brief Create a SISSOClassifier object
*
- * @param feat_space The FeatureSpace object that will be associated with this classifier (Defines SIS)
- * @param prop_label The label used to identify the property column in the DataFrame
+ * @param feat_space Feature Space for the problem
+ * @param prop_label The label for the property
* @param prop_unit The unit of the property
- * @param prop The property vector for all of the training samples
- * @param prpo_test The property vector for all of the test samples
- * @param task_sizes_train The number of training samples per task
- * @param task_sizes_test The number of test samples per task
- * @param leave_out_inds List of indexes from the initial data file in the test set
- * @param n_dim The highest possible dimension for the linear model
- * @param n_residual The number of residuals each iteration of SIS acts on
- * @param n_models_store The number of models to store in output files
+ * @param prop The value of the property to evaluate the loss function against for the training set
+ * @param prpo_test The value of the property to evaluate the loss function against for the test set
+ * @param task_sizes_train Number of training samples per task
+ * @param task_sizes_test Number of testing samples per task
+ * @param leave_out_inds List of indexes from the initial data file in the test set
+ * @param n_dim The maximum number of features allowed in the linear model
+ * @param n_residual Number of residuals to pass to the next SIS operation
+ * @param n_models_store The number of models to output to files
*/
SISSOClassifier(
const std::shared_ptr<FeatureSpace> feat_space,
@@ -96,32 +95,31 @@ public:
/**
* @brief Gets the max error index for the classification problem
- * @details [long description]
*
* @param n_models number of models to be stored
* @param n_convex_overlap number of points in the overlapping convex hull regions (in all models)
* @param svm_score the number of points misclassified by SVM (in all models)
* @param svm_margin The margin of the SVM model (in all models)
* @param scores the pointer to the scores array
- * @return Get the index with the maximum error
+ * @return The index with the maximum error
*/
int get_max_error_ind(const int n_models, const int* n_convex_overlap, const double* svm_score, const double* svm_margin, double* scores) const;
/**
- * @brief sets up _sample_inds_to_sorted_dmat_inds
+ * @brief Sort the property vector by class and store the mapped indexes to _sample_inds_to_sorted_dmat_inds
*/
void setup_d_mat_transfer();
/**
- * @brief Preform the l0 normalization for a property or the residual
+ * @brief Preform an l0-Regularization to find the best n_dim dimensional model
*
- * @param n_dim the dimensionality of the model
+ * @param n_dim The number of features to use in the linear model
*/
- void l0_norm(const int n_dim);
+ void l0_regularization(const int n_dim);
// DocString: sisso_class_fit
/**
- * @brief Perform the SISSO calculation to find the best model for each dimension up to n_dim dimensions equal to _n_dim
+ * @brief Iteratively run SISSO on the FeatureSpace an Property vector until the maximum dimenisonality is reached
*/
void fit();
@@ -136,17 +134,17 @@ public:
/**
* @brief Create a SISSOClassifier object
*
- * @param feat_space (FeatureSpace) The FeatureSpace object that will be associated with this classifier (Defines SIS)
- * @param prop_label (str) The label used to identify the property column in the DataFrame
+ * @param feat_space (FeatureSpace) Feature Space for the problem
+ * @param prop_label (str) The label for the property
* @param prop_unit (Unit) The unit of the property
- * @param prop (np.ndarray) The property vector for all of the training samples
- * @param prpo_test (np.ndarray) The property vector for all of the test samples
- * @param task_sizes_train (list) The number of training samples per task
- * @param task_sizes_test (list) The number of test samples per task
- * @param leave_out_inds (list) List of indexes from the initial data file in the test set
- * @param n_dim (int) The highest possible dimension for the linear model
- * @param n_residual (int) The number of residuals each iteration of SIS acts on
- * @param n_models_store (int) The number of models to store in output files
+ * @param prop (np.ndarray) The value of the property to evaluate the loss function against for the training set
+ * @param prpo_test (np.ndarray) The value of the property to evaluate the loss function against for the test set
+ * @param task_sizes_train (list) Number of training samples per task
+ * @param task_sizes_test (list) Number of testing samples per task
+ * @param leave_out_inds (list) List of indexes from the initial data file in the test set
+ * @param n_dim (int) The maximum number of features allowed in the linear model
+ * @param n_residual (int) Number of residuals to pass to the next SIS operation
+ * @param n_models_store (int) The number of models to output to files
*/
SISSOClassifier(
std::shared_ptr<FeatureSpace> feat_space,
@@ -166,17 +164,17 @@ public:
/**
* @brief Create a SISSOClassifier object
*
- * @param feat_space (FeatureSpace) The FeatureSpace object that will be associated with this classifier (Defines SIS)
- * @param prop_label (str) The label used to identify the property column in the DataFrame
+ * @param feat_space (FeatureSpace) Feature Space for the problem
+ * @param prop_label (str) The label for the property
* @param prop_unit (Unit) The unit of the property
- * @param prop (list) The property vector for all of the training samples
- * @param prpo_test (list) The property vector for all of the test samples
- * @param task_sizes_train (list) The number of training samples per task
- * @param task_sizes_test (list) The number of test samples per task
- * @param leave_out_inds (list) List of indexes from the initial data file in the test set
- * @param n_dim (int) The highest possible dimension for the linear model
- * @param n_residual (int) The number of residuals each iteration of SIS acts on
- * @param n_models_store (int) The number of models to store in output files
+ * @param prop (list) The value of the property to evaluate the loss function against for the training set
+ * @param prpo_test (list) The value of the property to evaluate the loss function against for the test set
+ * @param task_sizes_train (list) Number of training samples per task
+ * @param task_sizes_test (list) Number of testing samples per task
+ * @param leave_out_inds (list) List of indexes from the initial data file in the test set
+ * @param n_dim (int) The maximum number of features allowed in the linear model
+ * @param n_residual (int) Number of residuals to pass to the next SIS operation
+ * @param n_models_store (int) The number of models to output to files
*/
SISSOClassifier(
std::shared_ptr<FeatureSpace> feat_space,
diff --git a/src/descriptor_identifier/SISSO_DI/SISSOLogRegressor.cpp b/src/descriptor_identifier/SISSO_DI/SISSOLogRegressor.cpp
index 2373961ae95f3835b1c6b9568181c22e74a044b1..d46b45b615e4e27695674f8b14331c621b38607d 100644
--- a/src/descriptor_identifier/SISSO_DI/SISSOLogRegressor.cpp
+++ b/src/descriptor_identifier/SISSO_DI/SISSOLogRegressor.cpp
@@ -87,7 +87,7 @@ void SISSOLogRegressor::add_models(const std::vector<std::vector<int>> indexes)
_loss->reset_projection_prop(min_nodes);
}
-void SISSOLogRegressor::l0_norm(const int n_dim)
+void SISSOLogRegressor::l0_regularization(const int n_dim)
{
int n_get_models = std::max(_n_residual, _n_models_store);
diff --git a/src/descriptor_identifier/SISSO_DI/SISSOLogRegressor.hpp b/src/descriptor_identifier/SISSO_DI/SISSOLogRegressor.hpp
index 01133fb4b15a0f9333c0e7d00555afb3bc191758..448aae513c4afaf973aac82f14cd76c522dc3b20 100644
--- a/src/descriptor_identifier/SISSO_DI/SISSOLogRegressor.hpp
+++ b/src/descriptor_identifier/SISSO_DI/SISSOLogRegressor.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file descriptor_identifier/SISSOLogRegressor.hpp
- * @brief Perform SISSO on a previously generated Feature Space
+ * @brief Perform SISSO on a previously generated Feature Space
*
* Takes in a feature space and performs SISSO on it while storing the selected features in _models
*
@@ -44,18 +44,18 @@ public:
/**
* @brief Create a SISSOLogRegressor object
*
- * @param feat_space The FeatureSpace object that will be associated with this classifier (Defines SIS)
- * @param prop_label The label used to identify the property column in the DataFrame
+ * @param feat_space Feature Space for the problem
+ * @param prop_label The label for the property
* @param prop_unit The unit of the property
- * @param prop The property vector for all of the training samples
- * @param prpo_test The property vector for all of the test samples
- * @param task_sizes_train The number of training samples per task
- * @param task_sizes_test The number of test samples per task
- * @param leave_out_inds List of indexes from the initial data file in the test set
- * @param n_dim The highest possible dimension for the linear model
- * @param n_residual The number of residuals each iteration of SIS acts on
- * @param n_models_store The number of models to store in output files
- * @param fix_intrecept If true fix bias term to 0
+ * @param prop The value of the property to evaluate the loss function against for the training set
+ * @param prpo_test The value of the property to evaluate the loss function against for the test set
+ * @param task_sizes_train Number of training samples per task
+ * @param task_sizes_test Number of testing samples per task
+ * @param leave_out_inds List of indexes from the initial data file in the test set
+ * @param n_dim The maximum number of features allowed in the linear model
+ * @param n_residual Number of residuals to pass to the next SIS operation
+ * @param n_models_store The number of models to output to files
+ * @param fix_intrecept If true the bias term is fixed at 0
*/
SISSOLogRegressor(
const std::shared_ptr<FeatureSpace> feat_space,
@@ -73,25 +73,23 @@ public:
);
/**
- * @brief Adds a model to the model list
+ * @brief Create a Model for a given set of features and store them in _models
*
- * @param indexes Vector storing all of the indexes of features in _phi_selected to use for the model
+ * @param indexes Vector storing all of the indexes of features in _feat_space->phi_selected() to use for the model
*/
void add_models(const std::vector<std::vector<int>> indexes);
/**
* @brief Output the models to files and copy the residuals
- *
- * @param residual vector where all of the residuals are stored
*/
void output_models();
/**
- * @brief Preform the l0 normalization for a property or the residual
+ * @brief Preform an l0-Regularization to find the best n_dim dimensional model
*
- * @param n_dim the dimensionality of the model
+ * @param n_dim The number of features to use in the linear model
*/
- void l0_norm(const int n_dim);
+ void l0_regularization(const int n_dim);
/**
* @brief The selected models (n_dim, n_models_store)
@@ -104,18 +102,18 @@ public:
/**
* @brief Create a SISSOLogRegressor object
*
- * @param feat_space (FeatureSpace) The FeatureSpace object that will be associated with this classifier (Defines SIS)
- * @param prop_label (str) The label used to identify the property column in the DataFrame
+ * @param feat_space (FeatureSpace) Feature Space for the problem
+ * @param prop_label (str) The label for the property
* @param prop_unit (Unit) The unit of the property
- * @param prop (np.ndarray) The property vector for all of the training samples
- * @param prpo_test (np.ndarray) The property vector for all of the test samples
- * @param task_sizes_train (list) The number of training samples per task
- * @param task_sizes_test (list) The number of test samples per task
- * @param leave_out_inds (list) List of indexes from the initial data file in the test set
- * @param n_dim (int) The highest possible dimension for the linear model
- * @param n_residual (int) The number of residuals each iteration of SIS acts on
- * @param n_models_store (int) The number of models to store in output files
- * @param fix_intrecept (bool) If true fix bias term to 0
+ * @param prop (np.ndarray) The value of the property to evaluate the loss function against for the training set
+ * @param prpo_test (np.ndarray) The value of the property to evaluate the loss function against for the test set
+ * @param task_sizes_train (list) Number of training samples per task
+ * @param task_sizes_test (list) Number of testing samples per task
+ * @param leave_out_inds (list) List of indexes from the initial data file in the test set
+ * @param n_dim (int) The maximum number of features allowed in the linear model
+ * @param n_residual (int) Number of residuals to pass to the next SIS operation
+ * @param n_models_store (int) The number of models to output to files
+ * @param fix_intrecept (bool) If true the bias term is fixed at 0
*/
SISSOLogRegressor(
std::shared_ptr<FeatureSpace> feat_space,
@@ -136,18 +134,18 @@ public:
/**
* @brief Create a SISSOLogRegressor object
*
- * @param feat_space (FeatureSpace) The FeatureSpace object that will be associated with this classifier (Defines SIS)
- * @param prop_label (str) The label used to identify the property column in the DataFrame
+ * @param feat_space (FeatureSpace) Feature Space for the problem
+ * @param prop_label (str) The label for the property
* @param prop_unit (Unit) The unit of the property
- * @param prop (list) The property vector for all of the training samples
- * @param prpo_test (list) The property vector for all of the test samples
- * @param task_sizes_train (list) The number of training samples per task
- * @param task_sizes_test (list) The number of test samples per task
- * @param leave_out_inds (list) List of indexes from the initial data file in the test set
- * @param n_dim (int) The highest possible dimension for the linear model
- * @param n_residual (int) The number of residuals each iteration of SIS acts on
- * @param n_models_store (int) The number of models to store in output files
- * @param fix_intrecept (bool) If true fix bias term to 0
+ * @param prop (list) The value of the property to evaluate the loss function against for the training set
+ * @param prpo_test (list) The value of the property to evaluate the loss function against for the test set
+ * @param task_sizes_train (list) Number of training samples per task
+ * @param task_sizes_test (list) Number of testing samples per task
+ * @param leave_out_inds (list) List of indexes from the initial data file in the test set
+ * @param n_dim (int) The maximum number of features allowed in the linear model
+ * @param n_residual (int) Number of residuals to pass to the next SIS operation
+ * @param n_models_store (int) The number of models to output to files
+ * @param fix_intrecept (bool) If true the bias term is fixed at 0
*/
SISSOLogRegressor(
std::shared_ptr<FeatureSpace> feat_space,
diff --git a/src/descriptor_identifier/SISSO_DI/SISSORegressor.cpp b/src/descriptor_identifier/SISSO_DI/SISSORegressor.cpp
index dac7160641725122fdaf180b20a5cfff4414bd35..bde3be0701dd1e0cec6658afea6293830bf4bd58 100644
--- a/src/descriptor_identifier/SISSO_DI/SISSORegressor.cpp
+++ b/src/descriptor_identifier/SISSO_DI/SISSORegressor.cpp
@@ -87,7 +87,7 @@ void SISSORegressor::output_models()
}
}
-void SISSORegressor::l0_norm(const int n_dim)
+void SISSORegressor::l0_regularization(const int n_dim)
{
int n_get_models = std::max(_n_residual, _n_models_store);
std::vector<double> coefs(n_dim + 1, 0.0);
@@ -209,7 +209,7 @@ void SISSORegressor::fit()
}
start = omp_get_wtime();
- l0_norm(dd);
+ l0_regularization(dd);
_mpi_comm->barrier();
duration = omp_get_wtime() - start;
diff --git a/src/descriptor_identifier/SISSO_DI/SISSORegressor.hpp b/src/descriptor_identifier/SISSO_DI/SISSORegressor.hpp
index 4b7f1f59065c1c989bc7a20950db73cca545e078..4a6ce8ee594948db08466a4d2271ec4db336cba0 100644
--- a/src/descriptor_identifier/SISSO_DI/SISSORegressor.hpp
+++ b/src/descriptor_identifier/SISSO_DI/SISSORegressor.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file descriptor_identifier/SISSORegressor.hpp
- * @brief Perform SISSO on a previously generated Feature Space
+ * @brief Perform SISSO on a previously generated Feature Space
*
* Takes in a feature space and performs SISSO on it while storing the selected features in _models
*
@@ -44,18 +44,18 @@ public:
/**
* @brief Create a SISSORegressor object
*
- * @param feat_space The FeatureSpace object that will be associated with this classifier (Defines SIS)
- * @param prop_label The label used to identify the property column in the DataFrame
+ * @param feat_space Feature Space for the problem
+ * @param prop_label The label for the property
* @param prop_unit The unit of the property
- * @param prop The property vector for all of the training samples
- * @param prpo_test The property vector for all of the test samples
- * @param task_sizes_train The number of training samples per task
- * @param task_sizes_test The number of test samples per task
- * @param leave_out_inds List of indexes from the initial data file in the test set
- * @param n_dim The highest possible dimension for the linear model
- * @param n_residual The number of residuals each iteration of SIS acts on
- * @param n_models_store The number of models to store in output files
- * @param fix_intrecept If true fix bias term to 0
+ * @param prop The value of the property to evaluate the loss function against for the training set
+ * @param prpo_test The value of the property to evaluate the loss function against for the test set
+ * @param task_sizes_train Number of training samples per task
+ * @param task_sizes_test Number of testing samples per task
+ * @param leave_out_inds List of indexes from the initial data file in the test set
+ * @param n_dim The maximum number of features allowed in the linear model
+ * @param n_residual Number of residuals to pass to the next SIS operation
+ * @param n_models_store The number of models to output to files
+ * @param fix_intrecept If true the bias term is fixed at 0
*/
SISSORegressor(
const std::shared_ptr<FeatureSpace> feat_space,
@@ -73,29 +73,27 @@ public:
);
/**
- * @brief Adds a model to the model list
+ * @brief Create a Model for a given set of features and store them in _models
*
- * @param indexes Vector storing all of the indexes of features in _phi_selected to use for the model
+ * @param indexes Vector storing all of the indexes of features in _feat_space->phi_selected() to use for the model
*/
virtual void add_models(const std::vector<std::vector<int>> indexes);
/**
* @brief Output the models to files and copy the residuals
- *
- * @param residual vector where all of the residuals are stored
*/
virtual void output_models();
/**
- * @brief Preform the l0 normalization for a property or the residual
+ * @brief Preform an l0-Regularization to find the best n_dim dimensional model
*
- * @param n_dim the dimensionality of the model
+ * @param n_dim The number of features to use in the linear model
*/
- virtual void l0_norm(const int n_dim);
+ virtual void l0_regularization(const int n_dim);
// DocString: sisso_reg_fit
/**
- * @brief Perform the SISSO calculation to find the best model for each dimension up to n_dim dimensions equal to _n_dim
+ * @brief Iteratively run SISSO on the FeatureSpace an Property vector until the maximum dimenisonality is reached
*/
virtual void fit();
@@ -111,18 +109,18 @@ public:
/**
* @brief Create a SISSORegressor object
*
- * @param feat_space (FeatureSpace) The FeatureSpace object that will be associated with this classifier (Defines SIS)
- * @param prop_label (str) The label used to identify the property column in the DataFrame
+ * @param feat_space (FeatureSpace) Feature Space for the problem
+ * @param prop_label (str) The label for the property
* @param prop_unit (Unit) The unit of the property
- * @param prop (np.ndarray) The property vector for all of the training samples
- * @param prpo_test (np.ndarray) The property vector for all of the test samples
- * @param task_sizes_train (list) The number of training samples per task
- * @param task_sizes_test (list) The number of test samples per task
- * @param leave_out_inds (list) List of indexes from the initial data file in the test set
- * @param n_dim (int) The highest possible dimension for the linear model
- * @param n_residual (int) The number of residuals each iteration of SIS acts on
- * @param n_models_store (int) The number of models to store in output files
- * @param fix_intrecept (bool) If true fix bias term to 0
+ * @param prop (np.ndarray) The value of the property to evaluate the loss function against for the training set
+ * @param prpo_test (np.ndarray) The value of the property to evaluate the loss function against for the test set
+ * @param task_sizes_train (list) Number of training samples per task
+ * @param task_sizes_test (list) Number of testing samples per task
+ * @param leave_out_inds (list) List of indexes from the initial data file in the test set
+ * @param n_dim (int) The maximum number of features allowed in the linear model
+ * @param n_residual (int) Number of residuals to pass to the next SIS operation
+ * @param n_models_store (int) The number of models to output to files
+ * @param fix_intrecept (bool) If true the bias term is fixed at 0
*/
SISSORegressor(
std::shared_ptr<FeatureSpace> feat_space,
@@ -143,18 +141,18 @@ public:
/**
* @brief Create a SISSORegressor object
*
- * @param feat_space (FeatureSpace) The FeatureSpace object that will be associated with this classifier (Defines SIS)
- * @param prop_label (str) The label used to identify the property column in the DataFrame
+ * @param feat_space (FeatureSpace) Feature Space for the problem
+ * @param prop_label (str) The label for the property
* @param prop_unit (Unit) The unit of the property
- * @param prop (list) The property vector for all of the training samples
- * @param prpo_test (list) The property vector for all of the test samples
- * @param task_sizes_train (list) The number of training samples per task
- * @param task_sizes_test (list) The number of test samples per task
- * @param leave_out_inds (list) List of indexes from the initial data file in the test set
- * @param n_dim (int) The highest possible dimension for the linear model
- * @param n_residual (int) The number of residuals each iteration of SIS acts on
- * @param n_models_store (int) The number of models to store in output files
- * @param fix_intrecept (bool) If true fix bias term to 0
+ * @param prop (list) The value of the property to evaluate the loss function against for the training set
+ * @param prpo_test (list) The value of the property to evaluate the loss function against for the test set
+ * @param task_sizes_train (list) Number of training samples per task
+ * @param task_sizes_test (list) Number of testing samples per task
+ * @param leave_out_inds (list) List of indexes from the initial data file in the test set
+ * @param n_dim (int) The maximum number of features allowed in the linear model
+ * @param n_residual (int) Number of residuals to pass to the next SIS operation
+ * @param n_models_store (int) The number of models to output to files
+ * @param fix_intrecept (bool) If true the bias term is fixed at 0
*/
SISSORegressor(
std::shared_ptr<FeatureSpace> feat_space,
diff --git a/src/descriptor_identifier/SISSO_DI/SISSO_DI.hpp b/src/descriptor_identifier/SISSO_DI/SISSO_DI.hpp
index f88c831422c302f8298941fa96d2cf551e1eb4b2..a386f3384471e59374c6dc77287c195b742aabbf 100644
--- a/src/descriptor_identifier/SISSO_DI/SISSO_DI.hpp
+++ b/src/descriptor_identifier/SISSO_DI/SISSO_DI.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file descriptor_identifier/SISSO_DI/SISSO_DI.hpp
- * @brief Base class used to perform SISSO on a previously generated Feature Space
+ * @brief Base class used to perform SISSO on a previously generated Feature Space
*
* Takes in a feature space and performs SISSO on it while storing the selected features in _models
*
@@ -42,36 +42,36 @@ protected:
const std::vector<int> _task_sizes_test; //!< Number of testing samples per task
const std::vector<int> _leave_out_inds; //!< List of indexes from the initial data file in the test set
- const Unit _prop_unit; //!< The Unit for the property
- const std::string _prop_label; //!< The label for the property
+ const Unit _prop_unit; //!< The Unit of the property
+ const std::string _prop_label; //!< The label of the property
const std::shared_ptr<FeatureSpace> _feat_space; //!< Feature Space for the problem
- std::shared_ptr<MPI_Interface> _mpi_comm; //!< MPI Communicator
- std::shared_ptr<LossFunction> _loss; //!< The loss function for the calculation
+ std::shared_ptr<MPI_Interface> _mpi_comm; //!< MPI Communicator for the calculation
+ std::shared_ptr<LossFunction> _loss; //!< The loss function used to evaluate models
const int _n_task; //!< The number of tasks
- const int _n_samp; //!< the number of samples per feature
- const int _n_dim; //!< Number of dimensions to calculate
+ const int _n_samp; //!< The number of training samples per feature
+ const int _n_dim; //!< The maximum number of features allowed in the linear model
const int _n_residual; //!< Number of residuals to pass to the next sis model
- const int _n_models_store; //!< The number of models to store in files
+ const int _n_models_store; //!< The number of models to output to files
- const bool _fix_intercept; //!< If true fix intercept to 0
+ const bool _fix_intercept; //!< If true the bias term is fixed at 0
public:
/**
- * @brief Constructor for the Regressor
+ * @brief Constructor for the Solver
*
- * @param feat_space The feature space to run SISSO on
+ * @param feat_space Feature Space for the problem
* @param prop_label The label for the property
* @param prop_unit The unit of the property
- * @param prop Vector storing all data to train the SISSO models with
- * @param prpo_test Vector storing all data to test the SISSO models with
+ * @param prop The value of the property to evaluate the loss function against for the training set
+ * @param prpo_test The value of the property to evaluate the loss function against for the test set
* @param task_sizes_train Number of training samples per task
* @param task_sizes_test Number of testing samples per task
- * @param leave_out_inds List of indexes from the initial data file in the test set
- * @param n_dim Maximum dimensionality of the generated models
+ * @param leave_out_inds List of indexes from the initial data file in the test set
+ * @param n_dim The maximum number of features allowed in the linear model
* @param n_residual Number of residuals to pass to the next SIS operation
- * @param n_models_store Number of features to store in files
- * @param fix_intrecept If true fix intercept to 0
+ * @param n_models_store The number of models to output to files
+ * @param fix_intrecept If true the bias term is fixed at 0
*/
SISSO_DI(
const std::string loss_type,
@@ -90,18 +90,16 @@ public:
);
/**
- * @brief Perform SISSO to generate the models
- * @details Iteratively perform SISSO on the Feature space and property until the model dimensionality is equal to _n_dim
+ * @brief Iteratively run SISSO on the FeatureSpace an Property vector until the maximum dimenisonality is reached
*/
virtual void fit() = 0;
/**
- * @brief Preform the l0 normalization for a property or the residual
+ * @brief Preform an l0-Regularization to find the best n_dim dimensional model
*
- * @param prop The property to fit
- * @param n_dim the dimensionality of the model
+ * @param n_dim The number of features to use in the linear model
*/
- virtual void l0_norm(const int n_dim) = 0;
+ virtual void l0_regularization(const int n_dim) = 0;
/**
* @brief The FeatureSpace object associated with this solver
@@ -131,7 +129,7 @@ public:
// DocString: sisso_di_n_dim
/**
- * @brief The highest possible dimension for the linear model
+ * @brief The maximum number of features allowed in the linear model
*/
inline int n_dim() const {return _n_dim;}
@@ -143,7 +141,7 @@ public:
// DocString: sisso_di_n_modles_store
/**
- * @brief The number of models to store in output files
+ * @brief The number of models to output to files
*/
inline int n_models_store() const {return _n_models_store;}
@@ -152,20 +150,20 @@ public:
// DocString: sisso_di_init_arr
/**
- * @brief Constructor for the Regressor that takes in python objects (cpp definition in <python/descriptor_identifier/SISSO_DI.cpp)
+ * @brief Constructor for the solver that takes in python objects
*
- * @param feat_space The feature space to run SISSO on
+ * @param feat_space Feature Space for the problem
* @param prop_label The label for the property
* @param prop_unit The unit of the property
- * @param prop Vector storing all data to train the SISSO models with
- * @param prpo_test Vector storing all data to test the SISSO models with
+ * @param prop The value of the property to evaluate the loss function against for the training set
+ * @param prpo_test The value of the property to evaluate the loss function against for the test set
* @param task_sizes_train Number of training samples per task
* @param task_sizes_test Number of testing samples per task
- * @param leave_out_inds List of indexes from the initial data file in the test set
- * @param n_dim Maximum dimensionality of the generated models
+ * @param leave_out_inds List of indexes from the initial data file in the test set
+ * @param n_dim The maximum number of features allowed in the linear model
* @param n_residual Number of residuals to pass to the next SIS operation
- * @param n_models_store Number of features to store in files
- * @param fix_intrecept If true fix intercept to 0
+ * @param n_models_store The number of models to output to files
+ * @param fix_intrecept If true the bias term is fixed at 0
*/
SISSO_DI(
std::string loss_type,
@@ -185,20 +183,20 @@ public:
// DocString: sisso_di_init_list
/**
- * @brief Constructor for the Regressor that takes in python objects (cpp definition in <python/descriptor_identifier/SISSO_DI.cpp)
+ * @brief Constructor for the solver that takes in python objects
*
- * @param feat_space The feature space to run SISSO on
+ * @param feat_space Feature Space for the problem
* @param prop_label The label for the property
* @param prop_unit The unit of the property
- * @param prop Vector storing all data to train the SISSO models with
- * @param prpo_test Vector storing all data to test the SISSO models with
+ * @param prop The value of the property to evaluate the loss function against for the training set
+ * @param prpo_test The value of the property to evaluate the loss function against for the test set
* @param task_sizes_train Number of training samples per task
* @param task_sizes_test Number of testing samples per task
- * @param leave_out_inds List of indexes from the initial data file in the test set
- * @param n_dim Maximum dimensionality of the generated models
+ * @param leave_out_inds List of indexes from the initial data file in the test set
+ * @param n_dim The maximum number of features allowed in the linear model
* @param n_residual Number of residuals to pass to the next SIS operation
- * @param n_models_store Number of features to store in files
- * @param fix_intrecept If true fix intercept to 0
+ * @param n_models_store The number of models to output to files
+ * @param fix_intrecept If true the bias term is fixed at 0
*/
SISSO_DI(
std::string loss_type,
diff --git a/src/feature_creation/feature_space/FeatureSpace.cpp b/src/feature_creation/feature_space/FeatureSpace.cpp
index 71aa00ae4416d7be7457bab49ea40e5ad5c4a261..535faf9a8e7627c8d0589aa70f2c39b46d5e4e31 100644
--- a/src/feature_creation/feature_space/FeatureSpace.cpp
+++ b/src/feature_creation/feature_space/FeatureSpace.cpp
@@ -59,9 +59,9 @@ FeatureSpace::FeatureSpace(
std::vector<double> prop,
std::vector<int> task_sizes,
std::string project_type,
- int max_phi,
+ int max_rung,
int n_sis_select,
- int max_store_rung,
+ int n_rung_store,
int n_rung_generate,
double cross_corr_max,
double min_abs_feat_val,
@@ -72,13 +72,13 @@ FeatureSpace::FeatureSpace(
_phi(phi_0),
_phi_0(phi_0),
_end_no_params(1, phi_0.size()),
- _start_gen_reparam(1, 0),
+ _start_rung_reparam(1, 0),
_allowed_param_ops(allowed_param_ops),
_allowed_ops(allowed_ops),
_prop(prop),
_scores(phi_0.size(), 0.0),
_task_sizes(task_sizes),
- _start_gen(1, 0),
+ _start_rung(1, 0),
_feature_space_file("feature_space/selected_features.txt"),
_feature_space_summary_file("feature_space/SIS_summary.txt"),
_project_type(project_type),
@@ -86,11 +86,11 @@ FeatureSpace::FeatureSpace(
_cross_cor_max(cross_corr_max),
_l_bound(min_abs_feat_val),
_u_bound(max_abs_feat_val),
- _max_phi(max_phi),
+ _max_rung(max_rung),
_n_sis_select(n_sis_select),
_n_samp(phi_0[0]->n_samp()),
_n_feat(phi_0.size()),
- _n_rung_store(max_store_rung),
+ _n_rung_store(n_rung_store),
_n_rung_generate(n_rung_generate),
_max_param_depth(max_param_depth),
_reparam_residual(reparam_residual)
@@ -105,9 +105,9 @@ FeatureSpace::FeatureSpace(
std::vector<double> prop,
std::vector<int> task_sizes,
std::string project_type,
- int max_phi,
+ int max_rung,
int n_sis_select,
- int max_store_rung,
+ int n_rung_store,
int n_rung_generate,
double cross_corr_max,
double min_abs_feat_val,
@@ -119,7 +119,7 @@ FeatureSpace::FeatureSpace(
_prop(prop),
_scores(phi_0.size(), 0.0),
_task_sizes(task_sizes),
- _start_gen(1, 0),
+ _start_rung(1, 0),
_feature_space_file("feature_space/selected_features.txt"),
_feature_space_summary_file("feature_space/SIS_summary.txt"),
_project_type(project_type),
@@ -127,11 +127,11 @@ FeatureSpace::FeatureSpace(
_cross_cor_max(cross_corr_max),
_l_bound(min_abs_feat_val),
_u_bound(max_abs_feat_val),
- _max_phi(max_phi),
+ _max_rung(max_rung),
_n_sis_select(n_sis_select),
_n_samp(phi_0[0]->n_samp()),
_n_feat(phi_0.size()),
- _n_rung_store(max_store_rung),
+ _n_rung_store(n_rung_store),
_n_rung_generate(n_rung_generate),
_max_param_depth(0),
_reparam_residual(false)
@@ -145,9 +145,9 @@ void FeatureSpace::initialize_fs()
#ifdef PARAMETERIZE
if(_max_param_depth == -1)
{
- _max_param_depth = _max_phi;
+ _max_param_depth = _max_rung;
}
- if((_max_param_depth < 0) || (_max_param_depth > _max_phi))
+ if((_max_param_depth < 0) || (_max_param_depth > _max_rung))
{
throw std::logic_error("Invalid parameter depth.");
}
@@ -177,20 +177,20 @@ void FeatureSpace::initialize_fs()
if(_n_rung_store == -1)
{
- _n_rung_store = _max_phi - 1;
+ _n_rung_store = _max_rung - 1;
}
if(_n_rung_generate > 1)
{
throw std::logic_error("A maximum of one rung can be generated on the fly.");
}
- else if(_max_phi - _n_rung_generate < _n_rung_store)
+ else if(_max_rung - _n_rung_generate < _n_rung_store)
{
throw std::logic_error("Requesting to store more rungs than what can be pre-generated.");
}
node_value_arrs::set_task_sz_train(_task_sizes);
int n_max_ops = 0;
- for(int rr = 0; rr < _max_phi - _n_rung_store; ++rr)
+ for(int rr = 0; rr < _max_rung - _n_rung_store; ++rr)
{
n_max_ops += std::pow(2, rr);
}
@@ -324,7 +324,7 @@ void FeatureSpace::generate_reparam_feats(
{
for(int rr = 0; rr <= cur_rung; ++rr)
{
- for(auto feat_2 = _phi.begin() + _start_gen[rr]; (feat_2 != feat) || (feat_2 != _phi.begin() + _end_no_params[rr]); ++feat_2)
+ for(auto feat_2 = _phi.begin() + _start_rung[rr]; (feat_2 != feat) || (feat_2 != _phi.begin() + _end_no_params[rr]); ++feat_2)
{
op(feat_set, *feat, *feat_2, feat_ind, l_bound, u_bound, optimizer);
}
@@ -338,7 +338,7 @@ void FeatureSpace::generate_reparam_feats(
{
for(int rr = 0; rr <= cur_rung; ++rr)
{
- for(auto feat_2 = _phi.begin() + _start_gen[rr]; (feat_2 != feat) || (feat_2 != _phi.begin() + _end_no_params[rr]); ++feat_2)
+ for(auto feat_2 = _phi.begin() + _start_rung[rr]; (feat_2 != feat) || (feat_2 != _phi.begin() + _end_no_params[rr]); ++feat_2)
{
op(feat_set, *feat, *feat_2, feat_ind, l_bound, u_bound, optimizer);
op(feat_set, *feat_2, *feat, feat_ind, l_bound, u_bound, optimizer);
@@ -359,10 +359,10 @@ void FeatureSpace::generate_reparam_feature_set(const std::vector<double>& prop)
double l_bound = 1e-50;
std::vector<int> inds;
- for(int nn = 1; nn <= _max_phi - _n_rung_generate; ++nn)
+ for(int nn = 1; nn <= _max_rung - _n_rung_generate; ++nn)
{
node_value_arrs::clear_temp_reg();
- if(nn == _max_phi)
+ if(nn == _max_rung)
{
u_bound = _u_bound;
l_bound = _l_bound;
@@ -377,15 +377,15 @@ void FeatureSpace::generate_reparam_feature_set(const std::vector<double>& prop)
#pragma omp parallel firstprivate(feat_ind, l_bound, u_bound)
{
std::vector<node_ptr> next_phi_private;
- std::shared_ptr<NLOptimizer> optimizer_param = nlopt_wrapper::get_optimizer(_project_type, _task_sizes, _prop, _max_phi, _max_param_depth);
- std::shared_ptr<NLOptimizer> optimizer_reparam = nlopt_wrapper::get_optimizer(_project_type, _task_sizes, prop, _max_phi, _max_param_depth);
+ std::shared_ptr<NLOptimizer> optimizer_param = nlopt_wrapper::get_optimizer(_project_type, _task_sizes, _prop, _max_rung, _max_param_depth);
+ std::shared_ptr<NLOptimizer> optimizer_reparam = nlopt_wrapper::get_optimizer(_project_type, _task_sizes, prop, _max_rung, _max_param_depth);
#ifdef OMP45
#pragma omp for schedule(monotonic: dynamic)
#else
#pragma omp for schedule(dynamic)
#endif
- for(auto feat_1 = _phi_reparam.begin() + _start_gen_reparam.back() + _mpi_comm->rank(); feat_1 < _phi_reparam.end(); feat_1 += _mpi_comm->size())
+ for(auto feat_1 = _phi_reparam.begin() + _start_rung_reparam.back() + _mpi_comm->rank(); feat_1 < _phi_reparam.end(); feat_1 += _mpi_comm->size())
{
generate_non_param_feats(feat_1, next_phi_private, _phi_reparam.begin(), feat_ind, l_bound, u_bound);
generate_param_feats(feat_1, next_phi_private, _phi_reparam.begin(), feat_ind, optimizer_param, l_bound, u_bound);
@@ -396,7 +396,7 @@ void FeatureSpace::generate_reparam_feature_set(const std::vector<double>& prop)
#else
#pragma omp for schedule(dynamic)
#endif
- for(auto feat_1 = _phi.begin() + _start_gen[nn-1] + _mpi_comm->rank(); feat_1 < _phi.begin() + _end_no_params[nn-1]; feat_1 += _mpi_comm->size())
+ for(auto feat_1 = _phi.begin() + _start_rung[nn-1] + _mpi_comm->rank(); feat_1 < _phi.begin() + _end_no_params[nn-1]; feat_1 += _mpi_comm->size())
{
generate_reparam_feats(feat_1, next_phi_private, feat_ind, optimizer_reparam, l_bound, u_bound);
}
@@ -404,9 +404,9 @@ void FeatureSpace::generate_reparam_feature_set(const std::vector<double>& prop)
#pragma omp critical
next_phi.insert(next_phi.end(), next_phi_private.begin(), next_phi_private.end());
}
- _start_gen_reparam.push_back(_phi_reparam.size());
+ _start_rung_reparam.push_back(_phi_reparam.size());
node_value_arrs::clear_temp_reg();
- if(nn < _max_phi)
+ if(nn < _max_rung)
{
int new_phi_size;
int phi_size_start = _phi_reparam.size();
@@ -423,14 +423,14 @@ void FeatureSpace::generate_reparam_feature_set(const std::vector<double>& prop)
// Sort the features to ensure consistent feature spaces for all MPI/OpenMP configurations
std::sort(
- _phi_reparam.begin() + _start_gen_reparam.back(),
+ _phi_reparam.begin() + _start_rung_reparam.back(),
_phi_reparam.end(),
[feat_ind](node_ptr n1, node_ptr n2){return n1->sort_score(feat_ind) < n2->sort_score(feat_ind);}
);
// Reindex sorted features
std::for_each(
- _phi_reparam.begin() + _start_gen_reparam.back(),
+ _phi_reparam.begin() + _start_rung_reparam.back(),
_phi_reparam.end(),
[&feat_ind](node_ptr n){n->reindex(feat_ind); ++feat_ind;}
);
@@ -460,13 +460,13 @@ void FeatureSpace::generate_reparam_feature_set(const std::vector<double>& prop)
}
node_value_arrs::clear_temp_reg();
- if(nn < _max_phi)
+ if(nn < _max_rung)
{
// Remove identical features
std::vector<double> scores(_phi_reparam.size());
_mpi_comm->barrier();
project_funcs::project_r(_prop.data(), scores.data(), _phi_reparam, _task_sizes, 1);
- scores.erase(scores.begin(), scores.begin() + _start_gen_reparam.back());
+ scores.erase(scores.begin(), scores.begin() + _start_rung_reparam.back());
inds = util_funcs::argsort<double>(scores);
std::vector<int> del_inds;
@@ -476,7 +476,7 @@ void FeatureSpace::generate_reparam_feature_set(const std::vector<double>& prop)
for(int sc = 0; sc < scores.size() - 1; ++sc)
{
#ifdef PARAMETERIZE
- if(_phi_reparam[inds[sc] + _start_gen_reparam.back()]->n_params() > 0)
+ if(_phi_reparam[inds[sc] + _start_rung_reparam.back()]->n_params() > 0)
{
continue;
}
@@ -486,8 +486,8 @@ void FeatureSpace::generate_reparam_feature_set(const std::vector<double>& prop)
{
double base_val = std::abs(
util_funcs::r(
- _phi_reparam[_start_gen_reparam.back() + inds[sc]]->value_ptr(),
- _phi_reparam[_start_gen_reparam.back() + inds[sc]]->value_ptr(),
+ _phi_reparam[_start_rung_reparam.back() + inds[sc]]->value_ptr(),
+ _phi_reparam[_start_rung_reparam.back() + inds[sc]]->value_ptr(),
_n_samp
)
);
@@ -496,15 +496,15 @@ void FeatureSpace::generate_reparam_feature_set(const std::vector<double>& prop)
double comp = std::abs(
base_val - std::abs(
util_funcs::r(
- _phi_reparam[_start_gen_reparam.back() + inds[sc]]->value_ptr(),
- _phi_reparam[_start_gen_reparam.back() + inds[sc2]]->value_ptr(0, true),
+ _phi_reparam[_start_rung_reparam.back() + inds[sc]]->value_ptr(),
+ _phi_reparam[_start_rung_reparam.back() + inds[sc2]]->value_ptr(0, true),
_n_samp
)
)
);
if(comp < 1e-10)
{
- del_inds.push_back(-1 * (inds[sc] + _start_gen_reparam.back()));
+ del_inds.push_back(-1 * (inds[sc] + _start_rung_reparam.back()));
break;
}
}
@@ -513,23 +513,23 @@ void FeatureSpace::generate_reparam_feature_set(const std::vector<double>& prop)
{
double base_val = std::abs(
util_funcs::r(
- _phi_reparam[_start_gen_reparam.back() + inds[sc]]->value_ptr(),
- _phi_reparam[_start_gen_reparam.back() + inds[sc]]->value_ptr(),
+ _phi_reparam[_start_rung_reparam.back() + inds[sc]]->value_ptr(),
+ _phi_reparam[_start_rung_reparam.back() + inds[sc]]->value_ptr(),
_n_samp
)
);
double comp = std::abs(
base_val - std::abs(
util_funcs::r(
- _phi_reparam[_start_gen_reparam.back() + inds[sc]]->value_ptr(),
- _phi_reparam[_start_gen_reparam.back() + inds[sc + 1]]->value_ptr(0, true),
+ _phi_reparam[_start_rung_reparam.back() + inds[sc]]->value_ptr(),
+ _phi_reparam[_start_rung_reparam.back() + inds[sc + 1]]->value_ptr(0, true),
_n_samp
)
)
);
if(comp < 1e-10)
{
- del_inds.push_back(-1 * (inds[sc] + _start_gen.back()));
+ del_inds.push_back(-1 * (inds[sc] + _start_rung.back()));
}
}
}
@@ -541,7 +541,7 @@ void FeatureSpace::generate_reparam_feature_set(const std::vector<double>& prop)
}
// Reindex
- for(int ff = _start_gen_reparam.back(); ff < _phi_reparam.size(); ++ff)
+ for(int ff = _start_rung_reparam.back(); ff < _phi_reparam.size(); ++ff)
{
_phi_reparam[ff]->reindex(ff + _n_feat);
}
@@ -592,10 +592,10 @@ void FeatureSpace::generate_feature_space()
double l_bound = 1e-50;
std::vector<int> inds;
- for(int nn = 1; nn <= _max_phi - _n_rung_generate; ++nn)
+ for(int nn = 1; nn <= _max_rung - _n_rung_generate; ++nn)
{
node_value_arrs::clear_temp_reg();
- if(nn == _max_phi)
+ if(nn == _max_rung)
{
u_bound = _u_bound;
l_bound = _l_bound;
@@ -611,14 +611,14 @@ void FeatureSpace::generate_feature_space()
#pragma omp parallel firstprivate(feat_ind, l_bound, u_bound)
{
std::vector<node_ptr> next_phi_private;
- std::shared_ptr<NLOptimizer> optimizer = nlopt_wrapper::get_optimizer(_project_type, _task_sizes, _prop, _max_phi, _max_param_depth);
+ std::shared_ptr<NLOptimizer> optimizer = nlopt_wrapper::get_optimizer(_project_type, _task_sizes, _prop, _max_rung, _max_param_depth);
#ifdef OMP45
#pragma omp for schedule(monotonic: dynamic)
#else
#pragma omp for schedule(dynamic)
#endif
- for(auto feat_1 = _phi.begin() + _start_gen.back() + _mpi_comm->rank(); feat_1 < _phi.end(); feat_1 += _mpi_comm->size())
+ for(auto feat_1 = _phi.begin() + _start_rung.back() + _mpi_comm->rank(); feat_1 < _phi.end(); feat_1 += _mpi_comm->size())
{
generate_non_param_feats(feat_1, next_phi_private, _phi.begin(), feat_ind, l_bound, u_bound);
generate_param_feats(feat_1, next_phi_private, _phi.begin(), feat_ind, optimizer, l_bound, u_bound);
@@ -636,7 +636,7 @@ void FeatureSpace::generate_feature_space()
#else
#pragma omp for schedule(dynamic)
#endif
- for(auto feat_1 = _phi.begin() + _start_gen.back() + _mpi_comm->rank(); feat_1 < _phi.end(); feat_1 += _mpi_comm->size())
+ for(auto feat_1 = _phi.begin() + _start_rung.back() + _mpi_comm->rank(); feat_1 < _phi.end(); feat_1 += _mpi_comm->size())
{
generate_non_param_feats(feat_1, next_phi_private, _phi.begin(), feat_ind, l_bound, u_bound);
}
@@ -645,9 +645,9 @@ void FeatureSpace::generate_feature_space()
next_phi.insert(next_phi.end(), next_phi_private.begin(), next_phi_private.end());
}
#endif
- _start_gen.push_back(_phi.size());
+ _start_rung.push_back(_phi.size());
node_value_arrs::clear_temp_reg();
- if((nn < _max_phi) || (nn <= _n_rung_store) || (_mpi_comm->size() == 1))
+ if((nn < _max_rung) || (nn <= _n_rung_store) || (_mpi_comm->size() == 1))
{
int new_phi_size;
int phi_size_start = _phi.size();
@@ -664,14 +664,14 @@ void FeatureSpace::generate_feature_space()
// Sort the features to ensure consistent feature spaces for all MPI/OpenMP configurations
std::sort(
- _phi.begin() + _start_gen.back(),
+ _phi.begin() + _start_rung.back(),
_phi.end(),
[feat_ind](node_ptr n1, node_ptr n2){return n1->sort_score(feat_ind) < n2->sort_score(feat_ind);}
);
// Reindex sorted features
std::for_each(
- _phi.begin() + _start_gen.back(),
+ _phi.begin() + _start_rung.back(),
_phi.end(),
[&feat_ind](node_ptr n){n->reindex(feat_ind); ++feat_ind;}
);
@@ -701,13 +701,13 @@ void FeatureSpace::generate_feature_space()
}
node_value_arrs::clear_temp_reg();
- if(nn < _max_phi)
+ if(nn < _max_rung)
{
// Remove identical features
_scores.resize(_phi.size());
_mpi_comm->barrier();
project_funcs::project_r(_prop.data(), _scores.data(), _phi, _task_sizes, 1);
- _scores.erase(_scores.begin(), _scores.begin() + _start_gen[_start_gen.size() - 1]);
+ _scores.erase(_scores.begin(), _scores.begin() + _start_rung[_start_rung.size() - 1]);
inds = util_funcs::argsort<double>(_scores);
std::vector<int> del_inds;
@@ -717,7 +717,7 @@ void FeatureSpace::generate_feature_space()
for(int sc = 0; sc < _scores.size() - 1; ++sc)
{
#ifdef PARAMETERIZE
- if(_phi[inds[sc] + _start_gen.back()]->n_params() > 0)
+ if(_phi[inds[sc] + _start_rung.back()]->n_params() > 0)
{
continue;
}
@@ -727,8 +727,8 @@ void FeatureSpace::generate_feature_space()
{
double base_val = std::abs(
util_funcs::r(
- _phi[_start_gen.back() + inds[sc]]->value_ptr(),
- _phi[_start_gen.back() + inds[sc]]->value_ptr(),
+ _phi[_start_rung.back() + inds[sc]]->value_ptr(),
+ _phi[_start_rung.back() + inds[sc]]->value_ptr(),
_n_samp
)
);
@@ -737,15 +737,15 @@ void FeatureSpace::generate_feature_space()
double comp = std::abs(
base_val - std::abs(
util_funcs::r(
- _phi[_start_gen.back() + inds[sc]]->value_ptr(),
- _phi[_start_gen.back() + inds[sc2]]->value_ptr(0, true),
+ _phi[_start_rung.back() + inds[sc]]->value_ptr(),
+ _phi[_start_rung.back() + inds[sc2]]->value_ptr(0, true),
_n_samp
)
)
);
if(comp < 1e-10)
{
- del_inds.push_back(-1 * (inds[sc] + _start_gen.back()));
+ del_inds.push_back(-1 * (inds[sc] + _start_rung.back()));
break;
}
}
@@ -754,23 +754,23 @@ void FeatureSpace::generate_feature_space()
{
double base_val = std::abs(
util_funcs::r(
- _phi[_start_gen.back() + inds[sc]]->value_ptr(),
- _phi[_start_gen.back() + inds[sc]]->value_ptr(),
+ _phi[_start_rung.back() + inds[sc]]->value_ptr(),
+ _phi[_start_rung.back() + inds[sc]]->value_ptr(),
_n_samp
)
);
double comp = std::abs(
base_val - std::abs(
util_funcs::r(
- _phi[_start_gen.back() + inds[sc]]->value_ptr(),
- _phi[_start_gen.back() + inds[sc + 1]]->value_ptr(0, true),
+ _phi[_start_rung.back() + inds[sc]]->value_ptr(),
+ _phi[_start_rung.back() + inds[sc + 1]]->value_ptr(0, true),
_n_samp
)
)
);
if(comp < 1e-10)
{
- del_inds.push_back(-1 * (inds[sc] + _start_gen.back()));
+ del_inds.push_back(-1 * (inds[sc] + _start_rung.back()));
}
}
}
@@ -782,13 +782,13 @@ void FeatureSpace::generate_feature_space()
}
// Reindex
- for(int ff = _start_gen.back(); ff < _phi.size(); ++ff)
+ for(int ff = _start_rung.back(); ff < _phi.size(); ++ff)
{
_phi[ff]->reindex(ff);
}
}
node_value_arrs::clear_temp_reg();
- for(int ff = _start_gen.back(); ff < _phi.size(); ++ff)
+ for(int ff = _start_rung.back(); ff < _phi.size(); ++ff)
{
_phi[ff]->reset_feats(_phi);
}
@@ -797,7 +797,7 @@ void FeatureSpace::generate_feature_space()
{
node_value_arrs::resize_values_arr(nn, _phi.size());
- for(int ff = _start_gen.back(); ff < _phi.size(); ++ff)
+ for(int ff = _start_rung.back(); ff < _phi.size(); ++ff)
{
_phi[ff]->set_value();
_phi[ff]->set_test_value();
@@ -941,7 +941,7 @@ void FeatureSpace::generate_feature_space()
}
#pragma omp parallel for
- for(int ff = _start_gen.back(); ff < _phi.size(); ++ff)
+ for(int ff = _start_rung.back(); ff < _phi.size(); ++ff)
{
_phi[ff]->reindex(ff + n_feat_below_rank, ff);
_phi[ff]->set_value();
@@ -951,23 +951,23 @@ void FeatureSpace::generate_feature_space()
#ifdef PARAMETERIZE
// Reorder features based on the number of parameters they have (none goes first)
- std::vector<int> feat_n_params(_phi.size() - _start_gen.back());
+ std::vector<int> feat_n_params(_phi.size() - _start_rung.back());
std::transform(
- _phi.begin() + _start_gen.back(),
+ _phi.begin() + _start_rung.back(),
_phi.end(),
feat_n_params.begin(),
[](node_ptr feat){return feat->n_params();}
);
inds = util_funcs::argsort<int>(feat_n_params);
next_phi.resize(feat_n_params.size());
- std::copy_n(_phi.begin() + _start_gen.back(), feat_n_params.size(), next_phi.begin());
+ std::copy_n(_phi.begin() + _start_rung.back(), feat_n_params.size(), next_phi.begin());
std::transform(
inds.begin(),
inds.end(),
- _phi.begin() + _start_gen.back(),
+ _phi.begin() + _start_rung.back(),
[&next_phi](int ind){return next_phi[ind];}
);
- for(int ff = _start_gen.back(); ff < _phi.size(); ++ff)
+ for(int ff = _start_rung.back(); ff < _phi.size(); ++ff)
{
_phi[ff]->reindex(ff);
_phi[ff]->set_value();
@@ -982,8 +982,8 @@ void FeatureSpace::generate_feature_space()
_n_feat = _phi.size();
}
-// void FeatureSpace::project_generated(const double* prop, const int size, std::vector<node_ptr>& phi_sel, std::vector<double>& scores_sel)
-void FeatureSpace::project_generated(std::shared_ptr<LossFunction> loss, std::vector<node_ptr>& phi_sel, std::vector<double>& scores_sel)
+// void FeatureSpace::generate_and_project(const double* prop, const int size, std::vector<node_ptr>& phi_sel, std::vector<double>& scores_sel)
+void FeatureSpace::generate_and_project(std::shared_ptr<LossFunction> loss, std::vector<node_ptr>& phi_sel, std::vector<double>& scores_sel)
{
std::vector<double> scores_sel_all(node_value_arrs::N_SELECTED);
if(node_value_arrs::N_SELECTED > _n_sis_select)
@@ -1004,12 +1004,12 @@ void FeatureSpace::project_generated(std::shared_ptr<LossFunction> loss, std::ve
int index_base = _phi.size() + _n_sis_select * (omp_get_thread_num() + _mpi_comm->size());
#ifdef PARAMETERIZE
- std::shared_ptr<NLOptimizer> optimizer = nlopt_wrapper::get_optimizer(_project_type, _task_sizes, _prop, _max_phi, _max_param_depth);
+ std::shared_ptr<NLOptimizer> optimizer = nlopt_wrapper::get_optimizer(_project_type, _task_sizes, _prop, _max_rung, _max_param_depth);
std::shared_ptr<NLOptimizer> reparam_optimizer;
if(_reparam_residual)
{
std::vector<double> prop_vec(loss_copy->prop_project());
- reparam_optimizer = nlopt_wrapper::get_optimizer(_project_type, _task_sizes, prop_vec, _max_phi, _max_param_depth);
+ reparam_optimizer = nlopt_wrapper::get_optimizer(_project_type, _task_sizes, prop_vec, _max_rung, _max_param_depth);
}
else
{
@@ -1022,7 +1022,7 @@ void FeatureSpace::project_generated(std::shared_ptr<LossFunction> loss, std::ve
#else
#pragma omp for schedule(dynamic)
#endif
- for(auto feat = _phi.begin() + _start_gen.back() + _mpi_comm->rank(); feat < _phi.end(); feat += _mpi_comm->size())
+ for(auto feat = _phi.begin() + _start_rung.back() + _mpi_comm->rank(); feat < _phi.end(); feat += _mpi_comm->size())
{
unsigned long int feat_ind = _phi.size() + _n_sis_select * (omp_get_num_threads() + _mpi_comm->size());
@@ -1118,7 +1118,7 @@ void FeatureSpace::project_generated(std::shared_ptr<LossFunction> loss, std::ve
{
scores_sel[worst_score_ind] = scores_sel_private[sc];
phi_sel[worst_score_ind] = phi_sel_private[sc];
- if(phi_sel[worst_score_ind]->rung() == _max_phi)
+ if(phi_sel[worst_score_ind]->rung() == _max_rung)
{
phi_sel[worst_score_ind]->reindex(index_base + worst_score_ind);
}
@@ -1127,7 +1127,7 @@ void FeatureSpace::project_generated(std::shared_ptr<LossFunction> loss, std::ve
{
scores_sel.push_back(scores_sel_private[sc]);
phi_sel.push_back(phi_sel_private[sc]);
- if(phi_sel.back()->rung() == _max_phi)
+ if(phi_sel.back()->rung() == _max_rung)
{
phi_sel.back()->reindex(index_base + phi_sel.size());
}
@@ -1162,7 +1162,7 @@ void FeatureSpace::sis(std::shared_ptr<LossFunction> loss)
double start_time = omp_get_wtime();
_phi.resize(_n_feat);
_phi_reparam.resize(0);
- _start_gen_reparam.resize(0);
+ _start_rung_reparam.resize(0);
generate_reparam_feature_set(loss->prop_project());
_phi.insert(_phi.end(), _phi_reparam.begin(), _phi_reparam.end());
_scores.resize(_phi.size());
@@ -1255,14 +1255,14 @@ void FeatureSpace::sis(std::shared_ptr<LossFunction> loss)
{
#ifdef PARAMETERIZE
_phi.resize(_n_feat);
- _phi.insert(_phi.end(), _phi_reparam.begin() + _start_gen_reparam.back(), _phi_reparam.end());
+ _phi.insert(_phi.end(), _phi_reparam.begin() + _start_rung_reparam.back(), _phi_reparam.end());
_scores.resize(_phi.size());
#endif
start_time = omp_get_wtime();
phi_sel.resize(cur_feat_local);
scores_sel.resize(cur_feat_local);
- project_generated(loss, phi_sel, scores_sel);
+ generate_and_project(loss, phi_sel, scores_sel);
_mpi_comm->barrier();
if(_mpi_comm->rank() == 0)
diff --git a/src/feature_creation/feature_space/FeatureSpace.hpp b/src/feature_creation/feature_space/FeatureSpace.hpp
index 0d587fa324f2b86e610d44de59216645dd31992f..6048fa0d8631569be4d7878757b499d59d577980 100644
--- a/src/feature_creation/feature_space/FeatureSpace.hpp
+++ b/src/feature_creation/feature_space/FeatureSpace.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/feature_space/FeatureSpace.hpp
- * @brief Create a feature space from an initial set of features and algebraic operators
+ * @brief Create a feature space from an initial set of features and algebraic operators
*
* Use an initial set of features and combine them to generate more complicated algebraical features. SIS is also performed here
*
@@ -48,78 +48,77 @@
*/
class FeatureSpace
{
- std::vector<node_ptr> _phi_selected; //!< selected features
- std::vector<node_ptr> _phi; //!< all features
- const std::vector<node_ptr> _phi_0; //!< initial feature space
+ std::vector<node_ptr> _phi_selected; //!< A vector containing all of the selected features
+ std::vector<node_ptr> _phi; //!< A vector containing all features generated (Not including those created on the Fly during SIS)
+ const std::vector<node_ptr> _phi_0; //!< A vector containing all of the Primary features
#ifdef PARAMETERIZE
- std::vector<node_ptr> _phi_reparam; //!< The list of nodes used for reparameterization
- std::vector<int> _end_no_params; //!< The list of indexes of each rung where parameterized nodes start
- std::vector<int> _start_gen_reparam; //!< The list of indexes of each rung where parameterized nodes start
-
- std::vector<un_param_op_node_gen> _un_param_operators; //!< list of all parameterized unary operators with free parameters
- std::vector<bin_param_op_node_gen> _com_bin_param_operators; //!< list of all parameterized commutable binary operators with free parameters
- std::vector<bin_param_op_node_gen> _bin_param_operators; //!< list of all parameterized binary operators with free parameters
- std::vector<std::string> _allowed_param_ops; //!< Map of parameterization operator set (set of operators and non-linear parameters used for a non-linear least squares fit to property)
+ std::vector<node_ptr> _phi_reparam; //!< A vector containing the features created when reparameterizating using the residuals
+ std::vector<int> _end_no_params; //!< A vector containing the indexes of each rung where parameterized nodes start
+ std::vector<int> _start_rung_reparam; //!< A vector containing the indexes of each rung where parameterized nodes start
+
+ std::vector<un_param_op_node_gen> _un_param_operators; //!< Vector containing all parameterized unary operators with free parameters
+ std::vector<bin_param_op_node_gen> _com_bin_param_operators; //!< Vector containing all parameterized commutable binary operators with free parameters
+ std::vector<bin_param_op_node_gen> _bin_param_operators; //!< Vector containing all parameterized binary operators with free parameters
+ std::vector<std::string> _allowed_param_ops; //!< Vector containing all allowed operators strings for operators with free parameters
#endif
- std::vector<std::string> _allowed_ops; //!< list of all allowed operators strings
- std::vector<un_op_node_gen> _un_operators; //!< list of all unary operators
- std::vector<bin_op_node_gen> _com_bin_operators; //!< list of all commutable binary operators
- std::vector<bin_op_node_gen> _bin_operators; //!< list of all binary operators
+ std::vector<std::string> _allowed_ops; //!< Vector containing all allowed operators strings
+ std::vector<un_op_node_gen> _un_operators; //!< Vector containing all unary operators
+ std::vector<bin_op_node_gen> _com_bin_operators; //!< Vector containing all commutable binary operators
+ std::vector<bin_op_node_gen> _bin_operators; //!< Vector containing all binary operators
- std::vector<double> _prop; //!< The property to fit
- std::vector<double> _scores; //!< projection scores for each feature
+ std::vector<double> _prop; //!< The value of the property vector for each training sample
+ std::vector<double> _scores; //!< The projection scores for each feature
- const std::vector<int> _task_sizes; //!< The number of elements in each task (training data)
- std::vector<int> _start_gen; //!< list of the indexes where each generation starts in _phi
- const std::string _project_type; //!< The type of projection that should be done during SIS
- const std::string _feature_space_file; //!< File to store information about the selected features
- const std::string _feature_space_summary_file; //!< File to store information about the selected features
+ const std::vector<int> _task_sizes; //!< Number of training samples per task
+ std::vector<int> _start_rung; //!< Vector containing the indexes where each rung starts in _phi
+ const std::string _project_type; //!< The type of LossFunction to use when projecting the features onto a property
+ const std::string _feature_space_file; //!< File to output the computer readable representation of the selected features to
+ const std::string _feature_space_summary_file; //!< File to output the human readable representation of the selected features to
- std::function<bool(const double*, const int, const double, const std::vector<double>&, const double, const int, const int)> _is_valid; //!< Function used to calculate the scores for SIS
- std::function<bool(const double*, const int, const double, const std::vector<node_ptr>&, const std::vector<double>&, const double)> _is_valid_feat_list; //!< Function used to calculate the scores for SIS without changing omp environment
+ std::function<bool(const double*, const int, const double, const std::vector<double>&, const double, const int, const int)> _is_valid; //!< Function used to determine of a feature is too correlated to previously selected features
+ std::function<bool(const double*, const int, const double, const std::vector<node_ptr>&, const std::vector<double>&, const double)> _is_valid_feat_list; //!< Function used to determine of a feature is too correlated to previously selected features within a given list
- std::shared_ptr<MPI_Interface> _mpi_comm; //!< MPI communicator
+ std::shared_ptr<MPI_Interface> _mpi_comm; //!< the MPI communicator for the calculation
const double _cross_cor_max; //!< Maximum cross-correlation used for selecting features
- const double _l_bound; //!< lower bound for absolute value of the features
- const double _u_bound; //!< upper bound for absolute value of the features
+ const double _l_bound; //!< The lower bound for the maximum absolute value of the features
+ const double _u_bound; //!< The upper bound for the maximum absolute value of the features
- int _n_rung_store; //!< Total rungs stored
- int _n_feat; //!< Total number of features
- int _max_phi; //!< Maximum rung for the feature creation
+ int _n_rung_store; //!< The number of rungs to calculate and store the value of the features for all samples
+ int _n_feat; //!< Total number of features in the feature space
+ int _max_rung; //!< Maximum rung for the feature creation
- const int _n_sis_select; //!< Number of features to select for each dimensions
- const int _n_samp; //!< Number of samples (training data)
- const int _n_rung_generate; //!< Total number of rungs to generate on the fly
+ const int _n_sis_select; //!< Number of features to select during each SIS iteration
+ const int _n_samp; //!< Number of samples in the training set
+ const int _n_rung_generate; //!< Either 0 or 1, and is the number of rungs to generate on the fly during SIS
- int _max_param_depth; //!< Max depth to parameterize a feature (default=_max_rung)
- const bool _reparam_residual; //!< If True then reparameterize using the residuals of each model
+ int _max_param_depth; //!< The maximum depth in the binary expression tree to set non-linear optimization
+ const bool _reparam_residual; //!< If True then reparameterize features using the residuals of each model
public:
#ifdef PARAMETERIZE
/**
- * @brief Constructor for the feature space
- * @details constructs the feature space from an initial set of features and a list of allowed operators
+ * @brief FeatureSpace constructor given a set of primary features and operators
*
* @param mpi_comm MPI communicator for the calculations
- * @param phi_0 The initial set of features to combine
- * @param allowed_ops list of allowed operators
- * @param allowed_param_ops dictionary of the parameterizable operators and their associated free parameters
- * @param prop The property to be learned (training data)
+ * @param phi_0 The set of primary features
+ * @param allowed_ops The list of allowed operators
+ * @param allowed_param_ops The list of allowed operators to be used with non-linear optimization
+ * @param prop List containing the property vector (training data only)
* @param task_sizes The number of samples per task
- * @param project_type The projection operator to use
- * @param max_phi highest rung value for the calculation
- * @param n_sis_select number of features to select during each SIS step
- * @param max_store_rung number of rungs to calculate and store the value of the features for all samples
- * @param n_rung_generate number of rungs to generate on the fly during SIS (this must be 1 or 0 right now, possible to be higher with recursive algorithm)
- * @param cross_corr_max Maximum cross-correlation used for selecting features
- * @param min_abs_feat_val minimum absolute feature value
- * @param max_abs_feat_val maximum absolute feature value
- * @param max_param_depth the maximum paremterization depths for features
- * @param reparam_residual If True then reparameterize using the residuals of each model
+ * @param project_type The type of loss function/projection operator to use
+ * @param max_rung The maximum rung of the feature (Height of the binary expression tree -1)
+ * @param n_sis_select The number of features to select during each SIS step
+ * @param n_rung_store The number of rungs whose feature's data is always stored in memory
+ * @param n_rung_generate Either 0 or 1, and is the number of rungs to generate on the fly during SIS
+ * @param cross_corr_max The maximum allowed cross-correlation value between selected features
+ * @param min_abs_feat_val The minimum allowed absolute feature value for a feature
+ * @param max_abs_feat_val The maximum allowed absolute feature value for a feature
+ * @param max_param_depth The maximum depth in the binary expression tree to set non-linear optimization
+ * @param reparam_residual If True then reparameterize features using the residuals of each model
*/
FeatureSpace(
std::shared_ptr<MPI_Interface> mpi_comm,
@@ -129,9 +128,9 @@ public:
std::vector<double> prop,
std::vector<int> task_sizes,
std::string project_type="regression",
- int max_phi=1,
+ int max_rung=1,
int n_sis_select=1,
- int max_store_rung=-1,
+ int n_rung_store=-1,
int n_rung_generate=0,
double cross_corr_max=1.0,
double min_abs_feat_val=1e-50,
@@ -141,22 +140,21 @@ public:
);
#else
/**
- * @brief Constructor for the feature space
- * @details constructs the feature space from an initial set of features and a list of allowed operators
+ * @brief FeatureSpace constructor given a set of primary features and operators
*
* @param mpi_comm MPI communicator for the calculations
- * @param phi_0 The initial set of features to combine
- * @param allowed_ops list of allowed operators
- * @param prop The property to be learned (training data)
+ * @param phi_0 The set of primary features
+ * @param allowed_ops The list of allowed operators
+ * @param prop List containing the property vector (training data only)
* @param task_sizes The number of samples per task
- * @param project_type The projection operator to use
- * @param max_phi highest rung value for the calculation
- * @param n_sis_select number of features to select during each SIS step
- * @param max_store_rung number of rungs to calculate and store the value of the features for all samples
- * @param n_rung_generate number of rungs to generate on the fly during SIS (this must be 1 or 0 right now, possible to be higher with recursive algorithm)
- * @param cross_corr_max Maximum cross-correlation used for selecting features
- * @param min_abs_feat_val minimum absolute feature value
- * @param max_abs_feat_val maximum absolute feature value
+ * @param project_type The type of loss function/projection operator to use
+ * @param max_rung The maximum rung of the feature (Height of the binary expression tree -1)
+ * @param n_sis_select The number of features to select during each SIS step
+ * @param n_rung_store The number of rungs whose feature's data is always stored in memory
+ * @param n_rung_generate Either 0 or 1, and is the number of rungs to generate on the fly during SIS
+ * @param cross_corr_max The maximum allowed cross-correlation value between selected features
+ * @param min_abs_feat_val The minimum allowed absolute feature value for a feature
+ * @param max_abs_feat_val The maximum allowed absolute feature value for a feature
*/
FeatureSpace(
std::shared_ptr<MPI_Interface> mpi_comm,
@@ -165,9 +163,9 @@ public:
std::vector<double> prop,
std::vector<int> task_sizes,
std::string project_type="regression",
- int max_phi=1,
+ int max_rung=1,
int n_sis_select=1,
- int max_store_rung=-1,
+ int n_rung_store=-1,
int n_rung_generate=0,
double cross_corr_max=1.0,
double min_abs_feat_val=1e-50,
@@ -175,42 +173,42 @@ public:
);
#endif
/**
- * @brief Initialize the feature set given a property vector
+ * @brief Initialize members of the FeatureSpace using _prop
*/
void initialize_fs();
/**
- * @brief Uses _allowed_ops to set the operator lists
+ * @brief Populate the operator lists using _allowed_ops and _allowed_param_ops
*/
void set_op_lists();
/**
- * @brief Initializes the output files for SIS
+ * @brief Create SIS output files and write their headers
*/
void initialize_fs_output_files() const;
+
/**
- * @brief Generate the full feature set from the allowed operators and initial feature set
- * @details populates phi with all features from an initial set and the allowed operators
+ * @brief Populate _phi using _phi_0 and the allowed operators up to (_max_rung - _n_rung_generate)^th rung
*/
void generate_feature_space();
/**
- * @brief The selected feature space
+ * @brief A vector containing all of the selected features
*/
inline std::vector<node_ptr> phi_selected() const {return _phi_selected;};
/**
- * @brief The full feature space
+ * @brief A vector containing all features generated (Not including those created on the Fly during SIS)
*/
inline std::vector<node_ptr> phi() const {return _phi;};
/**
- * @brief The initial feature space
+ * @brief A vector containing all of the Primary features
*/
inline std::vector<node_ptr> phi0() const {return _phi_0;};
/**
- * @brief The vector of projection scores for SIS
+ * @brief The projection scores for each feature in _phi
*/
inline std::vector<double> scores() const {return _scores;}
@@ -220,55 +218,61 @@ public:
inline std::shared_ptr<MPI_Interface> mpi_comm() const {return _mpi_comm;}
/**
- * @brief The vector storing the number of samples in each task
+ * @brief Number of training samples per task
*/
inline std::vector<int> task_sizes() const {return _task_sizes;}
// DocString: feat_space_feature_space_file
/**
- * @brief The feature space filename
+ * @brief Filename of the file to output the computer readable representation of the selected features to
*/
inline std::string feature_space_file() const {return _feature_space_file;}
+ // DocString: feat_space_feature_space_file
+ /**
+ * @brief Filename of the file to output the human readable representation of the selected features to
+ */
+ inline std::string feature_space_summary_file() const {return _feature_space_summary_file;}
+
// DocString: feat_space_l_bound
/**
- * @brief The minimum absolute value of the feature
+ * @brief The mlower bound for the maximum absolute value of the features
*/
inline double l_bound() const {return _l_bound;}
// DocString: feat_space_u_bound
/**
- * @brief The maximum absolute value of the feature
+ * @brief The upper bound for the maximum absolute value of the features
*/
inline double u_bound() const {return _u_bound;}
- // DocString: feat_space_max_phi
+ // DocString: feat_space_max_rung
/**
- * @brief The maximum rung of the feature space
+ * @brief The maximum rung for the feature creation
*/
- inline int max_phi() const {return _max_phi;}
+ inline int max_rung() const {return _max_rung;}
// DocString: feat_space_n_sis_select
/**
- * @brief The number of features selected in each SIS step
+ * @brief The number of features to select during each SIS iteration
*/
inline int n_sis_select() const {return _n_sis_select;}
// DocString: feat_space_n_samp
/**
- * @brief The number of samples per feature
+ * @brief The nuumber of samples in the training set
*/
inline int n_samp() const {return _n_samp;}
// DocString: feat_space_n_feat
/**
- * @brief The number of features in the feature space
+ * @brief The total number of features in the feature space
*/
inline int n_feat() const {return _n_feat;}
// DocString: feat_space_n_rung_store
/**
- * @brief The number of rungs whose feature's data is always stored in memory
+ * @brief The number of rungs to calculate and store the value of the features for all samples
*/
inline int n_rung_store() const {return _n_rung_store;}
@@ -280,14 +284,14 @@ public:
/**
* @brief Generate a new set of non-parameterized features from a single feature
- * @details Take in the feature and perform all valid algebraic operations on it.
+ * @details Perform all valid algebraic operations on the passed feature and all features that appear before it in _phi.
*
* @param feat The feature to spawn new features from
- * @param feat_set The feature set to pull features from for combinations
+ * @param feat_set The feature set to pull features from for binary operations
+ * @param start The point in feat_set to begin pulling features from for binary operations
* @param feat_ind starting index for the next feature generated
- * @param optimizer The object used to optimize the parameterized features
- * @param l_bound lower bound for the absolute value of the feature
- * @param u_bound upper bound for the abosulte value of the feature
+ * @param l_bound lower bound for the maximum absolute value of the feature
+ * @param u_bound upper bound for the maximum abosulte value of the feature
*/
void generate_non_param_feats(
std::vector<node_ptr>::iterator& feat,
@@ -301,14 +305,15 @@ public:
#ifdef PARAMETERIZE
/**
* @brief Generate a new set of parameterized features from a single feature
- * @details Take in the feature and perform all valid algebraic operations on it.
+ * @details Perform all valid algebraic operations on the passed feature and all features that appear before it in _phi.
*
* @param feat The feature to spawn new features from
- * @param feat_set The feature set to pull features from for combinations
+ * @param feat_set The feature set to pull features from for binary operations
+ * @param start The point in feat_set to begin pulling features from for binary operations
* @param feat_ind starting index for the next feature generated
* @param optimizer The object used to optimize the parameterized features
- * @param l_bound lower bound for the absolute value of the feature
- * @param u_bound upper bound for the abosulte value of the feature
+ * @param l_bound lower bound for the maximum absolute value of the feature
+ * @param u_bound upper bound for the maximum abosulte value of the feature
*/
void generate_param_feats(
std::vector<node_ptr>::iterator& feat,
@@ -324,11 +329,11 @@ public:
* @brief Generate a new set of parameterized features for the residuals
*
* @param feat The feature to spawn new features from
- * @param feat_set The feature set to pull features from for combinations
+ * @param feat_set The feature set to pull features from for binary operations
* @param feat_ind starting index for the next feature generated
* @param optimizer The object used to optimize the parameterized features
- * @param l_bound lower bound for the absolute value of the feature
- * @param u_bound upper bound for the abosulte value of the feature
+ * @param l_bound lower bound for the maximum absolute value of the feature
+ * @param u_bound upper bound for the maximum abosulte value of the feature
*/
void generate_reparam_feats(
std::vector<node_ptr>::iterator& feat,
@@ -348,15 +353,13 @@ public:
#endif
/**
- * @brief Calculate the SIS Scores for feature generated on the fly
- * @details Create the next rung of features and calculate their projection scores. Only keep those that can be selected by SIS.
+ * @brief Generate the final rung of features on the fly and calculate their projection scores for SISat can be selected by SIS.
*
- * @param prop Pointer to the start of the vector storing the data to project the features onto
- * @param size The size of the data to project over
- * @param phi_selected The features that would be selected from the previous rungs
- * @param scores_selected The projection scores of the features that would be selected from the previous rungs
+ * @param loss The LossFunction used to project over all of the features
+ * @param phi_selected The set of features that would be selected excluding the final rung
+ * @param scores_selected The projection scores of all features in phi_selected
*/
- void project_generated(std::shared_ptr<LossFunction> loss, std::vector<node_ptr>& phi_selected, std::vector<double>& scores_selected);
+ void generate_and_project(std::shared_ptr<LossFunction> loss, std::vector<node_ptr>& phi_selected, std::vector<double>& scores_selected);
/**
* @brief Perform Sure-Independence Screening over the FeatureSpace. The features are ranked using a projection operator constructed using _project_type and the Property vector
@@ -403,9 +406,9 @@ public:
* @param allowed_param_ops (list) The list of allowed operators to be used with non-linear optimization
* @param prop (list) List containing the property vector (training data only)
* @param project_type (str) The type of loss function/projection operator to use
- * @param max_phi (int) The maximum rung of the feature (Height of the binary expression tree -1)
+ * @param max_rung (int) The maximum rung of the feature (Height of the binary expression tree -1)
* @param n_sis_select (int) The number of features to select during each SIS step
- * @param max_store_rung (int) The number of rungs whose feature's data is always stored in memory
+ * @param n_rung_store (int) The number of rungs whose feature's data is always stored in memory
* @param n_rung_generate (int) Either 0 or 1, and is the number of rungs to generate on the fly during SIS
* @param cross_corr_max (double) The maximum allowed cross-correlation value between selected features
* @param min_abs_feat_val (double) The minimum allowed absolute feature value for a feature
@@ -419,9 +422,9 @@ public:
py::list allowed_param_ops,
py::list prop,
std::string project_type="regression",
- int max_phi=1,
+ int max_rung=1,
int n_sis_select=1,
- int max_store_rung=-1,
+ int n_rung_store=-1,
int n_rung_generate=0,
double cross_corr_max=1.0,
double min_abs_feat_val=1e-50,
@@ -439,9 +442,9 @@ public:
* @param allowed_param_ops (list) The list of allowed operators to be used with non-linear optimization
* @param prop (np.ndarray) List containing the property vector (training data only)
* @param project_type (str) The type of loss function/projection operator to use
- * @param max_phi (int) The maximum rung of the feature (Height of the binary expression tree -1)
+ * @param max_rung (int) The maximum rung of the feature (Height of the binary expression tree -1)
* @param n_sis_select (int) The number of features to select during each SIS step
- * @param max_store_rung (int) The number of rungs whose feature's data is always stored in memory
+ * @param n_rung_store (int) The number of rungs whose feature's data is always stored in memory
* @param n_rung_generate (int) Either 0 or 1, and is the number of rungs to generate on the fly during SIS
* @param cross_corr_max (double) The maximum allowed cross-correlation value between selected features
* @param min_abs_feat_val (double) The minimum allowed absolute feature value for a feature
@@ -455,9 +458,9 @@ public:
py::list allowed_param_ops,
np::ndarray prop,
std::string project_type="regression",
- int max_phi=1,
+ int max_rung=1,
int n_sis_select=1,
- int max_store_rung=-1,
+ int n_rung_store=-1,
int n_rung_generate=0,
double cross_corr_max=1.0,
double min_abs_feat_val=1e-50,
@@ -476,9 +479,9 @@ public:
* @param allowed_ops (list) The list of allowed operators
* @param prop (list) List containing the property vector (training data only)
* @param project_type (str) The type of loss function/projection operator to use
- * @param max_phi (int) The maximum rung of the feature (Height of the binary expression tree -1)
+ * @param max_rung (int) The maximum rung of the feature (Height of the binary expression tree -1)
* @param n_sis_select (int) The number of features to select during each SIS step
- * @param max_store_rung (int) The number of rungs whose feature's data is always stored in memory
+ * @param n_rung_store (int) The number of rungs whose feature's data is always stored in memory
* @param n_rung_generate (int) Either 0 or 1, and is the number of rungs to generate on the fly during SIS
* @param cross_corr_max (double) The maximum allowed cross-correlation value between selected features
* @param min_abs_feat_val (double) The minimum allowed absolute feature value for a feature
@@ -489,9 +492,9 @@ public:
py::list allowed_ops,
py::list prop,
std::string project_type="regression",
- int max_phi=1,
+ int max_rung=1,
int n_sis_select=1,
- int max_store_rung=-1,
+ int n_rung_store=-1,
int n_rung_generate=0,
double cross_corr_max=1.0,
double min_abs_feat_val=1e-50,
@@ -506,9 +509,9 @@ public:
* @param allowed_ops (list) The list of allowed operators
* @param prop (np.ndarray) List containing the property vector (training data only)
* @param project_type (str) The type of loss function/projection operator to use
- * @param max_phi (int) The maximum rung of the feature (Height of the binary expression tree -1)
+ * @param max_rung (int) The maximum rung of the feature (Height of the binary expression tree -1)
* @param n_sis_select (int) The number of features to select during each SIS step
- * @param max_store_rung (int) The number of rungs whose feature's data is always stored in memory
+ * @param n_rung_store (int) The number of rungs whose feature's data is always stored in memory
* @param n_rung_generate (int) Either 0 or 1, and is the number of rungs to generate on the fly during SIS
* @param cross_corr_max (double) The maximum allowed cross-correlation value between selected features
* @param min_abs_feat_val (double) The minimum allowed absolute feature value for a feature
@@ -519,9 +522,9 @@ public:
py::list allowed_ops,
np::ndarray prop,
std::string project_type="regression",
- int max_phi=1,
+ int max_rung=1,
int n_sis_select=1,
- int max_store_rung=-1,
+ int n_rung_store=-1,
int n_rung_generate=0,
double cross_corr_max=1.0,
double min_abs_feat_val=1e-50,
@@ -599,21 +602,21 @@ public:
// DocString: feat_space_phi_selected_py
/**
- * @brief A list of all of the selected features
+ * @brief A list containing all of the selected features
*/
py::list phi_selected_py();
// DocString: feat_space_phi0_py
/**
- * @brief A list containing all of the primary features
+ * @brief A list containing all features generated (Not including those created on the Fly during SIS)
*/
- py::list phi0_py();
+ py::list phi_py();
// DocString: feat_space_phi_py
/**
- * @brief A list of all features in the FeatureSpace
+ * @brief A list containing all of the Primary features
*/
- py::list phi_py();
+ py::list phi0_py();
// DocString: feat_space_scores_py
/**
@@ -633,11 +636,11 @@ public:
*/
inline py::list allowed_ops_py(){return python_conv_utils::to_list<std::string>(_allowed_ops);}
- // DocString: feat_space_start_gen_py
+ // DocString: feat_space_start_rung_py
/**
* @brief A list containing the index of the first feature of each rung in the feature space.
*/
- inline py::list start_gen_py(){return python_conv_utils::to_list<int>(_start_gen);}
+ inline py::list start_rung_py(){return python_conv_utils::to_list<int>(_start_rung);}
// DocString: feat_space_get_feature
/**
diff --git a/src/feature_creation/node/FeatureNode.hpp b/src/feature_creation/node/FeatureNode.hpp
index c56ad79ba7aaa9749d33e895d99f94e2387e51df..f5cb3e900ca424efedc620273d57b0b4199a5a28 100644
--- a/src/feature_creation/node/FeatureNode.hpp
+++ b/src/feature_creation/node/FeatureNode.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/FeatureNode.hpp
- * @brief The class for describing the initial features
+ * @brief The class for describing the initial features
*
* This class represents the leaves of all of the binary expression trees.
*
@@ -56,8 +56,8 @@ class FeatureNode: public Node
}
protected:
- std::vector<double> _value; //!< values for the feature
- std::vector<double> _test_value; //!< test values for the feature
+ std::vector<double> _value; //!< The value for this feature's training data
+ std::vector<double> _test_value; //!< The value for this feature's test data
Unit _unit; //!< Unit for the feature
std::string _expr; //!< Expression of the feature
@@ -71,12 +71,12 @@ public:
/**
* @brief Constructs a feature node
*
- * @param feat_ind index of the feature
+ * @param feat_ind The index of the feature
* @param expr Expression for the feature
- * @param value Value of the feature for each sample
- * @param test_value Value of the feature for each test sample
+ * @param value The value for this feature's training data
+ * @param test_value The value for this feature's test data
* @param unit Unit of the feature
- * @param set_val if true set the value inside the VALUES_ARR object
+ * @param set_val If true set the value inside the VALUES_ARR object
*/
FeatureNode(
const unsigned long int feat_ind,
@@ -86,15 +86,16 @@ public:
const Unit unit,
const bool set_val=true
);
+
#ifdef PY_BINDINGS
// DocString: feat_node_init_arr
/**
* @brief Constructs a feature node using numpy arrays (cpp definition in python/feature_creation/FeatureNode.cpp)
*
- * @param feat_ind (int) index of the feature
+ * @param feat_ind (int) The index of the feature
* @param expr (str) Expression for the feature
- * @param value (np.ndarray) Value of the feature for each sample
- * @param test_value (np.ndarray) Value of the feature for each test sample
+ * @param value (np.ndarray) The value for this feature's training data
+ * @param test_value (np.ndarray) The value for this feature's test data
* @param unit (Unit) Unit of the feature
*/
FeatureNode(unsigned long int feat_ind, std::string expr, np::ndarray value, np::ndarray test_value, Unit unit);
@@ -103,10 +104,10 @@ public:
/**
* @brief Constructs a feature node using Python lists (cpp definition in python/feature_creation/FeatureNode.cpp)
*
- * @param feat_ind (int) index of the feature
+ * @param feat_ind (int) The index of the feature
* @param expr (str) Expression for the feature
- * @param value (list) Value of the feature for each sample
- * @param test_value (list) Value of the feature for each test sample
+ * @param value (list) The value for this feature's training data
+ * @param test_value (list) The value for this feature's test data
* @param unit (Unit) Unit of the feature
*/
FeatureNode(unsigned long int feat_ind, std::string expr, py::list value, py::list test_value, Unit unit);
@@ -145,16 +146,15 @@ public:
~FeatureNode();
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
virtual inline node_ptr hard_copy()const {return std::make_shared<FeatureNode>(*this);}
/**
- * @brief Resets the features of a operator to ones in the feature_space
- * @details Relink all feature to those in the FeatureSpace in case MPI copies break the links
+ * @brief Relinks all features of an OperatorNodes's _feat member to Nodes inside of phi
*
- * @param phi feature space vector
+ * @param phi A vector containing all Nodes that this Node can relink to
*/
void reset_feats(std::vector<node_ptr>& phi){};
@@ -171,16 +171,16 @@ public:
virtual inline int n_leaves(int cur_n_leaves = 0) const {return cur_n_leaves + 1;}
/**
- * @brief Get the score used to sort the features
+ * @brief Get the score used to sort the features in the feature space (Based on the type of node and _feat_ind)
*
- * @param max_ind the maximum index of the input nodes
- * @return the score used to sort the feature space
+ * @param max_ind The maximum index of the input nodes
+ * @return The score used to sort the feature space
*/
inline unsigned long long sort_score(unsigned int max_ind) const {return _feat_ind;}
// DocString: feat_node_expr_const
/**
- * @brief Get the expression for the feature (From root node down)
+ * @brief A human readable equation representing the feature
*/
inline std::string expr()const{return _expr;}
@@ -197,14 +197,12 @@ public:
inline Unit unit() const {return _unit;}
/**
- * @brief An array containing the values of the training set samples for the feature
-
+ * @brief A vector containing the values of the training set samples for the feature
*/
inline std::vector<double> value() const {return _value;}
/**
- * @brief An array containing the values of the test set samples for the feature
-
+ * @brief A vector containing the values of the test set samples for the feature
*/
inline std::vector<double> test_value() const {return _test_value;}
@@ -231,7 +229,7 @@ public:
{
if(!_selected)
{
- std::copy_n(_test_value.data(), _n_test_samp, test_value_ptr());
+ std::copy_n(_test_value.data(), _n_samp_test, test_value_ptr());
}
}
@@ -283,13 +281,14 @@ public:
// DocString: feat_node_rung
/**
- * @brief return the rung of the feature (Height of the binary expression tree + 1)
+ * @brief return the rung of the feature (Height of the binary expression tree - 1)
*
*/
inline int rung(int cur_rung = 0) const {return cur_rung;}
/**
* @brief Get the primary feature decomposition of a feature
+ *
* @return A map representing the primary feature comprising a feature
*/
std::map<int, int> primary_feature_decomp() const;
@@ -316,8 +315,7 @@ public:
}
/**
- * @brief Get the three character representation of the operator
- * @return the three character representation of the operator
+ * @brief Get the term used in the postfix expression for this Node
*/
inline std::string get_postfix_term() const {return std::to_string(_feat_ind);}
@@ -355,57 +353,56 @@ public:
}
/**
- * @brief update the dictionary used to check if an Add/Sub node is valid
+ * @brief update the dictionary used to check if an Add/Sub/AbsDiff node is valid
*
* @param add_sub_leaves the dictionary used to check if an Add/Sub node is valid
- * @param pl_mn if for an addition node: 1 if for a subtraction node: -1
+ * @param pl_mn 1 for addition and -1 for subtraction
* @param expected_abs_tot The expected absolute sum of all values in add_sub_leaves
*/
void update_add_sub_leaves(std::map<std::string, int>& add_sub_leaves, const int pl_mn, int& expected_abs_tot) const;
/**
- * @brief update the dictionary used to check if a mult/div node is valid
+ * @brief update the dictionary used to check if an Mult/Div node is valid
*
* @param div_mult_leaves the dictionary used to check if an Mult/Div node is valid
- * @param fact Factor used to update leaves (power of the feature)
+ * @param fact amount to increment the element (a primary features) of the dictionary by
* @param expected_abs_tot The expected absolute sum of all values in div_mult_leaves
- *
*/
void update_div_mult_leaves(std::map<std::string, double>& div_mult_leaves, const double fact, double& expected_abs_tot) const;
#ifdef PARAMETERIZE
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
inline std::vector<double> parameters() const {return std::vector<double>();};
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
* @param check_sz If true check the size of the parameters
*/
inline void set_parameters(const std::vector<double> params, const bool check_sz=true){};
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
inline void set_parameters(const double* params){};
// DocString: feat_node_get_params
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to get the paremeters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
void get_parameters(std::shared_ptr<NLOptimizer> optimizer){};
/**
* @brief returns the number of parameters for this feature
*
- * @param n_cur the current number of parameters for the feature
- * @param depth the current depth of the node on the binary expression tree
+ * @param n_cur A recursive helper variable to count the number of parameters of the Node
+ * @param depth How far down a given Node is from the root OperatorNode
* @return the number of parameters (_params.size())
*/
inline int n_params(int n_cur=0, int depth = 1) const {return n_cur;};
@@ -414,9 +411,9 @@ public:
* @brief Set the value of all test samples for the feature inside the central data storage array
*
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param for_comp(bool) If true then the evaluation is used for comparing features
- * @param depth the current depth of the node on the binary expression tree
+ * @param depth How far down a given Node is from the root OperatorNode
*/
inline void set_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const {set_value(offset);};
@@ -424,10 +421,10 @@ public:
* @brief The pointer to where the feature's training data is stored
*
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the current depth of the node on the binary expression tree
- * @returns the pointer to the feature's data
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @returns The pointer to the feature's data
*/
inline double* value_ptr(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const {return value_ptr(offset);};
@@ -435,9 +432,9 @@ public:
* @brief Set the value of all test samples for the feature inside the central data storage array
*
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the current depth of the node on the binary expression tree
+ * @param depth How far down a given Node is from the root OperatorNode
*/
inline void set_test_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const {set_test_value(offset);};
@@ -445,28 +442,28 @@ public:
* @brief The pointer to where the feature's test data is stored
*
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the current depth of the node on the binary expression tree
- * @returns the pointer to the feature's data
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @returns The pointer to the feature's data
*/
inline double* test_value_ptr(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const {return test_value_ptr(offset);};
/**
* @brief The expression of the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return feature expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return A human readable equation representing the feature
*/
inline std::string expr(const double* params, const int depth = 1) const {return _expr;};
/**
* @brief Get the valid LaTeX expression that represents the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return the latexified expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return Get the valid LaTeX expression that represents the feature
*/
inline std::string get_latex_expr(const double* params, const int depth=1) const {return str_utils::latexify(_expr);}
@@ -474,8 +471,8 @@ public:
/**
* @brief Get the string that corresponds to the code needed to evaluate the node in matlab
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
* @return The matlab code for the feature
*/
virtual inline std::string matlab_fxn_expr(const double* params, const int depth=1) const
@@ -484,27 +481,26 @@ public:
}
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Set the upper and lower bounds for the scale and bias term of this Node and its children
*
- * @param lb pointer to the lower bounds data
- * @param ub pointer to the upper bounds data
- * @param from_parent How many parameters are between the start of this node's parameters and its parent
- * @param depth the current depth of the node on the binary expression tree
+ * @param lb A pointer to the location where the lower bounds for the scale and bias term of this Node is set
+ * @param ub A pointer to the location where the upper bounds for the scale and bias term of this Node is set
+ * @param depth How far down a given Node is from the root OperatorNode
*/
inline void set_bounds(double* lb, double* ub, const int depth=1) const {};
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Initialize the scale and bias terms for this Node and its children
*
- * @param params pointer to the parameters vector
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
*/
virtual void initialize_params(double* params, const int depth = 1) const {};
/**
* @brief Calculates the derivative of an operation with respect to the parameters for a given sample
*
- * @param params pointer to the parameters
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param dfdp pointer to where the feature derivative pointers are located
*/
void param_derivative(const double* params, double* dfdp) const {}
@@ -525,7 +521,7 @@ public:
*
* @param grad pointer to the gradient storage
* @param dfdp pointer to where the feature derivative pointers are located
- * @param params pointer to the parameters vector
+ * @param params A pointer to the bias and scale terms for this Node and its children
*/
inline void gradient(double* grad, double* dfdp, const double* params) const {};
#endif
diff --git a/src/feature_creation/node/ModelNode.cpp b/src/feature_creation/node/ModelNode.cpp
index 0ff3883435708347d8d6f37bff269e719b1dea53..b72010520db5db26e6e14a1a6e51e6c95da47c0c 100644
--- a/src/feature_creation/node/ModelNode.cpp
+++ b/src/feature_creation/node/ModelNode.cpp
@@ -21,7 +21,7 @@ ModelNode::ModelNode(
const unsigned long int rung,
const std::string expr,
const std::string latex_expr,
- const std::string expr_postfix,
+ const std::string postfix_expr,
const std::string matlab_fxn_expr,
const std::vector<double> value,
const std::vector<double> test_value,
@@ -30,10 +30,10 @@ ModelNode::ModelNode(
) :
FeatureNode(feat_ind, expr, value, test_value, unit, false),
_value_svm(_n_samp),
- _test_value_svm(_n_test_samp),
+ _test_value_svm(_n_samp_test),
_x_in_expr_list(x_in_expr_list),
_latex_expr(latex_expr),
- _expr_postfix(expr_postfix),
+ _postfix_expr(postfix_expr),
_matlab_fxn_expr(matlab_fxn_expr),
_b_remap_svm(0.0),
_w_remap_svm(1.0),
@@ -66,10 +66,10 @@ ModelNode::ModelNode(
ModelNode::ModelNode(node_ptr in_node) :
FeatureNode(in_node->feat_ind(), in_node->expr(), in_node->value(), in_node->test_value(), in_node->unit(), false),
_value_svm(_n_samp),
- _test_value_svm(_n_test_samp),
+ _test_value_svm(_n_samp_test),
_x_in_expr_list(in_node->get_x_in_expr_list()),
_latex_expr(in_node->latex_expr()),
- _expr_postfix(in_node->postfix_expr()),
+ _postfix_expr(in_node->postfix_expr()),
_matlab_fxn_expr(in_node->matlab_fxn_expr()),
_b_remap_svm(0.0),
_w_remap_svm(1.0),
@@ -101,7 +101,7 @@ ModelNode::ModelNode(node_ptr in_node) :
void ModelNode::generate_fxn_list()
{
- std::vector<std::string> postfix_split = str_utils::split_string_trim(_expr_postfix, "|");
+ std::vector<std::string> postfix_split = str_utils::split_string_trim(_postfix_expr, "|");
std::vector<double> params;
int pp = 0;
std::vector<fxn_nterms_pair> fxn_list_rev;
@@ -864,7 +864,7 @@ void ModelNode::generate_fxn_list()
double ModelNode::eval(double* x_in)
{
std::vector<double*> stack;
- std::vector<std::string> postfix_split = str_utils::split_string_trim(_expr_postfix, "|");
+ std::vector<std::string> postfix_split = str_utils::split_string_trim(_postfix_expr, "|");
int ff = 0;
int xx = 0;
@@ -923,7 +923,7 @@ double ModelNode::eval(std::map<std::string, double> x_in_dct)
std::vector<double> ModelNode::eval(std::vector<double>* x_in)
{
std::vector<double*> stack;
- std::vector<std::string> postfix_split = str_utils::split_string_trim(_expr_postfix, "|");
+ std::vector<std::string> postfix_split = str_utils::split_string_trim(_postfix_expr, "|");
int ff = 0;
int xx = 0;
@@ -1027,7 +1027,7 @@ void ModelNode::update_div_mult_leaves(std::map<std::string, double>& div_mult_l
std::map<int, int> ModelNode::primary_feature_decomp() const
{
std::map<int, int> pf_decomp;
- std::vector<std::string> split_postfix = str_utils::split_string_trim(_expr_postfix, "|");
+ std::vector<std::string> split_postfix = str_utils::split_string_trim(_postfix_expr, "|");
for(auto& part : split_postfix)
{
try
@@ -1053,7 +1053,7 @@ std::map<int, int> ModelNode::primary_feature_decomp() const
void ModelNode::update_primary_feature_decomp(std::map<int, int>& pf_decomp) const
{
pf_decomp.clear();
- std::vector<std::string> split_postfix = str_utils::split_string_trim(_expr_postfix, "|");
+ std::vector<std::string> split_postfix = str_utils::split_string_trim(_postfix_expr, "|");
for(auto& part : split_postfix)
{
try
diff --git a/src/feature_creation/node/ModelNode.hpp b/src/feature_creation/node/ModelNode.hpp
index 36d791e77d68626c9206b142e10865f12b464e37..212ec2632a7eb400c2919d99e5fb3e01dfe2882a 100644
--- a/src/feature_creation/node/ModelNode.hpp
+++ b/src/feature_creation/node/ModelNode.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/ModelNode.hpp
- * @brief The base class for describing features selected in SISSO Models
+ * @brief The base class for describing features selected in SISSO Models
*
* This is a standardized feature type for all features selected in SISSO Models. This is created to allow for easier I/O of the models
*
@@ -46,7 +46,7 @@ class ModelNode: public FeatureNode
ar & _value_svm;
ar & _test_value_svm;
ar & _x_in_expr_list;
- ar & _expr_postfix;
+ ar & _postfix_expr;
ar & _latex_expr;
ar & _matlab_fxn_expr;
ar & _b_remap_svm;
@@ -61,16 +61,16 @@ protected:
std::vector<fxn_nterms_pair> _fxn_list; //!< list of functions used for evaluating feature values
- std::vector<std::string> _x_in_expr_list; //!< vector storing the expressions for all primary features that show up in feature in the order they appear in the postfix notation
- std::string _expr_postfix; //!< postfix expression for the feature
- std::string _latex_expr; //!< postfix expression for the feature
- std::string _matlab_fxn_expr; //!< The Matlab code needed to evaluate the feature
+ std::vector<std::string> _x_in_expr_list; //!< A vector storing the expressions for all primary features in the order they appear in the postfix expression
+ std::string _postfix_expr; //!< A computer readable representation of the feature. Primary features represented by their index in phi_0, node types are represented by abbreviations, and the order is the same as the postfix representation of the expression
+ std::string _latex_expr; //!< The valid LaTeX expression that represents the feature
+ std::string _matlab_fxn_expr; //!< The string that corresponds to the code needed to evaluate the node in matlab
- double _b_remap_svm; //!< value to remap the b from SVM to real life
- double _w_remap_svm; //!< value to remap the w from SVM to real life
+ double _b_remap_svm; //!< value to remap the b from SVM to the real value
+ double _w_remap_svm; //!< value to remap the w from SVM to the real value
- int _rung; //!< rung of the feature
- int _n_leaves; //!< Total number of primary features of the feature
+ int _rung; //!< The rung of the feature (Height of the binary expression tree - 1)
+ int _n_leaves; //!< The number of primary features (non-unique) this feature contains (The number of leaves of the Binary Expression Tree)
public:
/**
* @brief Base Constructor
@@ -82,13 +82,13 @@ public:
/**
* @brief Constructs a feature node
*
- * @param feat_ind index of the feature
+ * @param feat_ind The index of the feature
* @param rung the rung of the feature
* @param expr Expression for the feature
- * @param latex_expr The LaTeXified expression for the feature
+ * @param latex_expr Get the valid LaTeX expression that represents the feature for the feature
* @param matlab_fxn_expr The code to evaluate the feature in matlab
- * @param value Value of the feature for each sample
- * @param test_value Value of the feature for each test sample
+ * @param value The value for this feature's training data
+ * @param test_value The value for this feature's test data
* @param x_in_expr_list vector storing the expressions for all primary features that show up in feature in the order they appear in the postfix notation
* @param unit Unit of the feature
*/
@@ -97,7 +97,7 @@ public:
const unsigned long int rung,
const std::string expr,
const std::string latex_expr,
- const std::string expr_postfix,
+ const std::string postfix_expr,
const std::string matlab_fxn_expr,
const std::vector<double> value,
const std::vector<double> test_value,
@@ -146,8 +146,8 @@ public:
~ModelNode();
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
inline node_ptr hard_copy()const {return std::make_shared<ModelNode>(*this);}
@@ -249,10 +249,10 @@ public:
// DocString: model_node_set_value
/**
- * @brief Set the value of all training samples for the feature inside the central data storage array
+ * @brief Set the value of all training samples for the feature inside the central data storage arrays
*
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
- * @param for_comp(bool) If true then the evaluation is used for comparing features
+ * @param for_comp (bool) If true then the evaluation is used for comparing features
*/
inline void set_value(int offset=-1, const bool for_comp=false) const {return;}
@@ -261,7 +261,7 @@ public:
* @brief Set the value of all test samples for the feature inside the central data storage array
*
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
- * @param for_comp(bool) If true then the evaluation is used for comparing features
+ * @param for_comp (bool) If true then the evaluation is used for comparing features
*/
inline void set_test_value(int offset=-1, const bool for_comp=false) const {return;}
@@ -328,8 +328,7 @@ public:
// DocString: model_node_rung
/**
- * @brief return the rung of the feature (Height of the binary expression tree + 1)
- *
+ * @brief return the rung of the feature (Height of the binary expression tree - 1)
*/
inline int rung(int cur_rung = 0) const {return _rung;}
@@ -342,15 +341,15 @@ public:
/**
* @brief Get the primary feature decomposition of a feature
+ *
* @return A map representing the primary feature comprising a feature
*/
std::map<int, int> primary_feature_decomp() const;
/**
- * @brief Get the three character representation of the operator
- * @return the three character representation of the operator
+ * @brief Get the term used in the postfix expression for this Node
*/
- inline std::string get_postfix_term() const {return _expr_postfix;}
+ inline std::string get_postfix_term() const {return _postfix_expr;}
// DocString: model_node_latex_expr
/**
@@ -370,21 +369,20 @@ public:
}
/**
- * @brief update the dictionary used to check if an Add/Sub node is valid
+ * @brief update the dictionary used to check if an Add/Sub/AbsDiff node is valid
*
* @param add_sub_leaves the dictionary used to check if an Add/Sub node is valid
- * @param pl_mn if for an addition node: 1 if for a subtraction node: -1
+ * @param pl_mn 1 for addition and -1 for subtraction
* @param expected_abs_tot The expected absolute sum of all values in add_sub_leaves
*/
void update_add_sub_leaves(std::map<std::string, int>& add_sub_leaves, const int pl_mn, int& expected_abs_tot) const;
/**
- * @brief update the dictionary used to check if a mult/div node is valid
+ * @brief update the dictionary used to check if an Mult/Div node is valid
*
* @param div_mult_leaves the dictionary used to check if an Mult/Div node is valid
- * @param fact Factor used to update leaves (power of the feature)
+ * @param fact amount to increment the element (a primary features) of the dictionary by
* @param expected_abs_tot The expected absolute sum of all values in div_mult_leaves
- *
*/
void update_div_mult_leaves(std::map<std::string, double>& div_mult_leaves, const double fact, double& expected_abs_tot) const;
@@ -393,14 +391,15 @@ public:
/**
* @brief Constructs a feature node
*
- * @param feat_ind (int) index of the feature
+ * @param feat_ind (int)The index of the feature
* @param rung (int) the rung of the feature
* @param expr (str) Expression for the feature
- * @param latex_expr (str) The LaTeXified expression for the feature
+ * @param latex_expr (str) Get the valid LaTeX expression that represents the feature for the feature
+ * @param postfix_expr (str) A computer readable representation of the feature. Primary features represented by their index in phi_0, node types are represented by abbreviations, and the order is the same as the postfix representation of the expression
* @param matlab_fxn_expr (str) The code to evaluate the feature in matlab
- * @param value (list) Value of the feature for each sample
- * @param test_value (list) Value of the feature for each test sample
- * @param x_in_expr_list (list) //!< vector storing the expressions for all primary features that show up in feature in the order they appear in the postfix notation
+ * @param value (list) The value for this feature's training data
+ * @param test_value (list) The value for this feature's test data
+ * @param x_in_expr_list (list) vector storing the expressions for all primary features that show up in feature in the order they appear in the postfix notation
* @param unit (Unit) Unit of the feature
*/
ModelNode(
@@ -408,7 +407,7 @@ public:
const unsigned long int rung,
const std::string expr,
const std::string latex_expr,
- const std::string expr_postfix,
+ const std::string postfix_expr,
const std::string matlab_fxn_expr,
const py::list value,
const py::list test_value,
@@ -420,14 +419,15 @@ public:
/**
* @brief Constructs a feature node
*
- * @param feat_ind (int) index of the feature
+ * @param feat_ind (int)The index of the feature
* @param rung (int) the rung of the feature
* @param expr (str) Expression for the feature
- * @param latex_expr (str) The LaTeXified expression for the feature
+ * @param latex_expr (str) Get the valid LaTeX expression that represents the feature for the feature
+ * @param postfix_expr (str) A computer readable representation of the feature. Primary features represented by their index in phi_0, node types are represented by abbreviations, and the order is the same as the postfix representation of the expression
* @param matlab_fxn_expr (str) The code to evaluate the feature in matlab
- * @param value (np.ndarray) Value of the feature for each sample
- * @param test_value (np.ndarray) Value of the feature for each test sample
- * @param x_in_expr_list (list) //!< vector storing the expressions for all primary features that show up in feature in the order they appear in the postfix notation
+ * @param value (np.ndarray) The value for this feature's training data
+ * @param test_value (np.ndarray) The value for this feature's test data
+ * @param x_in_expr_list (np.ndarray) vector storing the expressions for all primary features that show up in feature in the order they appear in the postfix notation
* @param unit (Unit) Unit of the feature
*/
ModelNode(
@@ -435,7 +435,7 @@ public:
const unsigned long int rung,
const std::string expr,
const std::string latex_expr,
- const std::string expr_postfix,
+ const std::string postfix_expr,
const std::string matlab_fxn_expr,
const np::ndarray value,
const np::ndarray test_value,
@@ -490,7 +490,6 @@ public:
/**
* @brief get the list of feature expressions for each of the leaves
- * @return Python list containing the expression for each leaf in the binary expression tree
*/
inline py::list x_in_expr_list_py(){return python_conv_utils::to_list<std::string>(_x_in_expr_list);}
#endif
@@ -500,8 +499,8 @@ public:
/**
* @brief Get the string that corresponds to the code needed to evaluate the node in matlab
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
* @return The matlab code for the feature
*/
inline std::string matlab_fxn_expr(const double* params, const int depth=1) const
diff --git a/src/feature_creation/node/Node.cpp b/src/feature_creation/node/Node.cpp
index beec8bb39857447a24ce69767b3003926a1ccd70..9fa6a13939cdc4abb9e7f4209726bdd44f465bfa 100644
--- a/src/feature_creation/node/Node.cpp
+++ b/src/feature_creation/node/Node.cpp
@@ -16,8 +16,8 @@
Node::Node()
{}
-Node::Node(const unsigned long int feat_ind, const int n_samp, const int n_test_samp) :
- _n_test_samp(n_test_samp),
+Node::Node(const unsigned long int feat_ind, const int n_samp, const int n_samp_test) :
+ _n_samp_test(n_samp_test),
_n_samp(n_samp),
_feat_ind(feat_ind),
_arr_ind(feat_ind),
diff --git a/src/feature_creation/node/Node.hpp b/src/feature_creation/node/Node.hpp
index 6b4e7a8a325484d2d8cf39750167cf33bf12e0a1..cf160dbb32945cad4835f35ab85265466f464e1b 100644
--- a/src/feature_creation/node/Node.hpp
+++ b/src/feature_creation/node/Node.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/Node.hpp
- * @brief The base class for describing features
+ * @brief The base class for describing features
*
* This package represents the features in SISSO as a binary expression tree and are accessible from the root node of that tree.
* The node class are the vertices of the tree and represent initial features (FeatureNode) and all algebraic operators (OperatorNodes).
@@ -58,7 +58,7 @@ class Node
template <typename Archive>
void serialize(Archive& ar, const unsigned int version)
{
- ar & _n_test_samp;
+ ar & _n_samp_test;
ar & _n_samp;
ar & _feat_ind;
ar & _arr_ind;
@@ -67,11 +67,11 @@ class Node
}
protected:
- int _n_test_samp; //!< Number of samples in the feature's test set
- int _n_samp; //!< Number of samples in the feature's test set
- unsigned long int _feat_ind; //!< Index of the feature
- unsigned long int _arr_ind; //!< Index of the feature for the value arrays
- int _d_mat_ind; //!< Index for the feature matrix
+ int _n_samp; //!< Number of samples in the training set
+ int _n_samp_test; //!< Number of samples in the test set
+ unsigned long int _feat_ind; //!< The unique index ID for the feature
+ unsigned long int _arr_ind; //!< The index that is used to access the data stored in the central data storage
+ int _d_mat_ind; //!< The index used to access the data stored in the description matrix
bool _selected; //!< True if the feature is selected;
public:
@@ -84,11 +84,11 @@ public:
/**
* @brief Constructor that specifies feature index and number of samples
*
- * @param feat_ind index of the feature
- * @param n_samp number of samples in the node
- * @param n_samp number of test samples in the node
+ * @param feat_ind The index of the feature
+ * @param n_samp The number of samples in the training set
+ * @param n_samp_test The number of samples in the test set
*/
- Node(const unsigned long int feat_ind, const int n_samp, int n_test_samp);
+ Node(const unsigned long int feat_ind, const int n_samp, int n_samp_test);
/**
* @brief Copy Constructor
@@ -124,16 +124,15 @@ public:
virtual ~Node();
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
virtual std::shared_ptr<Node> hard_copy()const = 0;
/**
- * @brief Resets the features of a operator to ones in the feature_space
- * @details Relink all feature to those in the FeatureSpace in case MPI copies break the links
+ * @brief Relinks all features of an OperatorNodes's _feat member to Nodes inside of phi
*
- * @param phi feature space vector
+ * @param phi A vector containing all Nodes that this Node can relink to
*/
virtual void reset_feats(std::vector<std::shared_ptr<Node>>& phi) = 0;
@@ -170,10 +169,10 @@ public:
inline void reindex(const unsigned long int feat_ind, const unsigned long int arr_ind){_feat_ind = feat_ind; _arr_ind = arr_ind;}
/**
- * @brief Get the score used to sort the features
+ * @brief Get the score used to sort the features in the feature space (Based on the type of node and _feat_ind)
*
- * @param max_ind the maximum index of the input nodes
- * @return the score used to sort the feature space
+ * @param max_ind The maximum index of the input nodes
+ * @return The score used to sort the feature space
*/
virtual unsigned long long sort_score(unsigned int max_ind) const = 0;
@@ -183,11 +182,11 @@ public:
*/
inline int n_samp() const {return _n_samp;}
- // DocString: node_n_test_samp
+ // DocString: node_n_samp_test
/**
* @brief The number of samples in the test set
*/
- inline int n_test_samp() const {return _n_test_samp;}
+ inline int n_samp_test() const {return _n_samp_test;}
// DocString: node_feat_ind
/**
@@ -210,7 +209,7 @@ public:
/**
* @brief Setter function for _selected
*
- * @param sel is the feature selected?
+ * @param sel True if the feature selected
*/
inline void set_selected(const bool sel){_selected = sel;}
@@ -251,21 +250,18 @@ public:
virtual Unit unit() const = 0;
/**
- * @brief An array containing the values of the training set samples for the feature
-
+ * @brief A vector containing the values of the training set samples for the feature
*/
virtual std::vector<double> value() const = 0;
/**
- * @brief An array containing the values of the test set samples for the feature
-
+ * @brief A vector containing the values of the test set samples for the feature
*/
virtual std::vector<double> test_value() const = 0;
-
// DocString: node_set_value
/**
- * @brief Set the value of all training samples for the feature inside the central data storage array
+ * @brief Set the value of all training samples for the feature inside the central data storage arrays
*
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
@@ -277,6 +273,8 @@ public:
*
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
+ *
+ * @return pointer to the feature's training value
*/
virtual double* value_ptr(int offset=-1, const bool for_comp=false) const = 0;
@@ -294,6 +292,8 @@ public:
*
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
+ *
+ * @return pointer to the feature's test values
*/
virtual double* test_value_ptr(int offset=-1, const bool for_comp=false) const = 0;
@@ -320,7 +320,7 @@ public:
// DocString: node_rung
/**
- * @brief return the rung of the feature (Height of the binary expression tree + 1)
+ * @brief Return the rung of the feature (Height of the binary expression tree - 1)
*
* @param cur_rung (int) A recursive helper counter for the rung
*/
@@ -328,6 +328,7 @@ public:
/**
* @brief Get the primary feature decomposition of a feature
+ *
* @return A map representing the primary feature comprising a feature
*/
virtual std::map<int, int> primary_feature_decomp() const = 0;
@@ -340,12 +341,12 @@ public:
virtual void update_primary_feature_decomp(std::map<int, int>& pf_decomp) const = 0;
/**
- * @brief Converts a feature into a postfix expression (reverse polish notation)
+ * @brief Converts a feature into a postfix expression
*
* @details Recursively creates a postfix representation of the string
*
* @param cur_expr The current expression
- * @param add_params If true include the parameters in teh postfix expression
+ * @param add_params If true include the parameters in the postfix expression
* @return The current postfix expression of the feature
*/
virtual void update_postfix(std::string& cur_expr, const bool add_params=true) const = 0;
@@ -357,14 +358,13 @@ public:
inline std::string postfix_expr() const {std::string cur_expr = ""; update_postfix(cur_expr); return cur_expr.substr(0, cur_expr.size() - 1);}
/**
- * @brief Get a computer readable representation of the feature. Primary features represented by their index in phi_0, node types are represented by abbreviations, and the order is the same as the postfix representation of the expression
- * @return the string representation of the node for the postfix expression
+ * @brief Get the term used in the postfix expression for this Node
*/
virtual std::string get_postfix_term() const = 0;
// DocString: node_matlab_expr
/**
- * @brief Get the string that corresponds to the code needed to evaluate the node in matlab
+ * @brief The string that corresponds to the code needed to evaluate the node in matlab
* @return The matlab code for the feature
*/
virtual std::string matlab_fxn_expr() const = 0;
@@ -373,7 +373,7 @@ public:
* @brief update the dictionary used to check if an Add/Sub/AbsDiff node is valid
*
* @param add_sub_leaves the dictionary used to check if an Add/Sub node is valid
- * @param pl_mn if for an addition node: 1 if for a subtraction node: -1
+ * @param pl_mn 1 for addition and -1 for subtraction
* @param expected_abs_tot The expected absolute sum of all values in add_sub_leaves
*/
virtual void update_add_sub_leaves(std::map<std::string, int>& add_sub_leaves, const int pl_mn, int& expected_abs_tot) const = 0;
@@ -382,7 +382,7 @@ public:
* @brief update the dictionary used to check if an Mult/Div node is valid
*
* @param div_mult_leaves the dictionary used to check if an Mult/Div node is valid
- * @param fact amount to increment the dictionary by
+ * @param fact amount to increment the element (a primary features) of the dictionary by
* @param expected_abs_tot The expected absolute sum of all values in div_mult_leaves
*/
virtual void update_div_mult_leaves(std::map<std::string, double>& div_mult_leaves, const double fact, double& expected_abs_tot) const = 0;
@@ -390,37 +390,37 @@ public:
#ifdef PARAMETERIZE
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
virtual std::vector<double> parameters() const = 0;
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
* @param check_sz If true check the size of the parameters
*/
virtual void set_parameters(const std::vector<double> params, const bool check_sz=true) = 0;
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
virtual void set_parameters(const double* params) = 0;
// DocString: node_get_params
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to get the paremeters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
virtual void get_parameters(std::shared_ptr<NLOptimizer> optimizer) = 0;
/**
* @brief returns the number of parameters for this feature
*
- * @param n_cur the current number of parameters for the feature
- * @param depth the current depth of the node on the binary expression tree
+ * @param n_cur A recursive helper variable to count the number of parameters of the Node
+ * @param depth How far down a given Node is from the root OperatorNode
* @return the number of parameters (_params.size())
*/
virtual int n_params(const int n_cur = 0, const int depth = 1) const = 0;
@@ -429,9 +429,9 @@ public:
* @brief Set the value of all training samples for the feature inside the central data storage array
*
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the current depth of the node in the binary expression tree (from the root)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
virtual void set_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const = 0;
@@ -439,10 +439,10 @@ public:
* @brief The pointer to where the feature's training data is stored
*
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the current depth of the node in the binary expression tree (from the root)
- * @returns the pointer to the feature's data
+ * @param depth (int) How far down a given Node is from the root OperatorNode
+ * @returns The pointer to the feature's data
*/
virtual double* value_ptr(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const = 0;
@@ -450,9 +450,9 @@ public:
* @brief Set the value of all test samples for the feature inside the central data storage array
*
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the current depth of the node in the binary expression tree (from the root)
+ * @param depth How far down a given Node is from the root OperatorNode
*/
virtual void set_test_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const = 0;
@@ -460,28 +460,28 @@ public:
* @brief The pointer to where the feature's test data is stored
*
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the current depth of the node in the binary expression tree (from the root)
- * @returns the pointer to the feature's data
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @returns The pointer to the feature's data
*/
virtual double* test_value_ptr(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const = 0;
/**
* @brief The expression of the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return feature expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return A human readable equation representing the feature
*/
virtual std::string expr(const double* params, const int depth = 1) const = 0;
/**
* @brief Get valid LaTeX equation that represents the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return the latexified expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return Get the valid LaTeX expression that represents the feature
*/
virtual std::string get_latex_expr(const double* params, const int depth=1) const = 0;
@@ -489,34 +489,33 @@ public:
/**
* @brief Get the string that corresponds to the code needed to evaluate the node in matlab
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
* @return The matlab code for the feature
*/
virtual std::string matlab_fxn_expr(const double* params, const int depth=1) const = 0;
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Set the upper and lower bounds for the scale and bias term of this Node and its children
*
- * @param lb pointer to the lower bounds data
- * @param ub pointer to the upper bounds data
- * @param from_parent How many parameters are between the start of this node's parameters and its parent
- * @param depth the current depth of the node on the binary expression tree
+ * @param lb A pointer to the location where the lower bounds for the scale and bias term of this Node is set
+ * @param ub A pointer to the location where the upper bounds for the scale and bias term of this Node is set
+ * @param depth How far down a given Node is from the root OperatorNode
*/
virtual void set_bounds(double* lb, double* ub, const int depth=1) const = 0;
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Initialize the scale and bias terms for this Node and its children
*
- * @param params pointer to the parameters vector
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
*/
virtual void initialize_params(double* params, const int depth = 1) const = 0;
/**
* @brief Calculates the derivative of an operation with respect to the parameters for a given sample
*
- * @param params pointer to the parameters
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param samp_ind sample index number
*/
virtual void param_derivative(const double* params, double* dfdp) const = 0;
@@ -534,7 +533,7 @@ public:
*
* @param grad pointer to the gradient storage
* @param dfdp pointer to where the feature derivative pointers are located
- * @param params pointer to the parameters vector
+ * @param params A pointer to the bias and scale terms for this Node and its children
*/
virtual void gradient(double* grad, double* dfdp, const double* params) const = 0;
#endif
@@ -572,7 +571,7 @@ public:
#ifdef PARAMETERIZE
// DocString: node_parameters_py
/**
- * @brief The vector storing all non-linear parameters of the feature
+ * @brief A list containing all scale and bias terms for this Node
*/
inline py::list parameters_py(){return python_conv_utils::to_list<double>(parameters());}
#endif
@@ -588,6 +587,10 @@ public:
typedef std::shared_ptr<Node> node_ptr;
+/**
+ * @brief Exception used when a Node Object can't be created for some reason
+ *
+ */
class InvalidFeatureException: public std::exception
{
inline const char * what () const throw () {
diff --git a/src/feature_creation/node/operator_nodes/OperatorNode.hpp b/src/feature_creation/node/operator_nodes/OperatorNode.hpp
index 0a368c0bb5df806edd4d2ddcbc2aae2667618878..2a18036bb8e271ba49ed4ed00cb87d48a703b99f 100644
--- a/src/feature_creation/node/operator_nodes/OperatorNode.hpp
+++ b/src/feature_creation/node/operator_nodes/OperatorNode.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/OperatorNode.hpp
- * @brief Base class for representing all algebraic operators in the binary expression tree
+ * @brief Base class for representing all algebraic operators in the binary expression tree
*
* This class represents the non-leaf verticies in the binary expression tree. This also stores all edges of the graph
*
@@ -67,7 +67,7 @@ class OperatorNode: public Node
ar & _feats;
}
protected:
- std::array<node_ptr, N> _feats; //!< The features for the operator nodes to act on (edges of the graph)
+ std::array<node_ptr, N> _feats; //!< The features for the operator nodes to act on (This Node's children)
public:
/**
@@ -81,11 +81,11 @@ public:
* @brief Constructor
* @details Constructs an operator node with a set of features
*
- * @param feats array of features that the operator will act on
- * @param feat_ind index of the feature
+ * @param feats Array of features that the operator will act on
+ * @param feat_ind The index of the feature
*/
OperatorNode(const std::array<node_ptr, N> feats, const unsigned long int feat_ind) :
- Node(feat_ind, feats[0]->n_samp(), feats[0]->n_test_samp()),
+ Node(feat_ind, feats[0]->n_samp(), feats[0]->n_samp_test()),
_feats(feats)
{}
@@ -123,16 +123,15 @@ public:
{}
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
virtual node_ptr hard_copy()const = 0;
/**
- * @brief Resets the features of a operator to ones in the feature_space
- * @details Relink all feature to those in the FeatureSpace in case MPI copies break the links
+ * @brief Relinks all features of an OperatorNodes's _feat member to Nodes inside of phi
*
- * @param phi feature space vector
+ * @param phi A vector containing all Nodes that this Node can relink to
*/
void reset_feats(std::vector<node_ptr>& phi)
{
@@ -167,10 +166,10 @@ public:
}
/**
- * @brief Get the score used to sort the features
+ * @brief Get the score used to sort the features in the feature space (Based on the type of node and _feat_ind)
*
- * @param max_ind the maximum index of the input nodes
- * @return the score used to sort the feature space
+ * @param max_ind The maximum index of the input nodes
+ * @return The score used to sort the feature space
*/
unsigned long long sort_score(unsigned int max_ind) const
{
@@ -198,8 +197,7 @@ public:
virtual std::string get_latex_expr() const = 0;
/**
- * @brief An array containing the values of the training set samples for the feature
-
+ * @brief A vector containing the values of the training set samples for the feature
*/
std::vector<double> value() const
{
@@ -209,13 +207,12 @@ public:
}
/**
- * @brief An array containing the values of the test set samples for the feature
-
+ * @brief A vector containing the values of the test set samples for the feature
*/
std::vector<double> test_value() const
{
- std::vector<double> val(_n_test_samp, 0.0);
- std::copy_n(test_value_ptr(), _n_test_samp, val.data());
+ std::vector<double> val(_n_samp_test, 0.0);
+ std::copy_n(test_value_ptr(), _n_samp_test, val.data());
return val;
}
@@ -241,7 +238,7 @@ public:
* @brief Get the pointer to the feature's training data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
* @return pointer to the feature's training value
@@ -272,7 +269,7 @@ public:
* @brief Get the pointer to the feature's test data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
* @return pointer to the feature's test values
@@ -328,7 +325,7 @@ public:
// DocString: op_node_rung
/**
- * @brief return the rung of the feature (Height of the binary expression tree + 1)
+ * @brief return the rung of the feature (Height of the binary expression tree - 1)
*
* @param cur_rung (int) A recursive helper counter for the rung
*/
@@ -341,6 +338,7 @@ public:
/**
* @brief Get the primary feature decomposition of a feature
+ *
* @return A map representing the primary feature comprising a feature
*/
std::map<int, int> primary_feature_decomp() const
@@ -384,8 +382,7 @@ public:
}
/**
- * @brief Get the string character representation of the node for the postfix expression
- * @return the string representation of the node for the postfix expression
+ * @brief Get the term used in the postfix expression for this Node
*/
virtual std::string get_postfix_term() const = 0;
@@ -420,36 +417,35 @@ public:
}
/**
- * @brief update the dictionary used to check if an Add/Sub node is valid
+ * @brief update the dictionary used to check if an Add/Sub/AbsDiff node is valid
*
* @param add_sub_leaves the dictionary used to check if an Add/Sub node is valid
- * @param pl_mn if for an addition node: 1 if for a subtraction node: -1
+ * @param pl_mn 1 for addition and -1 for subtraction
* @param expected_abs_tot The expected absolute sum of all values in add_sub_leaves
*/
virtual void update_add_sub_leaves(std::map<std::string, int>& add_sub_leaves, const int pl_mn, int& expected_abs_tot) const = 0;
/**
- * @brief update the dictionary used to check if a mult/div node is valid
+ * @brief update the dictionary used to check if an Mult/Div node is valid
*
* @param div_mult_leaves the dictionary used to check if an Mult/Div node is valid
- * @param fact Factor used to update leaves (power of the feature)
+ * @param fact amount to increment the element (a primary features) of the dictionary by
* @param expected_abs_tot The expected absolute sum of all values in div_mult_leaves
- *
*/
virtual void update_div_mult_leaves(std::map<std::string, double>& div_mult_leaves, const double fact, double& expected_abs_tot) const = 0;
#ifdef PARAMETERIZE
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
virtual std::vector<double> parameters() const = 0;
// DocString: op_node_get_params
/**
- * @brief Optimize the non-linear parameters of the feature
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer (NLOptimizer) The optimizer used to get the parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
virtual void get_parameters(std::shared_ptr<NLOptimizer> optimizer) = 0;
@@ -462,7 +458,7 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
virtual void set_parameters(const double* params) = 0;
@@ -471,7 +467,7 @@ public:
* @brief returns the number of theoretical parameters for this feature
*
* @param n_cur the current number of parameters
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
* @return the number of theoretical parameters
*/
virtual inline int n_params(const int n_cur = 0, const int depth = 1) const
@@ -487,21 +483,21 @@ public:
/**
* @brief Set the value of all training samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
virtual void set_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const = 0;
/**
* @brief The pointer to where the feature's training data is stored
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
- * @returns the pointer to the feature's data
+ * @param depth (int) How far down a given Node is from the root OperatorNode
+ * @returns The pointer to the feature's data
*/
double* value_ptr(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const
{
@@ -518,21 +514,21 @@ public:
/**
* @brief Set the value of all test samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
virtual void set_test_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const = 0;
/**
* @brief The pointer to where the feature's test data is stored
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
- * @returns the pointer to the feature's data
+ * @param depth (int) How far down a given Node is from the root OperatorNode
+ * @returns The pointer to the feature's test data
*/
double* test_value_ptr(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const
{
@@ -545,52 +541,51 @@ public:
/**
* @brief A human readable equation representing the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return feature expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return A human readable equation representing the feature
*/
virtual std::string expr(const double* params, const int depth=1) const = 0;
/**
* @brief Get the valid LaTeX expression that represents the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return the latexified expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return Get the valid LaTeX expression that represents the feature
*/
virtual std::string get_latex_expr(const double* params, const int depth=1) const = 0;
/**
* @brief Get the string that corresponds to the code needed to evaluate the node in matlab
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
* @return The matlab code for the feature
*/
virtual std::string matlab_fxn_expr(const double* params, const int depth=1) const = 0;
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Set the upper and lower bounds for the scale and bias term of this Node and its children
*
- * @param lb pointer to the lower bounds data
- * @param ub pointer to the upper bounds data
- * @param from_parent How many parameters are between the start of this node's parameters and its parent
- * @param depth the current depth of the node on the binary expression tree
+ * @param lb A pointer to the location where the lower bounds for the scale and bias term of this Node is set
+ * @param ub A pointer to the location where the upper bounds for the scale and bias term of this Node is set
+ * @param depth How far down a given Node is from the root OperatorNode
*/
virtual void set_bounds(double* lb, double* ub, const int depth=1) const = 0;
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Initialize the scale and bias terms for this Node and its children
*
- * @param params pointer to the parameters vector
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
*/
virtual void initialize_params(double* params, const int depth = 1) const = 0;
/**
* @brief Calculates the derivative of an operation with respect to the parameters for a given sample
*
- * @param params pointer to the parameters
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param dfdp pointer to where the feature derivative pointers are located
*/
virtual void param_derivative(const double* params, double* dfdp) const = 0;
@@ -615,7 +610,7 @@ public:
*
* @param grad pointer to the gradient storage
* @param dfdp pointer to where the feature derivative are located
- * @param params pointer to the parameters vector
+ * @param params A pointer to the bias and scale terms for this Node and its children
*/
void gradient(double* grad, double* dfdp, const double* params) const
{
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs/absolute_value.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs/absolute_value.cpp
index dcee8ea2de7d2a4389646d32ed5a35a470e511ee..d63a154b4f8dca47f46c830660c59a5b7617ffb1 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs/absolute_value.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs/absolute_value.cpp
@@ -159,7 +159,7 @@ void AbsNode::set_test_value(int offset, const bool for_comp) const
{
offset += (offset == -1);
allowed_op_funcs::abs(
- _n_test_samp,
+ _n_samp_test,
_feats[0]->test_value_ptr(2 * offset, for_comp),
1.0,
0.0,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs/absolute_value.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs/absolute_value.hpp
index dbc724269b9d0605630c4822471aaa451c38eb78..26926db8d70a0616f6d526328d520d3441fdd33c 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs/absolute_value.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs/absolute_value.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/absolute_value.hpp
- * @brief Class describing the absolute value operator
+ * @brief Class describing the absolute value operator
*
* This class represents the unary operator -> |A|
*
@@ -52,8 +52,7 @@ public:
AbsNode();
/**
- * @brief Constructor
- * @details Constructs the Node from an array of features to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -62,8 +61,7 @@ public:
// DocString: abs_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -73,8 +71,8 @@ public:
AbsNode(const node_ptr feat, const unsigned long int feat_ind, const double l_bound, const double u_bound);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
virtual node_ptr hard_copy()const;
@@ -128,7 +126,7 @@ public:
// DocString: abs_node_rung
/**
- * @brief return the rung of the feature (Height of the binary expression tree + 1)
+ * @brief return the rung of the feature (Height of the binary expression tree - 1)
*
*/
inline int rung(const int cur_rung=0) const {return _feats[0]->rung(cur_rung + 1);}
@@ -139,8 +137,7 @@ public:
virtual inline NODE_TYPE type() const {return NODE_TYPE::ABS;}
/**
- * @brief Get the string character representation of the node for the postfix expression
- * @return the string representation of the node for the postfix expression
+ * @brief Get the term used in the postfix expression for this Node
*/
inline std::string get_postfix_term() const {return "abs";}
@@ -159,35 +156,34 @@ public:
}
/**
- * @brief update the dictionary used to check if an Add/Sub node is valid
+ * @brief update the dictionary used to check if an Add/Sub/AbsDiff node is valid
*
* @param add_sub_leaves the dictionary used to check if an Add/Sub node is valid
- * @param pl_mn if for an addition node: 1 if for a subtraction node: -1
+ * @param pl_mn 1 for addition and -1 for subtraction
* @param expected_abs_tot The expected absolute sum of all values in add_sub_leaves
*/
void update_add_sub_leaves(std::map<std::string, int>& add_sub_leaves, const int pl_mn, int& expected_abs_tot) const;
/**
- * @brief update the dictionary used to check if a mult/div node is valid
+ * @brief update the dictionary used to check if an Mult/Div node is valid
*
* @param div_mult_leaves the dictionary used to check if an Mult/Div node is valid
- * @param fact Factor used to update leaves (power of the feature)
+ * @param fact amount to increment the element (a primary features) of the dictionary by
* @param expected_abs_tot The expected absolute sum of all values in div_mult_leaves
- *
*/
void update_div_mult_leaves(std::map<std::string, double>& div_mult_leaves, const double fact, double& expected_abs_tot) const;
#ifdef PARAMETERIZE
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
virtual std::vector<double> parameters() const {return {};}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
virtual void get_parameters(std::shared_ptr<NLOptimizer> optimizer){return;}
@@ -200,36 +196,36 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
virtual void set_parameters(const double* params){return;}
/**
* @brief Set the value of all training samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief Set the value of all test samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_test_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief A human readable equation representing the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return feature expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return A human readable equation representing the feature
*/
inline std::string expr(const double* params, const int depth=1) const
{
@@ -244,9 +240,9 @@ public:
/**
* @brief Get the valid LaTeX expression that represents the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return the latexified expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return Get the valid LaTeX expression that represents the feature
*/
inline std::string get_latex_expr(const double* params, const int depth=1) const
{
@@ -261,8 +257,8 @@ public:
/**
* @brief Get the string that corresponds to the code needed to evaluate the node in matlab
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
* @return The matlab code for the feature
*/
inline std::string matlab_fxn_expr(const double* params, const int depth=1) const
@@ -276,27 +272,26 @@ public:
}
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Set the upper and lower bounds for the scale and bias term of this Node and its children
*
- * @param lb pointer to the lower bounds data
- * @param ub pointer to the upper bounds data
- * @param from_parent How many parameters are between the start of this node's parameters and its parent
- * @param depth the current depth of the node on the binary expression tree
+ * @param lb A pointer to the location where the lower bounds for the scale and bias term of this Node is set
+ * @param ub A pointer to the location where the upper bounds for the scale and bias term of this Node is set
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void set_bounds(double* lb, double* ub, const int depth=1) const;
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Initialize the scale and bias terms for this Node and its children
*
- * @param params pointer to the parameters vector
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void initialize_params(double* params, const int depth = 1) const;
/**
* @brief Calculates the derivative of an operation with respect to the parameters for a given sample
*
- * @param params pointer to the parameters
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param dfdp pointer to where the feature derivative pointers are located
*/
inline void param_derivative(const double* params, double* dfdp) const
@@ -312,9 +307,9 @@ public:
*
* @param feat_list list of features already generated
* @param feat feature to attempt to take the absolute value of
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
*/
void generateAbsNode(
std::vector<node_ptr>& feat_list,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs/parameterized_absolute_value.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs/parameterized_absolute_value.cpp
index 107b90ea9273bf2af201dc76cf308cdeb9ec890f..5b4d2a630d08401d97af37a4711b1931234a2f9e 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs/parameterized_absolute_value.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs/parameterized_absolute_value.cpp
@@ -147,7 +147,7 @@ void AbsNode::set_test_value(const double* params, int offset, const bool for_co
}
allowed_op_funcs::abs(
- _n_test_samp, vp_0, params[0], params[1], node_value_arrs::access_param_storage_test(rung(), offset, for_comp)
+ _n_samp_test, vp_0, params[0], params[1], node_value_arrs::access_param_storage_test(rung(), offset, for_comp)
);
}
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs/parameterized_absolute_value.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs/parameterized_absolute_value.hpp
index 2245168598e64997147ddb7bd6f2d84c2eeda897..8ec5026b5908adaec593f52bbb68867b0021d3ce 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs/parameterized_absolute_value.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs/parameterized_absolute_value.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/abs/parameterized_absolute_value.hpp
- * @brief Class describing the parameterized absolute value operator
+ * @brief Class describing the parameterized absolute value operator
*
* This class represents the parameterized unary operator -> |alpha * A + a|
*
@@ -55,7 +55,7 @@ protected:
using AbsNode::get_latex_expr;
using AbsNode::matlab_fxn_expr;
- std::vector<double> _params; //!< The parameters vector
+ std::vector<double> _params; //!< The vector containing the scale and bias terms for the Node
public:
/**
@@ -65,14 +65,13 @@ public:
AbsParamNode();
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
* @param l_bound (double) Minimum absolute value allowed for the feature.
* @param u_bound (double) Maximum absolute value allowed for the feature.
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
AbsParamNode(
const node_ptr feat,
@@ -83,19 +82,17 @@ public:
);
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
AbsParamNode(const node_ptr feat, const unsigned long int feat_ind, std::shared_ptr<NLOptimizer> optimizer);
// DocString: abs_param_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -105,8 +102,8 @@ public:
AbsParamNode(const node_ptr feat, const unsigned long int feat_ind, double const l_bound=1e-50, const double u_bound=1e50);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
node_ptr hard_copy()const;
@@ -132,10 +129,10 @@ public:
* @brief Get the pointer to the feature's training data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
- * @return pointer to the feature's test values
+ * @return A pointer to the feature's training data
*/
inline double* value_ptr(int offset=-1, const bool for_comp=false) const {return value_ptr(_params.data(), offset, for_comp);}
@@ -143,7 +140,7 @@ public:
* @brief Get the pointer to the feature's test data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
* @return pointer to the feature's test values
@@ -176,15 +173,15 @@ public:
inline std::string matlab_fxn_expr() const {return matlab_fxn_expr(_params.data());}
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
inline std::vector<double> parameters() const {return _params;}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
void get_parameters(std::shared_ptr<NLOptimizer> optimizer);
@@ -205,7 +202,7 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
inline void set_parameters(const double* params)
{
@@ -229,9 +226,9 @@ public:
*
* @param feat_list list of features already generated
* @param feat feature to attempt to take the absolute value of
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
* @param optimizer The optimizer used to find the parameters of the node
*/
void generateAbsParamNode(
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs_diff/absolute_difference.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs_diff/absolute_difference.cpp
index 1e573558770e602d95b2e76406d2e507958f16b3..ca73bda552b75146955c5587ff4dd55ad1376f59 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs_diff/absolute_difference.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs_diff/absolute_difference.cpp
@@ -196,7 +196,7 @@ void AbsDiffNode::set_test_value(int offset, const bool for_comp) const
{
offset += (offset == -1);
allowed_op_funcs::abs_diff(
- _n_test_samp,
+ _n_samp_test,
_feats[0]->test_value_ptr(2 * offset, for_comp),
_feats[1]->test_value_ptr(2 * offset + 1, for_comp),
1.0,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs_diff/absolute_difference.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs_diff/absolute_difference.hpp
index 19d45fc9660673686daf22136e38c2fc0479b09b..86533e75c3f903aec02bc8455fdcb1c2f08e015d 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs_diff/absolute_difference.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs_diff/absolute_difference.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/absolute_difference.hpp
- * @brief Class describing the absolute difference operator
+ * @brief Class describing the absolute difference operator
*
* This class represents the binary operator -> |A - B|
*
@@ -53,21 +53,17 @@ public:
AbsDiffNode();
/**
- * @brief Constructor
- * @details Constructs the Node from an array of features to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat_1 (Node) shared_ptr of the first feature to operate on (A)
* @param feat_2 (Node) shared_ptr of the second feature to operate on (B)
* @param feat_ind (int) Index of the new feature
- * @param l_bound (double) Minimum absolute value allowed for the feature.
- * @param u_bound (double) Maximum absolute value allowed for the feature.
*/
AbsDiffNode(const node_ptr feat_1, const node_ptr feat_2, const unsigned long int feat_ind);
// DocString: abs_diff_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the features to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat_1 (Node) shared_ptr of the first feature to operate on (A)
* @param feat_2 (Node) shared_ptr of the second feature to operate on (B)
@@ -78,8 +74,8 @@ public:
AbsDiffNode(const node_ptr feat_1, const node_ptr feat_2, const unsigned long int feat_ind, const double l_bound, const double u_bound);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
virtual node_ptr hard_copy()const;
@@ -136,7 +132,7 @@ public:
// DocString: abs_diff_node_rung
/**
- * @brief return the rung of the feature (Height of the binary expression tree + 1)
+ * @brief return the rung of the feature (Height of the binary expression tree - 1)
*
*/
inline int rung(const int cur_rung=0) const {return std::max(_feats[0]->rung(cur_rung + 1), _feats[1]->rung(cur_rung + 1));}
@@ -147,8 +143,7 @@ public:
virtual inline NODE_TYPE type() const {return NODE_TYPE::ABS_DIFF;}
/**
- * @brief Get the string character representation of the node for the postfix expression
- * @return the string representation of the node for the postfix expression
+ * @brief Get the term used in the postfix expression for this Node
*/
inline std::string get_postfix_term() const {return "abd";}
@@ -174,35 +169,34 @@ public:
void check_feats();
/**
- * @brief update the dictionary used to check if an Add/Sub node is valid
+ * @brief update the dictionary used to check if an Add/Sub/AbsDiff node is valid
*
* @param add_sub_leaves the dictionary used to check if an Add/Sub node is valid
- * @param pl_mn if for an addition node: 1 if for a subtraction node: -1
+ * @param pl_mn 1 for addition and -1 for subtraction
* @param expected_abs_tot The expected absolute sum of all values in add_sub_leaves
*/
void update_add_sub_leaves(std::map<std::string, int>& add_sub_leaves, const int pl_mn, int& expected_abs_tot) const;
/**
- * @brief update the dictionary used to check if a mult/div node is valid
+ * @brief update the dictionary used to check if an Mult/Div node is valid
*
* @param div_mult_leaves the dictionary used to check if an Mult/Div node is valid
- * @param fact Factor used to update leaves (power of the feature)
+ * @param fact amount to increment the element (a primary features) of the dictionary by
* @param expected_abs_tot The expected absolute sum of all values in div_mult_leaves
- *
*/
void update_div_mult_leaves(std::map<std::string, double>& div_mult_leaves, const double fact, double& expected_abs_tot) const;
#ifdef PARAMETERIZE
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
virtual std::vector<double> parameters() const {return {};}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
virtual void get_parameters(std::shared_ptr<NLOptimizer> optimizer){return;}
@@ -215,36 +209,36 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
virtual void set_parameters(const double* params){return;}
/**
* @brief Set the value of all training samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief Set the value of all test samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_test_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief A human readable equation representing the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return feature expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return A human readable equation representing the feature
*/
inline std::string expr(const double* params, const int depth=1) const
{
@@ -260,9 +254,9 @@ public:
/**
* @brief Get the valid LaTeX expression that represents the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return the latexified expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return Get the valid LaTeX expression that represents the feature
*/
inline std::string get_latex_expr(const double* params, const int depth=1) const
{
@@ -278,8 +272,8 @@ public:
/**
* @brief Get the string that corresponds to the code needed to evaluate the node in matlab
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
* @return The matlab code for the feature
*/
inline std::string matlab_fxn_expr(const double* params, const int depth=1) const
@@ -294,27 +288,26 @@ public:
}
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Set the upper and lower bounds for the scale and bias term of this Node and its children
*
- * @param lb pointer to the lower bounds data
- * @param ub pointer to the upper bounds data
- * @param from_parent How many parameters are between the start of this node's parameters and its parent
- * @param depth the current depth of the node on the binary expression tree
+ * @param lb A pointer to the location where the lower bounds for the scale and bias term of this Node is set
+ * @param ub A pointer to the location where the upper bounds for the scale and bias term of this Node is set
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void set_bounds(double* lb, double* ub, const int depth=1) const;
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Initialize the scale and bias terms for this Node and its children
*
- * @param params pointer to the parameters vector
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void initialize_params(double* params, const int depth = 1) const;
/**
* @brief Calculates the derivative of an operation with respect to the parameters for a given sample
*
- * @param params pointer to the parameters
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param dfdp pointer to where the feature derivative pointers are located
*/
inline void param_derivative(const double* params, double* dfdp) const
@@ -338,9 +331,9 @@ public:
* @param feat_list list of features already generated
* @param feat_1 The (A) feature
* @param feat_2 The (B) feature
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
*/
void generateAbsDiffNode(std::vector<node_ptr>& feat_list, const node_ptr feat_1, const node_ptr feat_2, unsigned long int& feat_ind, const double l_bound, const double u_bound);
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs_diff/parameterized_absolute_difference.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs_diff/parameterized_absolute_difference.cpp
index 30c0ed0f5fec1c0fb26cc8816e643ce2327b86ed..8db24edb020a2fa530683e70a91cb01de48013d7 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs_diff/parameterized_absolute_difference.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs_diff/parameterized_absolute_difference.cpp
@@ -210,7 +210,7 @@ void AbsDiffNode::set_test_value(const double* params, int offset, const bool fo
}
allowed_op_funcs::abs_diff(
- _n_test_samp, vp_0, vp_1, params[0], params[1], node_value_arrs::access_param_storage_test(rung(), offset, for_comp)
+ _n_samp_test, vp_0, vp_1, params[0], params[1], node_value_arrs::access_param_storage_test(rung(), offset, for_comp)
);
}
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs_diff/parameterized_absolute_difference.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs_diff/parameterized_absolute_difference.hpp
index fc8f89c396b9bc10c24c104aaca69ea93a9786e1..40a591792fa37b1a73c0964bc7acbe8795b9153d 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs_diff/parameterized_absolute_difference.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/abs_diff/parameterized_absolute_difference.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/abs_diff/parameterized_absolute_difference.hpp
- * @brief Class describing the parameterized absolute difference operator
+ * @brief Class describing the parameterized absolute difference operator
*
* This class represents the parameterized unary operator -> |A - alpha * B + a|
*
@@ -63,15 +63,14 @@ public:
AbsDiffParamNode();
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
- * @param feat_1 shared_ptr of the feature to operate on (A)
- * @param feat_2 shared_ptr of the feature to operate on (B)
+ * @param feat_1 (Node) shared_ptr of the feature to operate on (A)
+ * @param feat_2 (Node) shared_ptr of the feature to operate on (B)
* @param feat_ind (int) Index of the new feature
* @param l_bound (double) Minimum absolute value allowed for the feature.
* @param u_bound (double) Maximum absolute value allowed for the feature.
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
AbsDiffParamNode(
const node_ptr feat_1,
@@ -83,23 +82,21 @@ public:
);
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
- * @param feat_1 shared_ptr of the feature to operate on (A)
- * @param feat_2 shared_ptr of the feature to operate on (B)
+ * @param feat_1 (Node) shared_ptr of the feature to operate on (A)
+ * @param feat_2 (Node) shared_ptr of the feature to operate on (B)
* @param feat_ind (int) Index of the new feature
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
AbsDiffParamNode(const node_ptr feat_1, const node_ptr feat_2, const unsigned long int feat_ind, std::shared_ptr<NLOptimizer> optimizer);
// DocString: abs_diff_param_node_init
/**
* @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
*
- * @param feat_1 shared_ptr of the feature to operate on (A)
- * @param feat_2 shared_ptr of the feature to operate on (B)
+ * @param feat_1 (Node) shared_ptr of the feature to operate on (A)
+ * @param feat_2 (Node) shared_ptr of the feature to operate on (B)
* @param feat_ind (int) Index of the new feature
* @param l_bound (double) Minimum absolute value allowed for the feature.
* @param u_bound (double) Maximum absolute value allowed for the feature.
@@ -107,8 +104,8 @@ public:
AbsDiffParamNode(const node_ptr feat_1, const node_ptr feat_2, const unsigned long int feat_ind, double const l_bound=1e-50, const double u_bound=1e50);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
node_ptr hard_copy()const;
@@ -134,10 +131,10 @@ public:
* @brief Get the pointer to the feature's training data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
- * @return pointer to the feature's test values
+ * @return A pointer to the feature's training data
*/
inline double* value_ptr(int offset=-1, const bool for_comp=false) const {return value_ptr(_params.data(), offset, for_comp);}
@@ -145,7 +142,7 @@ public:
* @brief Get the pointer to the feature's test data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
* @return pointer to the feature's test values
@@ -178,15 +175,15 @@ public:
inline std::string matlab_fxn_expr() const {return matlab_fxn_expr(_params.data());}
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
inline std::vector<double> parameters() const {return _params;}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
void get_parameters(std::shared_ptr<NLOptimizer> optimizer);
@@ -207,7 +204,7 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
inline void set_parameters(const double* params)
{
@@ -232,9 +229,9 @@ public:
* @param feat_list list of features already generated
* @param feat_1 The (A) feature
* @param feat_2 The (B) feature
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
* @param optimizer The optimizer used to find the parameters of the node
*/
void generateAbsDiffParamNode(
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/add/add.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/add/add.cpp
index 053785876373729cbe65af94d6a8b8304f04a08b..48c639337fb35534f8b5ac5cd39a011bc059c601 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/add/add.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/add/add.cpp
@@ -189,7 +189,7 @@ void AddNode::set_test_value(int offset, const bool for_comp) const
{
offset += (offset == -1);
allowed_op_funcs::add(
- _n_test_samp,
+ _n_samp_test,
_feats[0]->test_value_ptr(2 * offset, for_comp),
_feats[1]->test_value_ptr(2 * offset + 1, for_comp),
1.0,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/add/add.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/add/add.hpp
index b0794f3145474c41bbea0a0501b361f003a0b375..5c521ecdbbc780542ebf654cacdc36802eede40d 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/add/add.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/add/add.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/add.hpp
- * @brief Class describing the addition operator
+ * @brief Class describing the addition operator
*
* This class represents the binary operator -> A + B
*
@@ -52,8 +52,7 @@ public:
AddNode();
/**
- * @brief Constructor
- * @details Constructs the Node from an array of features to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat_1 (Node) shared_ptr of the first feature to operate on (A)
* @param feat_2 (Node) shared_ptr of the second feature to operate on (B)
@@ -63,8 +62,7 @@ public:
// DocString: add_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the features to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat_1 (Node) shared_ptr of the first feature to operate on (A)
* @param feat_2 (Node) shared_ptr of the second feature to operate on (B)
@@ -75,8 +73,8 @@ public:
AddNode(const node_ptr feat_1, const node_ptr feat_2, const unsigned long int feat_ind, const double l_bound, const double u_bound);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
virtual node_ptr hard_copy()const;
@@ -118,7 +116,7 @@ public:
// DocString: add_node_rung
/**
- * @brief return the rung of the feature (Height of the binary expression tree + 1)
+ * @brief return the rung of the feature (Height of the binary expression tree - 1)
*
*/
inline int rung(const int cur_rung=0) const {return std::max(_feats[0]->rung(cur_rung + 1), _feats[1]->rung(cur_rung + 1));}
@@ -142,8 +140,7 @@ public:
virtual inline NODE_TYPE type() const {return NODE_TYPE::ADD;}
/**
- * @brief Get the string character representation of the node for the postfix expression
- * @return the string representation of the node for the postfix expression
+ * @brief Get the term used in the postfix expression for this Node
*/
inline std::string get_postfix_term() const {return "add";}
@@ -168,35 +165,34 @@ public:
void check_feats();
/**
- * @brief update the dictionary used to check if an Add/Sub node is valid
+ * @brief update the dictionary used to check if an Add/Sub/AbsDiff node is valid
*
* @param add_sub_leaves the dictionary used to check if an Add/Sub node is valid
- * @param pl_mn if for an addition node: 1 if for a subtraction node: -1
+ * @param pl_mn 1 for addition and -1 for subtraction
* @param expected_abs_tot The expected absolute sum of all values in add_sub_leaves
*/
void update_add_sub_leaves(std::map<std::string, int>& add_sub_leaves, const int pl_mn, int& expected_abs_tot) const;
/**
- * @brief update the dictionary used to check if a mult/div node is valid
+ * @brief update the dictionary used to check if an Mult/Div node is valid
*
* @param div_mult_leaves the dictionary used to check if an Mult/Div node is valid
- * @param fact Factor used to update leaves (power of the feature)
+ * @param fact amount to increment the element (a primary features) of the dictionary by
* @param expected_abs_tot The expected absolute sum of all values in div_mult_leaves
- *
*/
void update_div_mult_leaves(std::map<std::string, double>& div_mult_leaves, const double fact, double& expected_abs_tot) const;
#ifdef PARAMETERIZE
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
virtual std::vector<double> parameters() const {return {};}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
virtual void get_parameters(std::shared_ptr<NLOptimizer> optimizer){return;}
@@ -209,36 +205,36 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
virtual void set_parameters(const double* params){return;}
/**
* @brief Set the value of all training samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief Set the value of all test samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_test_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief A human readable equation representing the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return feature expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return A human readable equation representing the feature
*/
inline std::string expr(const double* params, const int depth=1) const
{
@@ -253,9 +249,9 @@ public:
/**
* @brief Get the valid LaTeX expression that represents the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return the latexified expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return Get the valid LaTeX expression that represents the feature
*/
inline std::string get_latex_expr(const double* params, const int depth=1) const
{
@@ -270,8 +266,8 @@ public:
/**
* @brief Get the string that corresponds to the code needed to evaluate the node in matlab
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
* @return The matlab code for the feature
*/
inline std::string matlab_fxn_expr(const double* params, const int depth=1) const
@@ -285,27 +281,26 @@ public:
}
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Set the upper and lower bounds for the scale and bias term of this Node and its children
*
- * @param lb pointer to the lower bounds data
- * @param ub pointer to the upper bounds data
- * @param from_parent How many parameters are between the start of this node's parameters and its parent
- * @param depth the current depth of the node on the binary expression tree
+ * @param lb A pointer to the location where the lower bounds for the scale and bias term of this Node is set
+ * @param ub A pointer to the location where the upper bounds for the scale and bias term of this Node is set
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void set_bounds(double* lb, double* ub, const int depth=1) const;
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Initialize the scale and bias terms for this Node and its children
*
- * @param params pointer to the parameters vector
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void initialize_params(double* params, const int depth = 1) const;
/**
* @brief Calculates the derivative of an operation with respect to the parameters for a given sample
*
- * @param params pointer to the parameters
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param dfdp pointer to where the feature derivative pointers are located
*/
inline void param_derivative(const double* params, double* dfdp) const {std::fill_n(dfdp, _n_samp, 1.0);}
@@ -318,9 +313,9 @@ public:
* @param feat_list list of features already generated
* @param feat_1 The (A) feature
* @param feat_2 The (B) feature
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
*/
void generateAddNode(std::vector<node_ptr>& feat_list, const node_ptr feat_1, const node_ptr feat_2, unsigned long int& feat_ind, const double l_bound, const double u_bound);
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/add/parameterized_add.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/add/parameterized_add.cpp
index 5aae7814ac9d0320b310e36039319d796fd885e2..2e6c7763de2441ffbc5c7d40a92feae51413a6e2 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/add/parameterized_add.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/add/parameterized_add.cpp
@@ -173,7 +173,7 @@ void AddNode::set_test_value(const double* params, int offset, const bool for_co
}
allowed_op_funcs::add(
- _n_test_samp,
+ _n_samp_test,
vp_0,
vp_1,
params[0],
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/add/parameterized_add.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/add/parameterized_add.hpp
index 923867899096c5fde6a95f6aed1e925a5083612d..27816574bdcdd9a213e5f76736627f815fc90744 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/add/parameterized_add.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/add/parameterized_add.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/add/parameterized_add.hpp
- * @brief Class describing the parameterized addition operator
+ * @brief Class describing the parameterized addition operator
*
* This class represents the parameterized unary operator -> A + alpha * B + a
*
@@ -65,15 +65,14 @@ public:
AddParamNode();
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
- * @param feat_1 shared_ptr of the feature to operate on (A)
- * @param feat_2 shared_ptr of the feature to operate on (B)
+ * @param feat_1 (Node) shared_ptr of the feature to operate on (A)
+ * @param feat_2 (Node) shared_ptr of the feature to operate on (B)
* @param feat_ind (int) Index of the new feature
* @param l_bound (double) Minimum absolute value allowed for the feature.
* @param u_bound (double) Maximum absolute value allowed for the feature.
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
AddParamNode(
const node_ptr feat_1,
@@ -85,20 +84,18 @@ public:
);
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
- * @param feat_1 shared_ptr of the feature to operate on (A)
- * @param feat_2 shared_ptr of the feature to operate on (B)
+ * @param feat_1 (Node) shared_ptr of the feature to operate on (A)
+ * @param feat_2 (Node) shared_ptr of the feature to operate on (B)
* @param feat_ind (int) Index of the new feature
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
AddParamNode(const node_ptr feat_1, const node_ptr feat_2, const unsigned long int feat_ind, std::shared_ptr<NLOptimizer> optimizer);
// DocString: add_param_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -108,8 +105,8 @@ public:
AddParamNode(const node_ptr feat_1, const node_ptr feat_2, const unsigned long int feat_ind, double const l_bound=1e-50, const double u_bound=1e50);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
node_ptr hard_copy()const;
@@ -135,10 +132,10 @@ public:
* @brief Get the pointer to the feature's training data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
- * @return pointer to the feature's test values
+ * @return A pointer to the feature's training data
*/
inline double* value_ptr(int offset=-1, const bool for_comp=false) const {return value_ptr(_params.data(), offset, for_comp);}
@@ -146,7 +143,7 @@ public:
* @brief Get the pointer to the feature's test data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
* @return pointer to the feature's test values
@@ -179,15 +176,15 @@ public:
inline std::string matlab_fxn_expr() const {return matlab_fxn_expr(_params.data());}
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
inline std::vector<double> parameters() const {return _params;}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
void get_parameters(std::shared_ptr<NLOptimizer> optimizer);
@@ -208,7 +205,7 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
inline void set_parameters(const double* params)
{
@@ -233,9 +230,9 @@ public:
* @param feat_list list of features already generated
* @param feat_1 The (A) feature
* @param feat_2 The (B) feature
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
* @param optimizer The optimizer used to find the parameters of the node
*/
void generateAddParamNode(
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cb/cube.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cb/cube.cpp
index 0a80999e40a8e4b553a054b5c1d0dc294f45afb4..bcd1b8299e846391edf12578ae49f8f26987e401 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cb/cube.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cb/cube.cpp
@@ -143,7 +143,7 @@ void CbNode::set_test_value(int offset, const bool for_comp) const
{
offset += (offset == -1);
allowed_op_funcs::cb(
- _n_test_samp,
+ _n_samp_test,
_feats[0]->test_value_ptr(2 * offset, for_comp),
1.0,
0.0,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cb/cube.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cb/cube.hpp
index 0507c0f1b685c03a8be1b0519b77d2e7e521c23a..1dc050d509c8c87616c4e41ba822dbb282b531c4 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cb/cube.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cb/cube.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/cube.hpp
- * @brief Class describing the cube operator
+ * @brief Class describing the cube operator
*
* This class represents the unary operator -> (A)^3
*
@@ -52,8 +52,7 @@ public:
CbNode();
/**
- * @brief Constructor
- * @details Constructs the Node from an array of features to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -62,8 +61,7 @@ public:
// DocString: cb_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -73,8 +71,8 @@ public:
CbNode(const node_ptr feat, const unsigned long int feat_ind, const double l_bound, const double u_bound);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
virtual node_ptr hard_copy()const;
@@ -128,7 +126,7 @@ public:
// DocString: cb_node_rung
/**
- * @brief return the rung of the feature (Height of the binary expression tree + 1)
+ * @brief return the rung of the feature (Height of the binary expression tree - 1)
*
*/
inline int rung(const int cur_rung=0) const {return _feats[0]->rung(cur_rung + 1);}
@@ -139,8 +137,7 @@ public:
virtual inline NODE_TYPE type() const {return NODE_TYPE::CB;}
/**
- * @brief Get the string character representation of the node for the postfix expression
- * @return the string representation of the node for the postfix expression
+ * @brief Get the term used in the postfix expression for this Node
*/
inline std::string get_postfix_term() const {return "cb";}
@@ -159,35 +156,34 @@ public:
}
/**
- * @brief update the dictionary used to check if an Add/Sub node is valid
+ * @brief update the dictionary used to check if an Add/Sub/AbsDiff node is valid
*
* @param add_sub_leaves the dictionary used to check if an Add/Sub node is valid
- * @param pl_mn if for an addition node: 1 if for a subtraction node: -1
+ * @param pl_mn 1 for addition and -1 for subtraction
* @param expected_abs_tot The expected absolute sum of all values in add_sub_leaves
*/
void update_add_sub_leaves(std::map<std::string, int>& add_sub_leaves, const int pl_mn, int& expected_abs_tot) const;
/**
- * @brief update the dictionary used to check if a mult/div node is valid
+ * @brief update the dictionary used to check if an Mult/Div node is valid
*
* @param div_mult_leaves the dictionary used to check if an Mult/Div node is valid
- * @param fact Factor used to update leaves (power of the feature)
+ * @param fact amount to increment the element (a primary features) of the dictionary by
* @param expected_abs_tot The expected absolute sum of all values in div_mult_leaves
- *
*/
void update_div_mult_leaves(std::map<std::string, double>& div_mult_leaves, const double fact, double& expected_abs_tot) const;
#ifdef PARAMETERIZE
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
virtual std::vector<double> parameters() const {return {};}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
virtual void get_parameters(std::shared_ptr<NLOptimizer> optimizer){return;}
@@ -200,36 +196,36 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
virtual void set_parameters(const double* params){return;}
/**
* @brief Set the value of all training samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief Set the value of all test samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_test_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief A human readable equation representing the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return feature expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return A human readable equation representing the feature
*/
inline std::string expr(const double* params, const int depth=1) const
{
@@ -243,9 +239,9 @@ public:
/**
* @brief Get the valid LaTeX expression that represents the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return the latexified expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return Get the valid LaTeX expression that represents the feature
*/
inline std::string get_latex_expr(const double* params, const int depth=1) const
{
@@ -259,8 +255,8 @@ public:
/**
* @brief Get the string that corresponds to the code needed to evaluate the node in matlab
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
* @return The matlab code for the feature
*/
inline std::string matlab_fxn_expr(const double* params, const int depth=1) const
@@ -273,27 +269,26 @@ public:
}
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Set the upper and lower bounds for the scale and bias term of this Node and its children
*
- * @param lb pointer to the lower bounds data
- * @param ub pointer to the upper bounds data
- * @param from_parent How many parameters are between the start of this node's parameters and its parent
- * @param depth the current depth of the node on the binary expression tree
+ * @param lb A pointer to the location where the lower bounds for the scale and bias term of this Node is set
+ * @param ub A pointer to the location where the upper bounds for the scale and bias term of this Node is set
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void set_bounds(double* lb, double* ub, const int depth=1) const;
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Initialize the scale and bias terms for this Node and its children
*
- * @param params pointer to the parameters vector
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void initialize_params(double* params, const int depth = 1) const;
/**
* @brief Calculates the derivative of an operation with respect to the parameters for a given sample
*
- * @param params pointer to the parameters
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param dfdp pointer to where the feature derivative pointers are located
*/
inline void param_derivative(const double* params, double* dfdp) const
@@ -309,9 +304,9 @@ public:
*
* @param feat_list list of features already generated
* @param feat feature to attempt to take the cube of
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
*/
void generateCbNode(
std::vector<node_ptr>& feat_list,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cb/parameterized_cube.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cb/parameterized_cube.cpp
index 2760a06cad72c5671aae808247b6701618b4d113..fd465f0a5126293dc91cfdfa19d99e058a244565 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cb/parameterized_cube.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cb/parameterized_cube.cpp
@@ -187,7 +187,7 @@ void CbNode::set_test_value(const double* params, int offset, const bool for_com
}
allowed_op_funcs::cb(
- _n_test_samp,
+ _n_samp_test,
vp_0,
params[0],
params[1],
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cb/parameterized_cube.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cb/parameterized_cube.hpp
index 1bf2287db58e22f039515730c04f048c8cde0214..ea095e25a6c4930bdd2dc625fe652f23db9b2de6 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cb/parameterized_cube.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cb/parameterized_cube.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/cb/parameterized_cube.hpp
- * @brief Class describing the parameterized cube operator
+ * @brief Class describing the parameterized cube operator
*
* This class represents the parameterized unary operator -> (alpha * A + a)^3
*
@@ -55,7 +55,7 @@ protected:
using CbNode::get_latex_expr;
using CbNode::matlab_fxn_expr;
- std::vector<double> _params; //!< The parameters vector
+ std::vector<double> _params; //!< The vector containing the scale and bias terms for the Node
public:
/**
@@ -65,14 +65,13 @@ public:
CbParamNode();
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
* @param l_bound (double) Minimum absolute value allowed for the feature.
* @param u_bound (double) Maximum absolute value allowed for the feature.
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
CbParamNode(
const node_ptr feat,
@@ -83,19 +82,17 @@ public:
);
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
CbParamNode(const node_ptr feat, const unsigned long int feat_ind, std::shared_ptr<NLOptimizer> optimizer);
// DocString: cb_param_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -105,8 +102,8 @@ public:
CbParamNode(const node_ptr feat, const unsigned long int feat_ind, double const l_bound=1e-50, const double u_bound=1e50);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
node_ptr hard_copy()const;
@@ -132,10 +129,10 @@ public:
* @brief Get the pointer to the feature's training data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
- * @return pointer to the feature's test values
+ * @return A pointer to the feature's training data
*/
inline double* value_ptr(int offset=-1, const bool for_comp=false) const {return value_ptr(_params.data(), offset, for_comp);}
@@ -143,7 +140,7 @@ public:
* @brief Get the pointer to the feature's test data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
* @return pointer to the feature's test values
@@ -176,15 +173,15 @@ public:
inline std::string matlab_fxn_expr() const {return matlab_fxn_expr(_params.data());}
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
inline std::vector<double> parameters() const {return _params;}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
void get_parameters(std::shared_ptr<NLOptimizer> optimizer);
@@ -205,7 +202,7 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
inline void set_parameters(const double* params)
{
@@ -229,9 +226,9 @@ public:
*
* @param feat_list list of features already generated
* @param feat feature to attempt to take the cube of
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
* @param optimizer The optimizer used to find the parameters of the node
*/
void generateCbParamNode(
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cbrt/cube_root.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cbrt/cube_root.cpp
index 241b68082a4721b32803d1d890eb3149b412f90d..7d31aec2339734ebb72fc5919268ecb968323af3 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cbrt/cube_root.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cbrt/cube_root.cpp
@@ -149,7 +149,7 @@ void CbrtNode::set_test_value(int offset, const bool for_comp) const
{
offset += (offset == -1);
allowed_op_funcs::cbrt(
- _n_test_samp,
+ _n_samp_test,
_feats[0]->test_value_ptr(2 * offset, for_comp),
1.0,
0.0,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cbrt/cube_root.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cbrt/cube_root.hpp
index 55f5a107314a835115da890dd3a3a26c6e525f3a..de6822569b0b763498538a17b7c58f548b607c5c 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cbrt/cube_root.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cbrt/cube_root.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/cube_root.hpp
- * @brief Class describing the cube root operator
+ * @brief Class describing the cube root operator
*
* This class represents the unary operator -> (A)^(1/3)
*
@@ -52,8 +52,7 @@ public:
CbrtNode();
/**
- * @brief Constructor
- * @details Constructs the Node from an array of features to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -62,8 +61,7 @@ public:
// DocString: cbrt_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -73,8 +71,8 @@ public:
CbrtNode(const node_ptr feat, const unsigned long int feat_ind, const double l_bound, const double u_bound);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
virtual node_ptr hard_copy()const;
@@ -128,7 +126,7 @@ public:
// DocString: cbrt_node_rung
/**
- * @brief return the rung of the feature (Height of the binary expression tree + 1)
+ * @brief return the rung of the feature (Height of the binary expression tree - 1)
*
*/
inline int rung(const int cur_rung=0) const {return _feats[0]->rung(cur_rung + 1);}
@@ -139,8 +137,7 @@ public:
virtual inline NODE_TYPE type() const {return NODE_TYPE::CBRT;}
/**
- * @brief Get the string character representation of the node for the postfix expression
- * @return the string representation of the node for the postfix expression
+ * @brief Get the term used in the postfix expression for this Node
*/
inline std::string get_postfix_term() const {return "cbrt";}
@@ -168,26 +165,25 @@ public:
void update_add_sub_leaves(std::map<std::string, int>& add_sub_leaves, const int pl_mn, int& expected_abs_tot) const;
/**
- * @brief update the dictionary used to check if a mult/div node is valid
+ * @brief update the dictionary used to check if an Mult/Div node is valid
*
* @param div_mult_leaves the dictionary used to check if an Mult/Div node is valid
- * @param fact Factor used to update leaves (power of the feature)
+ * @param fact amount to increment the element (a primary features) of the dictionary by
* @param expected_abs_tot The expected absolute sum of all values in div_mult_leaves
- *
*/
void update_div_mult_leaves(std::map<std::string, double>& div_mult_leaves, const double fact, double& expected_abs_tot) const;
#ifdef PARAMETERIZE
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
virtual std::vector<double> parameters() const {return {};}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
virtual void get_parameters(std::shared_ptr<NLOptimizer> optimizer){return;}
@@ -200,35 +196,35 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
virtual void set_parameters(const double* params){return;}
/**
* @brief Set the value of all training samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief Set the value of all test samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_test_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief A human readable equation representing the feature
*
- * @param params parameter values for non-linear operations
- * @return feature expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @return A human readable equation representing the feature
*/
inline std::string expr(const double* params, const int depth=1) const
{
@@ -242,9 +238,9 @@ public:
/**
* @brief Get the valid LaTeX expression that represents the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return the latexified expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return Get the valid LaTeX expression that represents the feature
*/
inline std::string get_latex_expr(const double* params, const int depth=1) const
{
@@ -258,8 +254,8 @@ public:
/**
* @brief Get the string that corresponds to the code needed to evaluate the node in matlab
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
* @return The matlab code for the feature
*/
inline std::string matlab_fxn_expr(const double* params, const int depth=1) const
@@ -272,27 +268,26 @@ public:
}
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Set the upper and lower bounds for the scale and bias term of this Node and its children
*
- * @param lb pointer to the lower bounds data
- * @param ub pointer to the upper bounds data
- * @param from_parent How many parameters are between the start of this node's parameters and its parent
- * @param depth the current depth of the node on the binary expression tree
+ * @param lb A pointer to the location where the lower bounds for the scale and bias term of this Node is set
+ * @param ub A pointer to the location where the upper bounds for the scale and bias term of this Node is set
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void set_bounds(double* lb, double* ub, const int depth=1) const;
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Initialize the scale and bias terms for this Node and its children
*
- * @param params pointer to the parameters vector
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void initialize_params(double* params, const int depth = 1) const;
/**
* @brief Calculates the derivative of an operation with respect to the parameters for a given sample
*
- * @param params pointer to the parameters
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param dfdp pointer to where the feature derivative pointers are located
*/
inline void param_derivative(const double* params, double* dfdp) const
@@ -308,9 +303,9 @@ public:
*
* @param feat_list list of features already generated
* @param feat feature to attempt to take the absolute value of
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
*/
void generateCbrtNode(
std::vector<node_ptr>& feat_list,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cbrt/parameterized_cube_root.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cbrt/parameterized_cube_root.cpp
index 753137f33f27850fd0c574134a0587e965a2a4cc..2841784d22dee0d6632bb91b25e43d5fb4e1f5d4 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cbrt/parameterized_cube_root.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cbrt/parameterized_cube_root.cpp
@@ -173,7 +173,7 @@ void CbrtNode::set_test_value(const double* params, int offset, const bool for_c
}
allowed_op_funcs::cbrt(
- _n_test_samp,
+ _n_samp_test,
vp_0,
params[0],
params[1],
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cbrt/parameterized_cube_root.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cbrt/parameterized_cube_root.hpp
index 23a0872d25f144bdf2978a7a83c4e9485d3d8bca..e1dfa283615a60cecc49671367231f58a8d315e3 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cbrt/parameterized_cube_root.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cbrt/parameterized_cube_root.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/cbrt/parameterized_cube_root.hpp
- * @brief Class describing the parameterized cube root operator
+ * @brief Class describing the parameterized cube root operator
*
* This class represents the parameterized unary operator -> cbrt(alpha * A + a)
*
@@ -55,7 +55,7 @@ protected:
using CbrtNode::get_latex_expr;
using CbrtNode::matlab_fxn_expr;
- std::vector<double> _params; //!< The parameters vector
+ std::vector<double> _params; //!< The vector containing the scale and bias terms for the Node
public:
/**
* @brief Base Constructor
@@ -64,14 +64,13 @@ public:
CbrtParamNode();
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
* @param l_bound (double) Minimum absolute value allowed for the feature.
* @param u_bound (double) Maximum absolute value allowed for the feature.
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
CbrtParamNode(
const node_ptr feat,
@@ -82,19 +81,17 @@ public:
);
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
CbrtParamNode(const node_ptr feat, const unsigned long int feat_ind, std::shared_ptr<NLOptimizer> optimizer);
// DocString: cbrt_param_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -104,8 +101,8 @@ public:
CbrtParamNode(const node_ptr feat, const unsigned long int feat_ind, double const l_bound=1e-50, const double u_bound=1e50);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
node_ptr hard_copy()const;
@@ -131,10 +128,10 @@ public:
* @brief Get the pointer to the feature's training data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
- * @return pointer to the feature's test values
+ * @return A pointer to the feature's training data
*/
inline double* value_ptr(int offset=-1, const bool for_comp=false) const {return value_ptr(_params.data(), offset, for_comp);}
@@ -142,7 +139,7 @@ public:
* @brief Get the pointer to the feature's test data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
* @return pointer to the feature's test values
@@ -175,15 +172,15 @@ public:
inline std::string matlab_fxn_expr() const {return matlab_fxn_expr(_params.data());}
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
inline std::vector<double> parameters() const {return _params;}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
void get_parameters(std::shared_ptr<NLOptimizer> optimizer);
@@ -204,7 +201,7 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
inline void set_parameters(const double* params)
{
@@ -228,9 +225,9 @@ public:
*
* @param feat_list list of features already generated
* @param feat feature to attempt to take the absolute value of
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
* @param optimizer The optimizer used to find the parameters of the node
*/
void generateCbrtParamNode(
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cos/cos.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cos/cos.cpp
index 3e6357b49fa33b7c500c0a3059005e9273713c62..81d1dff2b5506937312581ce72d4b42b23908e78 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cos/cos.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cos/cos.cpp
@@ -144,7 +144,7 @@ void CosNode::set_test_value(int offset, const bool for_comp) const
{
offset += (offset == -1);
allowed_op_funcs::cos(
- _n_test_samp,
+ _n_samp_test,
_feats[0]->test_value_ptr(2 * offset, for_comp),
1.0,
0.0,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cos/cos.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cos/cos.hpp
index c21ccd18337e9211f56e0fdcb6974dad2bd5e7bc..63a18150a00085a54ad3880c2fe919f07a0c5bc4 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cos/cos.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cos/cos.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/cos.hpp
- * @brief Class describing the cosine operator
+ * @brief Class describing the cosine operator
*
* This class represents the unary operator -> cos(A)
*
@@ -52,8 +52,7 @@ public:
CosNode();
/**
- * @brief Constructor
- * @details Constructs the Node from an array of features to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -62,8 +61,7 @@ public:
// DocString: cos_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -73,8 +71,8 @@ public:
CosNode(const node_ptr feat, const unsigned long int feat_ind, const double l_bound, const double u_bound);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
virtual node_ptr hard_copy()const;
@@ -128,7 +126,7 @@ public:
// DocString: cos_node_rung
/**
- * @brief return the rung of the feature (Height of the binary expression tree + 1)
+ * @brief return the rung of the feature (Height of the binary expression tree - 1)
*
*/
inline int rung(const int cur_rung=0) const {return _feats[0]->rung(cur_rung + 1);}
@@ -139,8 +137,7 @@ public:
virtual inline NODE_TYPE type() const {return NODE_TYPE::COS;}
/**
- * @brief Get the string character representation of the node for the postfix expression
- * @return the string representation of the node for the postfix expression
+ * @brief Get the term used in the postfix expression for this Node
*/
inline std::string get_postfix_term() const {return "cos";}
@@ -159,35 +156,34 @@ public:
}
/**
- * @brief update the dictionary used to check if an Add/Sub node is valid
+ * @brief update the dictionary used to check if an Add/Sub/AbsDiff node is valid
*
* @param add_sub_leaves the dictionary used to check if an Add/Sub node is valid
- * @param pl_mn if for an addition node: 1 if for a subtraction node: -1
+ * @param pl_mn 1 for addition and -1 for subtraction
* @param expected_abs_tot The expected absolute sum of all values in add_sub_leaves
*/
void update_add_sub_leaves(std::map<std::string, int>& add_sub_leaves, const int pl_mn, int& expected_abs_tot) const;
/**
- * @brief update the dictionary used to check if a mult/div node is valid
+ * @brief update the dictionary used to check if an Mult/Div node is valid
*
* @param div_mult_leaves the dictionary used to check if an Mult/Div node is valid
- * @param fact Factor used to update leaves (power of the feature)
+ * @param fact amount to increment the element (a primary features) of the dictionary by
* @param expected_abs_tot The expected absolute sum of all values in div_mult_leaves
- *
*/
void update_div_mult_leaves(std::map<std::string, double>& div_mult_leaves, const double fact, double& expected_abs_tot) const;
#ifdef PARAMETERIZE
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
virtual std::vector<double> parameters() const {return {};}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
virtual void get_parameters(std::shared_ptr<NLOptimizer> optimizer){return;}
@@ -200,35 +196,35 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
virtual void set_parameters(const double* params){return;}
/**
* @brief Set the value of all training samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief Set the value of all test samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_test_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief A human readable equation representing the feature
*
- * @param params parameter values for non-linear operations
- * @return feature expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @return A human readable equation representing the feature
*/
inline std::string expr(const double* params, const int depth=1) const
{
@@ -243,9 +239,9 @@ public:
/**
* @brief Get the valid LaTeX expression that represents the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return the latexified expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return Get the valid LaTeX expression that represents the feature
*/
inline std::string get_latex_expr(const double* params, const int depth=1) const
{
@@ -260,8 +256,8 @@ public:
/**
* @brief Get the string that corresponds to the code needed to evaluate the node in matlab
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
* @return The matlab code for the feature
*/
inline std::string matlab_fxn_expr(const double* params, const int depth=1) const
@@ -275,27 +271,26 @@ public:
}
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Set the upper and lower bounds for the scale and bias term of this Node and its children
*
- * @param lb pointer to the lower bounds data
- * @param ub pointer to the upper bounds data
- * @param from_parent How many parameters are between the start of this node's parameters and its parent
- * @param depth the current depth of the node on the binary expression tree
+ * @param lb A pointer to the location where the lower bounds for the scale and bias term of this Node is set
+ * @param ub A pointer to the location where the upper bounds for the scale and bias term of this Node is set
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void set_bounds(double* lb, double* ub, const int depth=1) const;
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Initialize the scale and bias terms for this Node and its children
*
- * @param params pointer to the parameters vector
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void initialize_params(double* params, const int depth = 1) const;
/**
* @brief Calculates the derivative of an operation with respect to the parameters for a given sample
*
- * @param params pointer to the parameters
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param dfdp pointer to where the feature derivative pointers are located
*/
inline void param_derivative(const double* params, double* dfdp) const
@@ -310,10 +305,10 @@ public:
* @brief Attempt to generate a new parameterized cosine node cos(A) and add it to feat_list
*
* @param feat_list list of features already generated
- * @param feat feature to attempt to take the cosine of
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat shared_ptr of the feature to operate on (A)
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
*/
void generateCosNode(
std::vector<node_ptr>& feat_list,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cos/parameterized_cos.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cos/parameterized_cos.cpp
index 1f3845bb1b129e43cb65495f056439cbbb0865d8..b7a51e4cc1a27cff1618003685fe0094e1647c28 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cos/parameterized_cos.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cos/parameterized_cos.cpp
@@ -173,7 +173,7 @@ void CosNode::set_test_value(const double* params, int offset, const bool for_co
}
allowed_op_funcs::cos(
- _n_test_samp,
+ _n_samp_test,
vp_0,
params[0],
params[1],
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cos/parameterized_cos.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cos/parameterized_cos.hpp
index 94ef25f352987a4307d4dd460e708f4574c2bbd6..aa5fe2b020bdc0d56f886281a0310329f9f21ee1 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cos/parameterized_cos.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/cos/parameterized_cos.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/cos/parameterized_cos.hpp
- * @brief Class describing the parameterized cos operator
+ * @brief Class describing the parameterized cos operator
*
* This class represents the parameterized unary operator -> cos(alpha * A + a)
*
@@ -55,7 +55,7 @@ protected:
using CosNode::get_latex_expr;
using CosNode::matlab_fxn_expr;
- std::vector<double> _params; //!< The parameters vector
+ std::vector<double> _params; //!< The vector containing the scale and bias terms for the Node
public:
/**
@@ -65,14 +65,13 @@ public:
CosParamNode();
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
* @param l_bound (double) Minimum absolute value allowed for the feature.
* @param u_bound (double) Maximum absolute value allowed for the feature.
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
CosParamNode(
const node_ptr feat,
@@ -83,19 +82,17 @@ public:
);
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
CosParamNode(const node_ptr feat, const unsigned long int feat_ind, std::shared_ptr<NLOptimizer> optimizer);
// DocString: cos_param_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -105,8 +102,8 @@ public:
CosParamNode(const node_ptr feat, const unsigned long int feat_ind, double const l_bound=1e-50, const double u_bound=1e50);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
node_ptr hard_copy()const;
@@ -132,10 +129,10 @@ public:
* @brief Get the pointer to the feature's training data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
- * @return pointer to the feature's test values
+ * @return A pointer to the feature's training data
*/
inline double* value_ptr(int offset=-1, const bool for_comp=false) const {return value_ptr(_params.data(), offset, for_comp);}
@@ -143,7 +140,7 @@ public:
* @brief Get the pointer to the feature's test data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
* @return pointer to the feature's test values
@@ -176,15 +173,15 @@ public:
inline std::string matlab_fxn_expr() const {return matlab_fxn_expr(_params.data());}
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
inline std::vector<double> parameters() const {return _params;}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
void get_parameters(std::shared_ptr<NLOptimizer> optimizer);
@@ -205,7 +202,7 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
inline void set_parameters(const double* params)
{
@@ -228,10 +225,10 @@ public:
* @brief Attempt to generate a new parameterized cosine node cos(\alpha A + a) and add it to feat_list
*
* @param feat_list list of features already generated
- * @param feat feature to attempt to take the cosine of
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat shared_ptr of the feature to operate on (A)
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
* @param optimizer The optimizer used to find the parameters of the node
*/
void generateCosParamNode(
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/div/divide.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/div/divide.cpp
index cbcca7e97e6e73300bd88def0dff64de5486335b..013d9eadb3bf7389ce4610abc43c4b9818fe3c57 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/div/divide.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/div/divide.cpp
@@ -191,7 +191,7 @@ void DivNode::set_test_value(int offset, const bool for_comp) const
{
offset += (offset == -1);
allowed_op_funcs::div(
- _n_test_samp,
+ _n_samp_test,
_feats[0]->test_value_ptr(2 * offset, for_comp),
_feats[1]->test_value_ptr(2 * offset + 1, for_comp),
1.0,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/div/divide.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/div/divide.hpp
index 7732e64842df3af02fae0d4086c4f3b6a4878af4..7d819ba1efe530a2c39285e5c15b4a030adc7be9 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/div/divide.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/div/divide.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/divide.hpp
- * @brief Class describing the division operator
+ * @brief Class describing the division operator
*
* This class represents the binary operator -> A / B
*
@@ -52,8 +52,7 @@ public:
DivNode();
/**
- * @brief Constructor
- * @details Constructs the Node from an array of features to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat_1 (Node) shared_ptr of the first feature to operate on (A)
* @param feat_2 (Node) shared_ptr of the second feature to operate on (B)
@@ -75,8 +74,8 @@ public:
DivNode(const node_ptr feat_1, const node_ptr feat_2, const unsigned long int feat_ind, const double l_bound, const double u_bound);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
virtual node_ptr hard_copy()const;
@@ -132,7 +131,7 @@ public:
// DocString: div_node_rung
/**
- * @brief return the rung of the feature (Height of the binary expression tree + 1)
+ * @brief return the rung of the feature (Height of the binary expression tree - 1)
*
*/
inline int rung(const int cur_rung=0) const {return std::max(_feats[0]->rung(cur_rung + 1), _feats[1]->rung(cur_rung + 1));}
@@ -143,8 +142,7 @@ public:
virtual inline NODE_TYPE type() const {return NODE_TYPE::DIV;}
/**
- * @brief Get the string character representation of the node for the postfix expression
- * @return the string representation of the node for the postfix expression
+ * @brief Get the term used in the postfix expression for this Node
*/
inline std::string get_postfix_term() const {return "div";}
@@ -170,35 +168,34 @@ public:
void check_feats();
/**
- * @brief update the dictionary used to check if an Add/Sub node is valid
+ * @brief update the dictionary used to check if an Add/Sub/AbsDiff node is valid
*
* @param add_sub_leaves the dictionary used to check if an Add/Sub node is valid
- * @param pl_mn if for an addition node: 1 if for a subtraction node: -1
+ * @param pl_mn 1 for addition and -1 for subtraction
* @param expected_abs_tot The expected absolute sum of all values in add_sub_leaves
*/
void update_add_sub_leaves(std::map<std::string, int>& add_sub_leaves, const int pl_mn, int& expected_abs_tot) const;
/**
- * @brief update the dictionary used to check if a mult/div node is valid
+ * @brief update the dictionary used to check if an Mult/Div node is valid
*
* @param div_mult_leaves the dictionary used to check if an Mult/Div node is valid
- * @param fact Factor used to update leaves (power of the feature)
+ * @param fact amount to increment the element (a primary features) of the dictionary by
* @param expected_abs_tot The expected absolute sum of all values in div_mult_leaves
- *
*/
void update_div_mult_leaves(std::map<std::string, double>& div_mult_leaves, const double fact, double& expected_abs_tot) const;
#ifdef PARAMETERIZE
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
virtual std::vector<double> parameters() const {return {};}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
virtual void get_parameters(std::shared_ptr<NLOptimizer> optimizer){return;}
@@ -211,36 +208,36 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
virtual void set_parameters(const double* params){return;}
/**
* @brief Set the value of all training samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief Set the value of all test samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_test_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief A human readable equation representing the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return feature expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return A human readable equation representing the feature
*/
inline std::string expr(const double* params, const int depth=1) const
{
@@ -255,9 +252,9 @@ public:
/**
* @brief Get the valid LaTeX expression that represents the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return the latexified expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return Get the valid LaTeX expression that represents the feature
*/
inline std::string get_latex_expr(const double* params, const int depth=1) const
{
@@ -272,8 +269,8 @@ public:
/**
* @brief Get the string that corresponds to the code needed to evaluate the node in matlab
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
* @return The matlab code for the feature
*/
inline std::string matlab_fxn_expr(const double* params, const int depth=1) const
@@ -287,27 +284,26 @@ public:
}
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Set the upper and lower bounds for the scale and bias term of this Node and its children
*
- * @param lb pointer to the lower bounds data
- * @param ub pointer to the upper bounds data
- * @param from_parent How many parameters are between the start of this node's parameters and its parent
- * @param depth the current depth of the node on the binary expression tree
+ * @param lb A pointer to the location where the lower bounds for the scale and bias term of this Node is set
+ * @param ub A pointer to the location where the upper bounds for the scale and bias term of this Node is set
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void set_bounds(double* lb, double* ub, const int depth=1) const;
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Initialize the scale and bias terms for this Node and its children
*
- * @param params pointer to the parameters vector
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void initialize_params(double* params, const int depth = 1) const;
/**
* @brief Calculates the derivative of an operation with respect to the parameters for a given sample
*
- * @param params pointer to the parameters
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param dfdp pointer to where the feature derivative pointers are located
*/
inline void param_derivative(const double* params, double* dfdp) const
@@ -331,9 +327,9 @@ public:
* @param feat_list list of features already generated
* @param feat_1 The (A) feature
* @param feat_2 The (B) feature
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
*/
void generateDivNode(std::vector<node_ptr>& feat_list, const node_ptr feat_1, const node_ptr feat_2, unsigned long int& feat_ind, const double l_bound, const double u_bound);
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/div/parameterized_divide.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/div/parameterized_divide.cpp
index 59de9952d39b683b2018ac700ea6e0ac74e7147f..67911e01ac03170762a18e6c7ccc70653902ca33 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/div/parameterized_divide.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/div/parameterized_divide.cpp
@@ -199,7 +199,7 @@ void DivNode::set_test_value(const double* params, int offset, const bool for_co
vp_1 = _feats[1]->test_value_ptr(2 * offset + 1, for_comp);
}
allowed_op_funcs::div(
- _n_test_samp,
+ _n_samp_test,
vp_0,
vp_1,
params[0],
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/div/parameterized_divide.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/div/parameterized_divide.hpp
index 0636915e7d84810920a4be7187c580df0766a011..3953d9220f7609bf60051092e63e8862ca46b45c 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/div/parameterized_divide.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/div/parameterized_divide.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/div/parameterized_divide.hpp
- * @brief Class describing the parameterized addition operator
+ * @brief Class describing the parameterized addition operator
*
* This class represents the parameterized unary operator -> A / (alpha * B + a)
*
@@ -65,15 +65,14 @@ public:
DivParamNode();
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
- * @param feat_1 shared_ptr of the feature to operate on (A)
- * @param feat_2 shared_ptr of the feature to operate on (B)
+ * @param feat_1 (Node) shared_ptr of the feature to operate on (A)
+ * @param feat_2 (Node) shared_ptr of the feature to operate on (B)
* @param feat_ind (int) Index of the new feature
* @param l_bound (double) Minimum absolute value allowed for the feature.
* @param u_bound (double) Maximum absolute value allowed for the feature.
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
DivParamNode(
const node_ptr feat_1,
@@ -85,20 +84,18 @@ public:
);
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
- * @param feat_1 shared_ptr of the feature to operate on (A)
- * @param feat_2 shared_ptr of the feature to operate on (B)
+ * @param feat_1 (Node) shared_ptr of the feature to operate on (A)
+ * @param feat_2 (Node) shared_ptr of the feature to operate on (B)
* @param feat_ind (int) Index of the new feature
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
DivParamNode(const node_ptr feat_1, const node_ptr feat_2, const unsigned long int feat_ind, std::shared_ptr<NLOptimizer> optimizer);
// DocString: div_param_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -108,8 +105,8 @@ public:
DivParamNode(const node_ptr feat_1, const node_ptr feat_2, const unsigned long int feat_ind, double const l_bound=1e-50, const double u_bound=1e50);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
node_ptr hard_copy()const;
@@ -135,10 +132,10 @@ public:
* @brief Get the pointer to the feature's training data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
- * @return pointer to the feature's test values
+ * @return A pointer to the feature's training data
*/
inline double* value_ptr(int offset=-1, const bool for_comp=false) const {return value_ptr(_params.data(), offset, for_comp);}
@@ -146,7 +143,7 @@ public:
* @brief Get the pointer to the feature's test data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
* @return pointer to the feature's test values
@@ -179,15 +176,15 @@ public:
inline std::string matlab_fxn_expr() const {return matlab_fxn_expr(_params.data());}
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
inline std::vector<double> parameters() const {return _params;}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
void get_parameters(std::shared_ptr<NLOptimizer> optimizer);
@@ -208,7 +205,7 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
inline void set_parameters(const double* params)
{
@@ -233,9 +230,9 @@ public:
* @param feat_list list of features already generated
* @param feat_1 The (A) feature
* @param feat_2 The (B) feature
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
* @param optimizer The optimizer used to find the parameters of the node
*/
void generateDivParamNode(
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/exp/exponential.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/exp/exponential.cpp
index ce0b644f3db905db7616f22ac2f5cc715e4393ea..8d39be61664d688d0567eb6562599a4e9fbaca39 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/exp/exponential.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/exp/exponential.cpp
@@ -156,7 +156,7 @@ void ExpNode::set_test_value(int offset, const bool for_comp) const
{
offset += (offset == -1);
allowed_op_funcs::exp(
- _n_test_samp,
+ _n_samp_test,
_feats[0]->test_value_ptr(2 * offset, for_comp),
1.0,
0.0,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/exp/exponential.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/exp/exponential.hpp
index c6510d21125bdd8c425d6ba9269a52449af2edf5..d0041b89599cde342f67c6ceb9badfd0bcfdea1c 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/exp/exponential.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/exp/exponential.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/exponential.hpp
- * @brief Class describing the exponential operator
+ * @brief Class describing the exponential operator
*
* This class represents the unary operator -> exp(A)
*
@@ -52,8 +52,7 @@ public:
ExpNode();
/**
- * @brief Constructor
- * @details Constructs the Node from an array of features to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -62,8 +61,7 @@ public:
// DocString: exp_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -73,8 +71,8 @@ public:
ExpNode(const node_ptr feat, const unsigned long int feat_ind, const double l_bound, const double u_bound);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
virtual node_ptr hard_copy()const;
@@ -128,7 +126,7 @@ public:
// DocString: exp_node_rung
/**
- * @brief return the rung of the feature (Height of the binary expression tree + 1)
+ * @brief return the rung of the feature (Height of the binary expression tree - 1)
*
*/
inline int rung(const int cur_rung=0) const {return _feats[0]->rung(cur_rung + 1);}
@@ -139,8 +137,7 @@ public:
virtual inline NODE_TYPE type() const {return NODE_TYPE::EXP;}
/**
- * @brief Get the string character representation of the node for the postfix expression
- * @return the string representation of the node for the postfix expression
+ * @brief Get the term used in the postfix expression for this Node
*/
inline std::string get_postfix_term() const {return "exp";}
@@ -159,35 +156,34 @@ public:
}
/**
- * @brief update the dictionary used to check if an Add/Sub node is valid
+ * @brief update the dictionary used to check if an Add/Sub/AbsDiff node is valid
*
* @param add_sub_leaves the dictionary used to check if an Add/Sub node is valid
- * @param pl_mn if for an addition node: 1 if for a subtraction node: -1
+ * @param pl_mn 1 for addition and -1 for subtraction
* @param expected_abs_tot The expected absolute sum of all values in add_sub_leaves
*/
void update_add_sub_leaves(std::map<std::string, int>& add_sub_leaves, const int pl_mn, int& expected_abs_tot) const;
/**
- * @brief update the dictionary used to check if a mult/div node is valid
+ * @brief update the dictionary used to check if an Mult/Div node is valid
*
* @param div_mult_leaves the dictionary used to check if an Mult/Div node is valid
- * @param fact Factor used to update leaves (power of the feature)
+ * @param fact amount to increment the element (a primary features) of the dictionary by
* @param expected_abs_tot The expected absolute sum of all values in div_mult_leaves
- *
*/
void update_div_mult_leaves(std::map<std::string, double>& div_mult_leaves, const double fact, double& expected_abs_tot) const;
#ifdef PARAMETERIZE
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
virtual std::vector<double> parameters() const {return {};}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
virtual void get_parameters(std::shared_ptr<NLOptimizer> optimizer){return;}
@@ -200,35 +196,35 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
virtual void set_parameters(const double* params){return;}
/**
* @brief Set the value of all training samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief Set the value of all test samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_test_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief A human readable equation representing the feature
*
- * @param params parameter values for non-linear operations
- * @return feature expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @return A human readable equation representing the feature
*/
inline std::string expr(const double* params, const int depth=1) const
{
@@ -242,9 +238,9 @@ public:
/**
* @brief Get the valid LaTeX expression that represents the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return the latexified expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return Get the valid LaTeX expression that represents the feature
*/
inline std::string get_latex_expr(const double* params, const int depth=1) const
{
@@ -258,8 +254,8 @@ public:
/**
* @brief Get the string that corresponds to the code needed to evaluate the node in matlab
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
* @return The matlab code for the feature
*/
inline std::string matlab_fxn_expr(const double* params, const int depth=1) const
@@ -272,27 +268,26 @@ public:
}
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Set the upper and lower bounds for the scale and bias term of this Node and its children
*
- * @param lb pointer to the lower bounds data
- * @param ub pointer to the upper bounds data
- * @param from_parent How many parameters are between the start of this node's parameters and its parent
- * @param depth the current depth of the node on the binary expression tree
+ * @param lb A pointer to the location where the lower bounds for the scale and bias term of this Node is set
+ * @param ub A pointer to the location where the upper bounds for the scale and bias term of this Node is set
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void set_bounds(double* lb, double* ub, const int depth=1) const;
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Initialize the scale and bias terms for this Node and its children
*
- * @param params pointer to the parameters vector
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void initialize_params(double* params, const int depth = 1) const;
/**
* @brief Calculates the derivative of an operation with respect to the parameters for a given sample
*
- * @param params pointer to the parameters
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param dfdp pointer to where the feature derivative pointers are located
*/
inline void param_derivative(const double* params, double* dfdp) const
@@ -309,9 +304,9 @@ public:
*
* @param feat_list list of features already generated
* @param feat feature to attempt to take the exponential of
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
*/
void generateExpNode(
std::vector<node_ptr>& feat_list,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/exp/parameterized_exponential.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/exp/parameterized_exponential.cpp
index 3d06b727b1f9fa5ad62a907b379c76d15afd2c50..a0a0e3a26990f355ef702ccaf27648da7e2ecf5e 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/exp/parameterized_exponential.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/exp/parameterized_exponential.cpp
@@ -185,7 +185,7 @@ void ExpNode::set_test_value(const double* params, int offset, const bool for_co
}
allowed_op_funcs::exp(
- _n_test_samp,
+ _n_samp_test,
vp_0,
params[0],
params[1],
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/exp/parameterized_exponential.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/exp/parameterized_exponential.hpp
index 14a0de44f38cd99eff8aa8fae5405aeee75fa11c..f0f97602aa430059d1c051191f7c7cf22c82a75c 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/exp/parameterized_exponential.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/exp/parameterized_exponential.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/exp/parameterized_exponetial.hpp
- * @brief Class describing the parameterized exp operator
+ * @brief Class describing the parameterized exp operator
*
* This class represents the parameterized unary operator -> exp(alpha * A + a)
*
@@ -55,7 +55,7 @@ protected:
using ExpNode::get_latex_expr;
using ExpNode::matlab_fxn_expr;
- std::vector<double> _params; //!< The parameters vector
+ std::vector<double> _params; //!< The vector containing the scale and bias terms for the Node
public:
/**
@@ -65,14 +65,13 @@ public:
ExpParamNode();
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
* @param l_bound (double) Minimum absolute value allowed for the feature.
* @param u_bound (double) Maximum absolute value allowed for the feature.
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
ExpParamNode(
const node_ptr feat,
@@ -83,18 +82,16 @@ public:
);
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
ExpParamNode(const node_ptr feat, const unsigned long int feat_ind, std::shared_ptr<NLOptimizer> optimizer);
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -104,8 +101,8 @@ public:
ExpParamNode(const node_ptr feat, const unsigned long int feat_ind, double const l_bound=1e-50, const double u_bound=1e50);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
node_ptr hard_copy()const;
@@ -131,10 +128,10 @@ public:
* @brief Get the pointer to the feature's training data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
- * @return pointer to the feature's test values
+ * @return A pointer to the feature's training data
*/
inline double* value_ptr(int offset=-1, const bool for_comp=false) const {return value_ptr(_params.data(), offset, for_comp);}
@@ -142,7 +139,7 @@ public:
* @brief Get the pointer to the feature's test data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
* @return pointer to the feature's test values
@@ -175,15 +172,15 @@ public:
inline std::string matlab_fxn_expr() const {return matlab_fxn_expr(_params.data());}
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
inline std::vector<double> parameters() const {return _params;}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
void get_parameters(std::shared_ptr<NLOptimizer> optimizer);
@@ -204,7 +201,7 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
inline void set_parameters(const double* params)
{
@@ -228,9 +225,9 @@ public:
*
* @param feat_list list of features already generated
* @param feat feature to attempt to take the exponential of
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
* @param optimizer The optimizer used to find the parameters of the node
*/
void generateExpParamNode(
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/inv/inverse.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/inv/inverse.cpp
index 95d2eb616384334a5aee1b1c80383f08269daa5c..dd759485618d99425acfe0e8485bc68604490755 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/inv/inverse.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/inv/inverse.cpp
@@ -139,7 +139,7 @@ void InvNode::set_test_value(int offset, const bool for_comp) const
{
offset += (offset == -1);
allowed_op_funcs::inv(
- _n_test_samp,
+ _n_samp_test,
_feats[0]->test_value_ptr(2 * offset, for_comp),
1.0,
0.0,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/inv/inverse.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/inv/inverse.hpp
index 6d153776c63421046ec716188351e3495ea464fe..6ad2945db76adf07df4b2483521c7f0deec28cef 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/inv/inverse.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/inv/inverse.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/inverse.hpp
- * @brief Class describing the logarithm operator
+ * @brief Class describing the logarithm operator
*
* This class represents the unary operator -> 1 / A
*
@@ -48,8 +48,7 @@ public:
InvNode();
/**
- * @brief Constructor
- * @details Constructs the Node from an array of features to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -69,8 +68,8 @@ public:
InvNode(const node_ptr feat, const unsigned long int feat_ind, const double l_bound, const double u_bound);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
virtual node_ptr hard_copy()const;
@@ -124,7 +123,7 @@ public:
// DocString: inv_node_rung
/**
- * @brief return the rung of the feature (Height of the binary expression tree + 1)
+ * @brief return the rung of the feature (Height of the binary expression tree - 1)
*
*/
inline int rung(const int cur_rung=0) const {return _feats[0]->rung(cur_rung + 1);}
@@ -135,8 +134,7 @@ public:
virtual inline NODE_TYPE type() const {return NODE_TYPE::INV;}
/**
- * @brief Get the string character representation of the node for the postfix expression
- * @return the string representation of the node for the postfix expression
+ * @brief Get the term used in the postfix expression for this Node
*/
inline std::string get_postfix_term() const {return "inv";}
@@ -155,35 +153,34 @@ public:
}
/**
- * @brief update the dictionary used to check if an Add/Sub node is valid
+ * @brief update the dictionary used to check if an Add/Sub/AbsDiff node is valid
*
* @param add_sub_leaves the dictionary used to check if an Add/Sub node is valid
- * @param pl_mn if for an addition node: 1 if for a subtraction node: -1
+ * @param pl_mn 1 for addition and -1 for subtraction
* @param expected_abs_tot The expected absolute sum of all values in add_sub_leaves
*/
void update_add_sub_leaves(std::map<std::string, int>& add_sub_leaves, const int pl_mn, int& expected_abs_tot) const;
/**
- * @brief update the dictionary used to check if a mult/div node is valid
+ * @brief update the dictionary used to check if an Mult/Div node is valid
*
* @param div_mult_leaves the dictionary used to check if an Mult/Div node is valid
- * @param fact Factor used to update leaves (power of the feature)
+ * @param fact amount to increment the element (a primary features) of the dictionary by
* @param expected_abs_tot The expected absolute sum of all values in div_mult_leaves
- *
*/
void update_div_mult_leaves(std::map<std::string, double>& div_mult_leaves, const double fact, double& expected_abs_tot) const;
#ifdef PARAMETERIZE
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
virtual std::vector<double> parameters() const {return {};}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
virtual void get_parameters(std::shared_ptr<NLOptimizer> optimizer){return;}
@@ -196,35 +193,35 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
virtual void set_parameters(const double* params){return;}
/**
* @brief Set the value of all training samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief Set the value of all test samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_test_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief A human readable equation representing the feature
*
- * @param params parameter values for non-linear operations
- * @return feature expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @return A human readable equation representing the feature
*/
inline std::string expr(const double* params, const int depth=1) const
{
@@ -238,9 +235,9 @@ public:
/**
* @brief Get the valid LaTeX expression that represents the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return the latexified expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return Get the valid LaTeX expression that represents the feature
*/
inline std::string get_latex_expr(const double* params, const int depth=1) const
{
@@ -254,8 +251,8 @@ public:
/**
* @brief Get the string that corresponds to the code needed to evaluate the node in matlab
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
* @return The matlab code for the feature
*/
inline std::string matlab_fxn_expr(const double* params, const int depth=1) const
@@ -268,27 +265,26 @@ public:
}
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Set the upper and lower bounds for the scale and bias term of this Node and its children
*
- * @param lb pointer to the lower bounds data
- * @param ub pointer to the upper bounds data
- * @param from_parent How many parameters are between the start of this node's parameters and its parent
- * @param depth the current depth of the node on the binary expression tree
+ * @param lb A pointer to the location where the lower bounds for the scale and bias term of this Node is set
+ * @param ub A pointer to the location where the upper bounds for the scale and bias term of this Node is set
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void set_bounds(double* lb, double* ub, const int depth=1) const;
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Initialize the scale and bias terms for this Node and its children
*
- * @param params pointer to the parameters vector
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void initialize_params(double* params, const int depth = 1) const;
/**
* @brief Calculates the derivative of an operation with respect to the parameters for a given sample
*
- * @param params pointer to the parameters
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param dfdp pointer to where the feature derivative pointers are located
*/
inline void param_derivative(const double* params, double* dfdp) const
@@ -304,9 +300,9 @@ public:
*
* @param feat_list list of features already generated
* @param feat feature to attempt to take the inverse of
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
*/
void generateInvNode(
std::vector<node_ptr>& feat_list,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/inv/parameterized_inverse.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/inv/parameterized_inverse.cpp
index c29dbc4b82261654507d0602fc156d3322702850..8356c716262e52784a4c08492e4324974a15de9f 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/inv/parameterized_inverse.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/inv/parameterized_inverse.cpp
@@ -180,7 +180,7 @@ void InvNode::set_test_value(const double* params, int offset, const bool for_co
vp_0 = _feats[0]->test_value_ptr(2 * offset, for_comp);
}
- allowed_op_funcs::inv(_n_test_samp, vp_0, params[0], params[1], node_value_arrs::access_param_storage_test(rung(), offset, for_comp));
+ allowed_op_funcs::inv(_n_samp_test, vp_0, params[0], params[1], node_value_arrs::access_param_storage_test(rung(), offset, for_comp));
}
void InvNode::set_bounds(double* lb, double* ub, const int depth) const
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/inv/parameterized_inverse.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/inv/parameterized_inverse.hpp
index 623f2187ee39599fc023b19327c5843b7407d17e..1d56bf4be89f11cbb044fa7386da5a249a333f39 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/inv/parameterized_inverse.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/inv/parameterized_inverse.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/inv/parameterized_inverse.hpp
- * @brief Class describing the parameterized inverse operator
+ * @brief Class describing the parameterized inverse operator
*
* This class represents the parameterized unary operator -> 1.0 / (alpha * A + a)
*
@@ -55,7 +55,7 @@ protected:
using InvNode::get_latex_expr;
using InvNode::matlab_fxn_expr;
- std::vector<double> _params; //!< The parameters vector
+ std::vector<double> _params; //!< The vector containing the scale and bias terms for the Node
public:
/**
@@ -65,14 +65,13 @@ public:
InvParamNode();
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
* @param l_bound (double) Minimum absolute value allowed for the feature.
* @param u_bound (double) Maximum absolute value allowed for the feature.
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
InvParamNode(
const node_ptr feat,
@@ -83,19 +82,17 @@ public:
);
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
InvParamNode(const node_ptr feat, const unsigned long int feat_ind, std::shared_ptr<NLOptimizer> optimizer);
// DocString: inv_param_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -105,8 +102,8 @@ public:
InvParamNode(const node_ptr feat, const unsigned long int feat_ind, double const l_bound=1e-50, const double u_bound=1e50);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
node_ptr hard_copy()const;
@@ -132,10 +129,10 @@ public:
* @brief Get the pointer to the feature's training data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
- * @return pointer to the feature's test values
+ * @return A pointer to the feature's training data
*/
inline double* value_ptr(int offset=-1, const bool for_comp=false) const {return value_ptr(_params.data(), offset, for_comp);}
@@ -143,7 +140,7 @@ public:
* @brief Get the pointer to the feature's test data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
* @return pointer to the feature's test values
@@ -176,15 +173,15 @@ public:
inline std::string matlab_fxn_expr() const {return matlab_fxn_expr(_params.data());}
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
inline std::vector<double> parameters() const {return _params;}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
void get_parameters(std::shared_ptr<NLOptimizer> optimizer);
@@ -205,7 +202,7 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
inline void set_parameters(const double* params)
{
@@ -229,9 +226,9 @@ public:
*
* @param feat_list list of features already generated
* @param feat feature to attempt to take the inverse of
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
* @param optimizer The optimizer used to find the parameters of the node
*/
void generateInvParamNode(
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/log/log.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/log/log.cpp
index f6145a038eed9c0c328d77783717932cbd950eab..39b031f33eed54c5ac7dcc6bca594785e1a8da9b 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/log/log.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/log/log.cpp
@@ -181,7 +181,7 @@ void LogNode::set_test_value(int offset, const bool for_comp) const
{
offset += (offset == -1);
allowed_op_funcs::log(
- _n_test_samp,
+ _n_samp_test,
_feats[0]->test_value_ptr(2 * offset, for_comp),
1.0,
0.0,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/log/log.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/log/log.hpp
index 0fcba49018320c4c1f602897bbcd3a76eb0faee6..76a066bbc52b1417c234484cb67251e19f94c7ff 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/log/log.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/log/log.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/log.hpp
- * @brief Class describing the logarithm operator
+ * @brief Class describing the logarithm operator
*
* This class represents the unary operator -> log(A)
*
@@ -52,8 +52,7 @@ public:
LogNode();
/**
- * @brief Constructor
- * @details Constructs the Node from an array of features to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -62,8 +61,7 @@ public:
// DocString: log_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -73,8 +71,8 @@ public:
LogNode(const node_ptr feat, const unsigned long int feat_ind, const double l_bound, const double u_bound);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
virtual node_ptr hard_copy()const;
@@ -119,7 +117,7 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
virtual void set_parameters(const double* params){return;}
@@ -134,7 +132,7 @@ public:
// DocString: log_node_rung
/**
- * @brief return the rung of the feature (Height of the binary expression tree + 1)
+ * @brief return the rung of the feature (Height of the binary expression tree - 1)
*
*/
inline int rung(const int cur_rung=0) const {return _feats[0]->rung(cur_rung + 1);}
@@ -145,8 +143,7 @@ public:
virtual inline NODE_TYPE type() const {return NODE_TYPE::LOG;}
/**
- * @brief Get the string character representation of the node for the postfix expression
- * @return the string representation of the node for the postfix expression
+ * @brief Get the term used in the postfix expression for this Node
*/
inline std::string get_postfix_term() const {return "log";}
@@ -165,35 +162,34 @@ public:
}
/**
- * @brief update the dictionary used to check if an Add/Sub node is valid
+ * @brief update the dictionary used to check if an Add/Sub/AbsDiff node is valid
*
* @param add_sub_leaves the dictionary used to check if an Add/Sub node is valid
- * @param pl_mn if for an addition node: 1 if for a subtraction node: -1
+ * @param pl_mn 1 for addition and -1 for subtraction
* @param expected_abs_tot The expected absolute sum of all values in add_sub_leaves
*/
void update_add_sub_leaves(std::map<std::string, int>& add_sub_leaves, const int pl_mn, int& expected_abs_tot) const;
/**
- * @brief update the dictionary used to check if a mult/div node is valid
+ * @brief update the dictionary used to check if an Mult/Div node is valid
*
* @param div_mult_leaves the dictionary used to check if an Mult/Div node is valid
- * @param fact Factor used to update leaves (power of the feature)
+ * @param fact amount to increment the element (a primary features) of the dictionary by
* @param expected_abs_tot The expected absolute sum of all values in div_mult_leaves
- *
*/
void update_div_mult_leaves(std::map<std::string, double>& div_mult_leaves, const double fact, double& expected_abs_tot) const;
#ifdef PARAMETERIZE
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
virtual std::vector<double> parameters() const {return {};}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
virtual void get_parameters(std::shared_ptr<NLOptimizer> optimizer){return;}
@@ -207,28 +203,28 @@ public:
/**
* @brief Set the value of all training samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief Set the value of all test samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_test_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief A human readable equation representing the feature
*
- * @param params parameter values for non-linear operations
- * @return feature expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @return A human readable equation representing the feature
*/
inline std::string expr(const double* params, const int depth=1) const
{
@@ -243,9 +239,9 @@ public:
/**
* @brief Get the valid LaTeX expression that represents the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return the latexified expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return Get the valid LaTeX expression that represents the feature
*/
inline std::string get_latex_expr(const double* params, const int depth=1) const
{
@@ -260,8 +256,8 @@ public:
/**
* @brief Get the string that corresponds to the code needed to evaluate the node in matlab
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
* @return The matlab code for the feature
*/
inline std::string matlab_fxn_expr(const double* params, const int depth=1) const
@@ -275,27 +271,26 @@ public:
}
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Set the upper and lower bounds for the scale and bias term of this Node and its children
*
- * @param lb pointer to the lower bounds data
- * @param ub pointer to the upper bounds data
- * @param from_parent How many parameters are between the start of this node's parameters and its parent
- * @param depth the current depth of the node on the binary expression tree
+ * @param lb A pointer to the location where the lower bounds for the scale and bias term of this Node is set
+ * @param ub A pointer to the location where the upper bounds for the scale and bias term of this Node is set
+ * @param depth How far down a given Node is from the root OperatorNode
*/
virtual void set_bounds(double* lb, double* ub, const int depth=1) const;
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Initialize the scale and bias terms for this Node and its children
*
- * @param params pointer to the parameters vector
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
*/
virtual void initialize_params(double* params, const int depth = 1) const;
/**
* @brief Calculates the derivative of an operation with respect to the parameters for a given sample
*
- * @param params pointer to the parameters
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param dfdp pointer to where the feature derivative pointers are located
*/
inline void param_derivative(const double* params, double* dfdp) const
@@ -311,9 +306,9 @@ public:
*
* @param feat_list list of features already generated
* @param feat feature to attempt to take the logarithm of
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
*/
void generateLogNode(
std::vector<node_ptr>& feat_list,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/log/parameterized_log.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/log/parameterized_log.cpp
index 110fcd5a3855a95fc2da5aa1a05190bf87bbe5a1..7d3ad1eabf02cb53440ae57d430d69298217c2e2 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/log/parameterized_log.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/log/parameterized_log.cpp
@@ -222,7 +222,7 @@ void LogNode::set_test_value(const double* params, int offset, const bool for_co
}
allowed_op_funcs::log(
- _n_test_samp,
+ _n_samp_test,
vp_0,
params[0],
params[1],
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/log/parameterized_log.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/log/parameterized_log.hpp
index 0482e12b31a9b9d61dc9e6b2886d5923733dcc91..ee77b803f050e15c417b9fc50c2cd314a25d9092 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/log/parameterized_log.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/log/parameterized_log.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/log/parameterized_log.hpp
- * @brief Class describing the parameterized log operator
+ * @brief Class describing the parameterized log operator
*
* This class represents the parameterized unary operator -> log(alpha * A + a)
*
@@ -56,7 +56,7 @@ protected:
using LogNode::get_latex_expr;
using LogNode::matlab_fxn_expr;
- std::vector<double> _params; //!< The parameters vector
+ std::vector<double> _params; //!< The vector containing the scale and bias terms for the Node
double _sign_alpha; //!< Used to alternate between +/- 1
public:
@@ -67,14 +67,13 @@ public:
LogParamNode();
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
* @param l_bound (double) Minimum absolute value allowed for the feature.
* @param u_bound (double) Maximum absolute value allowed for the feature.
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
LogParamNode(
const node_ptr feat,
@@ -85,19 +84,17 @@ public:
);
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
LogParamNode(const node_ptr feat, const unsigned long int feat_ind, std::shared_ptr<NLOptimizer> optimizer);
// DocString: log_param_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -107,8 +104,8 @@ public:
LogParamNode(const node_ptr feat, const unsigned long int feat_ind, double const l_bound=1e-50, const double u_bound=1e50);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
node_ptr hard_copy()const;
@@ -134,10 +131,10 @@ public:
* @brief Get the pointer to the feature's training data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
- * @return pointer to the feature's test values
+ * @return A pointer to the feature's training data
*/
inline double* value_ptr(int offset=-1, const bool for_comp=false) const {return value_ptr(_params.data(), offset, for_comp);}
@@ -145,7 +142,7 @@ public:
* @brief Get the pointer to the feature's test data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
* @return pointer to the feature's test values
@@ -178,33 +175,32 @@ public:
inline std::string matlab_fxn_expr() const {return matlab_fxn_expr(_params.data());}
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Set the upper and lower bounds for the scale and bias term of this Node and its children
*
- * @param lb pointer to the lower bounds data
- * @param ub pointer to the upper bounds data
- * @param from_parent How many parameters are between the start of this node's parameters and its parent
- * @param depth the current depth of the node on the binary expression tree
+ * @param lb A pointer to the location where the lower bounds for the scale and bias term of this Node is set
+ * @param ub A pointer to the location where the upper bounds for the scale and bias term of this Node is set
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void set_bounds(double* lb, double* ub, const int depth=1) const;
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Initialize the scale and bias terms for this Node and its children
*
- * @param params pointer to the parameters vector
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void initialize_params(double* params, const int depth = 1) const;
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
inline std::vector<double> parameters() const {return _params;}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
void get_parameters(std::shared_ptr<NLOptimizer> optimizer);
@@ -225,7 +221,7 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
inline void set_parameters(const double* params)
{
@@ -249,9 +245,9 @@ public:
*
* @param feat_list list of features already generated
* @param feat feature to attempt to take the logarithm of
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
* @param optimizer The optimizer used to find the parameters of the node
*/
void generateLogParamNode(
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/mult/multiply.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/mult/multiply.cpp
index 3b982a4a28e129b1d9e1bb028092e31b3000385e..fac10bb3607a843562b5645fd16b5d2212b8e9ef 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/mult/multiply.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/mult/multiply.cpp
@@ -190,7 +190,7 @@ void MultNode::set_test_value(int offset, const bool for_comp) const
{
offset += (offset == -1);
allowed_op_funcs::mult(
- _n_test_samp,
+ _n_samp_test,
_feats[0]->test_value_ptr(2 * offset, for_comp),
_feats[1]->test_value_ptr(2 * offset + 1, for_comp),
1.0,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/mult/multiply.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/mult/multiply.hpp
index 64e1cf4ae25cf9fe4811bfd0f9ea0aabe1b7f8b5..a9d4592f746512339a2c105354897a08093585e7 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/mult/multiply.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/mult/multiply.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/multiply.hpp
- * @brief Class describing the multiplication operator
+ * @brief Class describing the multiplication operator
*
* This class represents the binary operator -> A * B
*
@@ -53,8 +53,7 @@ public:
MultNode();
/**
- * @brief Constructor
- * @details Constructs the Node from an array of features to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat_1 (Node) shared_ptr of the first feature to operate on (A)
* @param feat_2 (Node) shared_ptr of the second feature to operate on (B)
@@ -64,8 +63,7 @@ public:
// DocString: mult_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the features to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat_1 (Node) shared_ptr of the first feature to operate on (A)
* @param feat_2 (Node) shared_ptr of the second feature to operate on (B)
@@ -76,8 +74,8 @@ public:
MultNode(const node_ptr feat_1, const node_ptr feat_2, const unsigned long int feat_ind, const double l_bound, const double u_bound);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
virtual node_ptr hard_copy()const;
@@ -133,7 +131,7 @@ public:
// DocString: mult_node_rung
/**
- * @brief return the rung of the feature (Height of the binary expression tree + 1)
+ * @brief return the rung of the feature (Height of the binary expression tree - 1)
*
*/
inline int rung(const int cur_rung=0) const {return std::max(_feats[0]->rung(cur_rung + 1), _feats[1]->rung(cur_rung + 1));}
@@ -144,8 +142,7 @@ public:
virtual inline NODE_TYPE type() const {return NODE_TYPE::MULT;}
/**
- * @brief Get the string character representation of the node for the postfix expression
- * @return the string representation of the node for the postfix expression
+ * @brief Get the term used in the postfix expression for this Node
*/
inline std::string get_postfix_term() const {return "mult";}
@@ -171,35 +168,34 @@ public:
void check_feats();
/**
- * @brief update the dictionary used to check if an Add/Sub node is valid
+ * @brief update the dictionary used to check if an Add/Sub/AbsDiff node is valid
*
* @param add_sub_leaves the dictionary used to check if an Add/Sub node is valid
- * @param pl_mn if for an addition node: 1 if for a subtraction node: -1
+ * @param pl_mn 1 for addition and -1 for subtraction
* @param expected_abs_tot The expected absolute sum of all values in add_sub_leaves
*/
void update_add_sub_leaves(std::map<std::string, int>& add_sub_leaves, const int pl_mn, int& expected_abs_tot) const;
/**
- * @brief update the dictionary used to check if a mult/div node is valid
+ * @brief update the dictionary used to check if an Mult/Div node is valid
*
* @param div_mult_leaves the dictionary used to check if an Mult/Div node is valid
- * @param fact Factor used to update leaves (power of the feature)
+ * @param fact amount to increment the element (a primary features) of the dictionary by
* @param expected_abs_tot The expected absolute sum of all values in div_mult_leaves
- *
*/
void update_div_mult_leaves(std::map<std::string, double>& div_mult_leaves, const double fact, double& expected_abs_tot) const;
#ifdef PARAMETERIZE
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
virtual std::vector<double> parameters() const {return {};}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
virtual void get_parameters(std::shared_ptr<NLOptimizer> optimizer){return;}
@@ -212,36 +208,36 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
virtual void set_parameters(const double* params){return;}
/**
* @brief Set the value of all training samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief Set the value of all test samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_test_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief A human readable equation representing the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return feature expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return A human readable equation representing the feature
*/
inline std::string expr(const double* params, const int depth=1) const
{
@@ -256,9 +252,9 @@ public:
/**
* @brief Get the valid LaTeX expression that represents the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return the latexified expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return Get the valid LaTeX expression that represents the feature
*/
inline std::string get_latex_expr(const double* params, const int depth=1) const
{
@@ -273,8 +269,8 @@ public:
/**
* @brief Get the string that corresponds to the code needed to evaluate the node in matlab
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
* @return The matlab code for the feature
*/
inline std::string matlab_fxn_expr(const double* params, const int depth=1) const
@@ -288,27 +284,26 @@ public:
}
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Set the upper and lower bounds for the scale and bias term of this Node and its children
*
- * @param lb pointer to the lower bounds data
- * @param ub pointer to the upper bounds data
- * @param from_parent How many parameters are between the start of this node's parameters and its parent
- * @param depth the current depth of the node on the binary expression tree
+ * @param lb A pointer to the location where the lower bounds for the scale and bias term of this Node is set
+ * @param ub A pointer to the location where the upper bounds for the scale and bias term of this Node is set
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void set_bounds(double* lb, double* ub, const int depth=1) const;
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Initialize the scale and bias terms for this Node and its children
*
- * @param params pointer to the parameters vector
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void initialize_params(double* params, const int depth = 1) const;
/**
* @brief Calculates the derivative of an operation with respect to the parameters for a given sample
*
- * @param params pointer to the parameters
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param dfdp pointer to where the feature derivative pointers are located
*/
inline void param_derivative(const double* params, double* dfdp) const {std::copy_n(_feats[0]->value_ptr(params, 2), _n_samp, dfdp);}
@@ -321,9 +316,9 @@ public:
* @param feat_list list of features already generated
* @param feat_1 The (A) feature
* @param feat_2 The (B) feature
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
*/
void generateMultNode(std::vector<node_ptr>& feat_list, const node_ptr feat_1, const node_ptr feat_2, unsigned long int& feat_ind, const double l_bound, const double u_bound);
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/mult/parameterized_multiply.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/mult/parameterized_multiply.cpp
index 2865945417a0b8c689eb4d5973851f50aa618f39..2af7d4b81c3b1eea115acc0e37d80c357638eb94 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/mult/parameterized_multiply.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/mult/parameterized_multiply.cpp
@@ -172,7 +172,7 @@ void MultNode::set_test_value(const double* params, int offset, const bool for_c
}
allowed_op_funcs::mult(
- _n_test_samp,
+ _n_samp_test,
vp_0,
vp_1,
params[0],
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/mult/parameterized_multiply.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/mult/parameterized_multiply.hpp
index 521c8870da5acdbdf11f21b74632e888bfc60d81..110134997b4f251bdd19cfa0f70fc8934a00527d 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/mult/parameterized_multiply.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/mult/parameterized_multiply.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/div/parameterized_multiply.hpp
- * @brief Class describing the parameterized addition operator
+ * @brief Class describing the parameterized addition operator
*
* This class represents the parameterized unary operator -> A * (alpha * B + a)
*
@@ -65,14 +65,14 @@ public:
MultParamNode();
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
- * @param feat (Node) shared_ptr of the feature to operate on (A)
+ * @param feat_1 (Node) shared_ptr of the feature to operate on (A)
+ * @param feat_2 (Node) shared_ptr of the feature to operate on (B)
* @param feat_ind (int) Index of the new feature
* @param l_bound (double) Minimum absolute value allowed for the feature.
* @param u_bound (double) Maximum absolute value allowed for the feature.
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
MultParamNode(
const node_ptr feat_1,
@@ -84,19 +84,18 @@ public:
);
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
- * @param feat (Node) shared_ptr of the feature to operate on (A)
+ * @param feat_1 (Node) shared_ptr of the feature to operate on (A)
+ * @param feat_2 (Node) shared_ptr of the feature to operate on (B)
* @param feat_ind (int) Index of the new feature
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
MultParamNode(const node_ptr feat_1, const node_ptr feat_2, const unsigned long int feat_ind, std::shared_ptr<NLOptimizer> optimizer);
// DocString: mult_param_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -106,8 +105,8 @@ public:
MultParamNode(const node_ptr feat_1, const node_ptr feat_2, const unsigned long int feat_ind, double const l_bound=1e-50, const double u_bound=1e50);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
node_ptr hard_copy()const;
@@ -133,10 +132,10 @@ public:
* @brief Get the pointer to the feature's training data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
- * @return pointer to the feature's test values
+ * @return A pointer to the feature's training data
*/
inline double* value_ptr(int offset=-1, const bool for_comp=false) const {return value_ptr(_params.data(), offset, for_comp);}
@@ -144,7 +143,7 @@ public:
* @brief Get the pointer to the feature's test data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
* @return pointer to the feature's test values
@@ -177,15 +176,15 @@ public:
inline std::string matlab_fxn_expr() const {return matlab_fxn_expr(_params.data());}
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
inline std::vector<double> parameters() const {return _params;}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
void get_parameters(std::shared_ptr<NLOptimizer> optimizer);
@@ -206,7 +205,7 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
inline void set_parameters(const double* params)
{
@@ -231,9 +230,9 @@ public:
* @param feat_list list of features already generated
* @param feat_1 The (A) feature
* @param feat_2 The (B) feature
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
* @param optimizer The optimizer used to find the parameters of the node
*/
void generateMultParamNode(
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/neg_exp/negative_exponential.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/neg_exp/negative_exponential.cpp
index 8a14190494c203e01ddc800becb316a2b283fa14..98573a7254606232d405c81987021cf132c792e3 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/neg_exp/negative_exponential.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/neg_exp/negative_exponential.cpp
@@ -155,7 +155,7 @@ void NegExpNode::set_test_value(int offset, const bool for_comp) const
{
offset += (offset == -1);
allowed_op_funcs::neg_exp(
- _n_test_samp,
+ _n_samp_test,
_feats[0]->test_value_ptr(2 * offset, for_comp),
1.0,
0.0,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/neg_exp/negative_exponential.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/neg_exp/negative_exponential.hpp
index a0e824806476f4cd2928e6e9b18288bc57136c2c..0abea08159f4dce62d521829ace3f7b9fde79137 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/neg_exp/negative_exponential.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/neg_exp/negative_exponential.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/negative_exponential.hpp
- * @brief Class describing the negative exponential operator
+ * @brief Class describing the negative exponential operator
*
* This class represents the unary operator -> exp(-1 * A)
*
@@ -53,8 +53,7 @@ public:
NegExpNode();
/**
- * @brief Constructor
- * @details Constructs the Node from an array of features to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -63,8 +62,7 @@ public:
// DocString: neg_exp_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -74,8 +72,8 @@ public:
NegExpNode(const node_ptr feat, const unsigned long int feat_ind, const double l_bound, const double u_bound);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
virtual node_ptr hard_copy()const;
@@ -129,7 +127,7 @@ public:
// DocString: neg_exp_node_rung
/**
- * @brief return the rung of the feature (Height of the binary expression tree + 1)
+ * @brief return the rung of the feature (Height of the binary expression tree - 1)
*
*/
inline int rung(const int cur_rung=0) const {return _feats[0]->rung(cur_rung + 1);}
@@ -140,8 +138,7 @@ public:
virtual inline NODE_TYPE type() const {return NODE_TYPE::NEG_EXP;}
/**
- * @brief Get the string character representation of the node for the postfix expression
- * @return the string representation of the node for the postfix expression
+ * @brief Get the term used in the postfix expression for this Node
*/
inline std::string get_postfix_term() const {return "nexp";}
@@ -160,35 +157,34 @@ public:
}
/**
- * @brief update the dictionary used to check if an Add/Sub node is valid
+ * @brief update the dictionary used to check if an Add/Sub/AbsDiff node is valid
*
* @param add_sub_leaves the dictionary used to check if an Add/Sub node is valid
- * @param pl_mn if for an addition node: 1 if for a subtraction node: -1
+ * @param pl_mn 1 for addition and -1 for subtraction
* @param expected_abs_tot The expected absolute sum of all values in add_sub_leaves
*/
void update_add_sub_leaves(std::map<std::string, int>& add_sub_leaves, const int pl_mn, int& expected_abs_tot) const;
/**
- * @brief update the dictionary used to check if a mult/div node is valid
+ * @brief update the dictionary used to check if an Mult/Div node is valid
*
* @param div_mult_leaves the dictionary used to check if an Mult/Div node is valid
- * @param fact Factor used to update leaves (power of the feature)
+ * @param fact amount to increment the element (a primary features) of the dictionary by
* @param expected_abs_tot The expected absolute sum of all values in div_mult_leaves
- *
*/
void update_div_mult_leaves(std::map<std::string, double>& div_mult_leaves, const double fact, double& expected_abs_tot) const;
#ifdef PARAMETERIZE
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
virtual std::vector<double> parameters() const {return {};}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
virtual void get_parameters(std::shared_ptr<NLOptimizer> optimizer){return;}
@@ -201,35 +197,35 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
virtual void set_parameters(const double* params){return;}
/**
* @brief Set the value of all training samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief Set the value of all test samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_test_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief A human readable equation representing the feature
*
- * @param params parameter values for non-linear operations
- * @return feature expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @return A human readable equation representing the feature
*/
inline std::string expr(const double* params, const int depth=1) const
{
@@ -243,9 +239,9 @@ public:
/**
* @brief Get the valid LaTeX expression that represents the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return the latexified expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return Get the valid LaTeX expression that represents the feature
*/
inline std::string get_latex_expr(const double* params, const int depth=1) const
{
@@ -259,8 +255,8 @@ public:
/**
* @brief Get the string that corresponds to the code needed to evaluate the node in matlab
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
* @return The matlab code for the feature
*/
inline std::string matlab_fxn_expr(const double* params, const int depth=1) const
@@ -273,27 +269,26 @@ public:
}
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Set the upper and lower bounds for the scale and bias term of this Node and its children
*
- * @param lb pointer to the lower bounds data
- * @param ub pointer to the upper bounds data
- * @param from_parent How many parameters are between the start of this node's parameters and its parent
- * @param depth the current depth of the node on the binary expression tree
+ * @param lb A pointer to the location where the lower bounds for the scale and bias term of this Node is set
+ * @param ub A pointer to the location where the upper bounds for the scale and bias term of this Node is set
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void set_bounds(double* lb, double* ub, const int depth=1) const;
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Initialize the scale and bias terms for this Node and its children
*
- * @param params pointer to the parameters vector
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void initialize_params(double* params, const int depth = 1) const;
/**
* @brief Calculates the derivative of an operation with respect to the parameters for a given sample
*
- * @param params pointer to the parameters
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param dfdp pointer to where the feature derivative pointers are located
*/
inline void param_derivative(const double* params, double* dfdp) const
@@ -309,9 +304,9 @@ public:
*
* @param feat_list list of features already generated
* @param feat feature to attempt to take the negative exponential of
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
*/
void generateNegExpNode(
std::vector<node_ptr>& feat_list,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/neg_exp/parameterized_negative_exponential.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/neg_exp/parameterized_negative_exponential.cpp
index 12beb08fe3b1eeec921489a72c7e33d0a8e319f9..678307c5abc09ba2bc5c2ea816d3eb9e9004f421 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/neg_exp/parameterized_negative_exponential.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/neg_exp/parameterized_negative_exponential.cpp
@@ -184,7 +184,7 @@ void NegExpNode::set_test_value(const double* params, int offset, const bool for
}
allowed_op_funcs::neg_exp(
- _n_test_samp,
+ _n_samp_test,
vp_0,
params[0],
params[1],
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/neg_exp/parameterized_negative_exponential.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/neg_exp/parameterized_negative_exponential.hpp
index 9fbe705a2ae7688629dc4f461a57f43fa17c8906..830a49565d8bfc6a9e8da570e946de75a0c4a870 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/neg_exp/parameterized_negative_exponential.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/neg_exp/parameterized_negative_exponential.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/neg_exp/parameterized_negative_exponetial.hpp
- * @brief Class describing the parameterized neg_exp operator
+ * @brief Class describing the parameterized neg_exp operator
*
* This class represents the parameterized unary operator -> exp(-alpha * A + a)
*
@@ -55,7 +55,7 @@ protected:
using NegExpNode::get_latex_expr;
using NegExpNode::matlab_fxn_expr;
- std::vector<double> _params; //!< The parameters vector
+ std::vector<double> _params; //!< The vector containing the scale and bias terms for the Node
public:
/**
@@ -65,14 +65,13 @@ public:
NegExpParamNode();
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
* @param l_bound (double) Minimum absolute value allowed for the feature.
* @param u_bound (double) Maximum absolute value allowed for the feature.
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
NegExpParamNode(
const node_ptr feat,
@@ -83,19 +82,17 @@ public:
);
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
NegExpParamNode(const node_ptr feat, const unsigned long int feat_ind, std::shared_ptr<NLOptimizer> optimizer);
// DocString: neg_exp_param_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -105,8 +102,8 @@ public:
NegExpParamNode(const node_ptr feat, const unsigned long int feat_ind, double const l_bound=1e-50, const double u_bound=1e50);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
node_ptr hard_copy()const;
@@ -140,10 +137,10 @@ public:
* @brief Get the pointer to the feature's training data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
- * @return pointer to the feature's test values
+ * @return A pointer to the feature's training data
*/
inline double* value_ptr(int offset=-1, const bool for_comp=false) const {return value_ptr(_params.data(), offset, for_comp);}
@@ -151,7 +148,7 @@ public:
* @brief Get the pointer to the feature's test data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
* @return pointer to the feature's test values
@@ -176,15 +173,15 @@ public:
inline std::string get_latex_expr() const {return get_latex_expr(_params.data());}
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
inline std::vector<double> parameters() const {return _params;}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
void get_parameters(std::shared_ptr<NLOptimizer> optimizer);
@@ -205,7 +202,7 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
inline void set_parameters(const double* params)
{
@@ -229,9 +226,9 @@ public:
*
* @param feat_list list of features already generated
* @param feat feature to attempt to take the negative exponential of
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
* @param optimizer The optimizer used to find the parameters of the node
*/
void generateNegExpParamNode(
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sin/parameterized_sin.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sin/parameterized_sin.cpp
index 0e2190eac2e2fdb5234aff9e50a3a03e2ed94f74..b782a9c6796652787fecaca807d894e147037bae 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sin/parameterized_sin.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sin/parameterized_sin.cpp
@@ -172,7 +172,7 @@ void SinNode::set_test_value(const double* params, int offset, const bool for_co
}
allowed_op_funcs::sin(
- _n_test_samp,
+ _n_samp_test,
vp_0,
params[0],
params[1],
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sin/parameterized_sin.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sin/parameterized_sin.hpp
index 9b3b79ca55f4139d239b9adc288a68096fec232e..b2558cbaf1ae1a423dd7d81c2e06255975b92c7e 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sin/parameterized_sin.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sin/parameterized_sin.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/sin/parameterized_sin.hpp
- * @brief Class describing the parameterized sin operator
+ * @brief Class describing the parameterized sin operator
*
* This class represents the parameterized unary operator -> sin(alpha * A + a)
*
@@ -55,7 +55,7 @@ protected:
using SinNode::get_latex_expr;
using SinNode::matlab_fxn_expr;
- std::vector<double> _params; //!< The parameters vector
+ std::vector<double> _params; //!< The vector containing the scale and bias terms for the Node
public:
/**
@@ -65,14 +65,13 @@ public:
SinParamNode();
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
* @param l_bound (double) Minimum absolute value allowed for the feature.
* @param u_bound (double) Maximum absolute value allowed for the feature.
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
SinParamNode(
const node_ptr feat,
@@ -83,19 +82,17 @@ public:
);
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
SinParamNode(const node_ptr feat, const unsigned long int feat_ind, std::shared_ptr<NLOptimizer> optimizer);
// DocString: sin_param_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -105,8 +102,8 @@ public:
SinParamNode(const node_ptr feat, const unsigned long int feat_ind, double const l_bound=1e-50, const double u_bound=1e50);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
node_ptr hard_copy()const;
@@ -132,10 +129,10 @@ public:
* @brief Get the pointer to the feature's training data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
- * @return pointer to the feature's test values
+ * @return A pointer to the feature's training data
*/
inline double* value_ptr(int offset=-1, const bool for_comp=false) const {return value_ptr(_params.data(), offset, for_comp);}
@@ -143,7 +140,7 @@ public:
* @brief Get the pointer to the feature's test data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
* @return pointer to the feature's test values
@@ -176,15 +173,15 @@ public:
inline std::string matlab_fxn_expr() const {return matlab_fxn_expr(_params.data());}
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
inline std::vector<double> parameters() const {return _params;}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
void get_parameters(std::shared_ptr<NLOptimizer> optimizer);
@@ -205,7 +202,7 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
inline void set_parameters(const double* params)
{
@@ -229,9 +226,9 @@ public:
*
* @param feat_list list of features already generated
* @param feat feature to attempt to take the sine of
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
* @param optimizer The optimizer used to find the parameters of the node
*/
void generateSinParamNode(
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sin/sin.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sin/sin.cpp
index db351a356e02e8e1765b1c5dc05f18a5131e9b37..67749f03d912dadf0c1e03932a3dc9f543f25a2c 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sin/sin.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sin/sin.cpp
@@ -133,7 +133,7 @@ void SinNode::set_test_value(int offset, const bool for_comp) const
{
offset += (offset == -1);
allowed_op_funcs::sin(
- _n_test_samp,
+ _n_samp_test,
_feats[0]->test_value_ptr(2 * offset, for_comp),
1.0,
0.0,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sin/sin.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sin/sin.hpp
index 4dbe38f495cd4299a2aeaa7f719202adbddb1d92..f790fd55159f3f8941fc875de63f27291e049923 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sin/sin.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sin/sin.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/sin.hpp
- * @brief Class describing the sine operator
+ * @brief Class describing the sine operator
*
* This class represents the unary operator -> sin(A)
*
@@ -53,8 +53,7 @@ public:
SinNode();
/**
- * @brief Constructor
- * @details Constructs the Node from an array of features to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -63,8 +62,7 @@ public:
// DocString: sin_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -74,8 +72,8 @@ public:
SinNode(const node_ptr feat, const unsigned long int feat_ind, const double l_bound, const double u_bound);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
virtual node_ptr hard_copy()const;
@@ -129,7 +127,7 @@ public:
// DocString: sin_node_rung
/**
- * @brief return the rung of the feature (Height of the binary expression tree + 1)
+ * @brief return the rung of the feature (Height of the binary expression tree - 1)
*
*/
inline int rung(const int cur_rung=0) const {return _feats[0]->rung(cur_rung + 1);}
@@ -140,8 +138,7 @@ public:
virtual inline NODE_TYPE type() const {return NODE_TYPE::SIN;}
/**
- * @brief Get the string character representation of the node for the postfix expression
- * @return the string representation of the node for the postfix expression
+ * @brief Get the term used in the postfix expression for this Node
*/
inline std::string get_postfix_term() const {return "sin";}
@@ -160,35 +157,34 @@ public:
}
/**
- * @brief update the dictionary used to check if an Add/Sub node is valid
+ * @brief update the dictionary used to check if an Add/Sub/AbsDiff node is valid
*
* @param add_sub_leaves the dictionary used to check if an Add/Sub node is valid
- * @param pl_mn if for an addition node: 1 if for a subtraction node: -1
+ * @param pl_mn 1 for addition and -1 for subtraction
* @param expected_abs_tot The expected absolute sum of all values in add_sub_leaves
*/
void update_add_sub_leaves(std::map<std::string, int>& add_sub_leaves, const int pl_mn, int& expected_abs_tot) const;
/**
- * @brief update the dictionary used to check if a mult/div node is valid
+ * @brief update the dictionary used to check if an Mult/Div node is valid
*
* @param div_mult_leaves the dictionary used to check if an Mult/Div node is valid
- * @param fact Factor used to update leaves (power of the feature)
+ * @param fact amount to increment the element (a primary features) of the dictionary by
* @param expected_abs_tot The expected absolute sum of all values in div_mult_leaves
- *
*/
void update_div_mult_leaves(std::map<std::string, double>& div_mult_leaves, const double fact, double& expected_abs_tot) const;
#ifdef PARAMETERIZE
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
virtual std::vector<double> parameters() const {return {};}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
virtual void get_parameters(std::shared_ptr<NLOptimizer> optimizer){return;}
@@ -201,35 +197,35 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
virtual void set_parameters(const double* params){return;}
/**
* @brief Set the value of all training samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief Set the value of all test samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_test_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief A human readable equation representing the feature
*
- * @param params parameter values for non-linear operations
- * @return feature expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @return A human readable equation representing the feature
*/
inline std::string expr(const double* params, const int depth=1) const
{
@@ -244,9 +240,9 @@ public:
/**
* @brief Get the valid LaTeX expression that represents the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return the latexified expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return Get the valid LaTeX expression that represents the feature
*/
inline std::string get_latex_expr(const double* params, const int depth=1) const
{
@@ -261,8 +257,8 @@ public:
/**
* @brief Get the string that corresponds to the code needed to evaluate the node in matlab
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
* @return The matlab code for the feature
*/
inline std::string matlab_fxn_expr(const double* params, const int depth=1) const
@@ -276,27 +272,26 @@ public:
}
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Set the upper and lower bounds for the scale and bias term of this Node and its children
*
- * @param lb pointer to the lower bounds data
- * @param ub pointer to the upper bounds data
- * @param from_parent How many parameters are between the start of this node's parameters and its parent
- * @param depth the current depth of the node on the binary expression tree
+ * @param lb A pointer to the location where the lower bounds for the scale and bias term of this Node is set
+ * @param ub A pointer to the location where the upper bounds for the scale and bias term of this Node is set
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void set_bounds(double* lb, double* ub, const int depth=1) const;
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Initialize the scale and bias terms for this Node and its children
*
- * @param params pointer to the parameters vector
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void initialize_params(double* params, const int depth = 1) const;
/**
* @brief Calculates the derivative of an operation with respect to the parameters for a given sample
*
- * @param params pointer to the parameters
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param dfdp pointer to where the feature derivative pointers are located
*/
inline void param_derivative(const double* params, double* dfdp) const
@@ -312,9 +307,9 @@ public:
*
* @param feat_list list of features already generated
* @param feat feature to attempt to take the sine of
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
*/
void generateSinNode(
std::vector<node_ptr>& feat_list,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/six_pow/parameterized_sixth_power.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/six_pow/parameterized_sixth_power.cpp
index dc0330676f1ae2dd417730e7a2688a3005c3fb57..164c84cbfc57f77c52ecda36153dfe89c6caa773 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/six_pow/parameterized_sixth_power.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/six_pow/parameterized_sixth_power.cpp
@@ -187,7 +187,7 @@ void SixPowNode::set_test_value(const double* params, int offset, const bool for
}
allowed_op_funcs::sixth_pow(
- _n_test_samp,
+ _n_samp_test,
vp_0,
params[0],
params[1],
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/six_pow/parameterized_sixth_power.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/six_pow/parameterized_sixth_power.hpp
index 1bc58f05f9b323298dc9690af329e65de2e94bc2..94228b00730aa7c17c05fdabae3b696d9170b882 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/six_pow/parameterized_sixth_power.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/six_pow/parameterized_sixth_power.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/six_pow/parameterized_sixth_pow.hpp
- * @brief Class describing the parameterized sixth power operator
+ * @brief Class describing the parameterized sixth power operator
*
* This class represents the parameterized unary operator -> (alpha * A + a)^6
*
@@ -55,7 +55,7 @@ protected:
using SixPowNode::get_latex_expr;
using SixPowNode::matlab_fxn_expr;
- std::vector<double> _params; //!< The parameters vector
+ std::vector<double> _params; //!< The vector containing the scale and bias terms for the Node
public:
/**
@@ -65,14 +65,13 @@ public:
SixPowParamNode();
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
* @param l_bound (double) Minimum absolute value allowed for the feature.
* @param u_bound (double) Maximum absolute value allowed for the feature.
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
SixPowParamNode(
const node_ptr feat,
@@ -83,19 +82,17 @@ public:
);
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
SixPowParamNode(const node_ptr feat, const unsigned long int feat_ind, std::shared_ptr<NLOptimizer> optimizer);
// DocString: six_pow_param_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -105,8 +102,8 @@ public:
SixPowParamNode(const node_ptr feat, const unsigned long int feat_ind, double const l_bound=1e-50, const double u_bound=1e50);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
node_ptr hard_copy()const;
@@ -132,10 +129,10 @@ public:
* @brief Get the pointer to the feature's training data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
- * @return pointer to the feature's test values
+ * @return A pointer to the feature's training data
*/
inline double* value_ptr(int offset=-1, const bool for_comp=false) const {return value_ptr(_params.data(), offset, for_comp);}
@@ -143,7 +140,7 @@ public:
* @brief Get the pointer to the feature's test data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
* @return pointer to the feature's test values
@@ -176,15 +173,15 @@ public:
inline std::string matlab_fxn_expr() const {return matlab_fxn_expr(_params.data());}
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
inline std::vector<double> parameters() const {return _params;}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
void get_parameters(std::shared_ptr<NLOptimizer> optimizer);
@@ -205,7 +202,7 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
inline void set_parameters(const double* params)
{
@@ -229,9 +226,9 @@ public:
*
* @param feat_list list of features already generated
* @param feat feature to attempt to take the sixth power of
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
* @param optimizer The optimizer used to find the parameters of the node
*/
void generateSixPowParamNode(
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/six_pow/sixth_power.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/six_pow/sixth_power.cpp
index 5293033e5b5f7526a8e52bc092e34aefd07b8253..503ccf65b4cfa9abd52145f83146984e16ffa3cd 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/six_pow/sixth_power.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/six_pow/sixth_power.cpp
@@ -152,7 +152,7 @@ void SixPowNode::set_test_value(int offset, const bool for_comp) const
{
offset += (offset == -1);
allowed_op_funcs::sixth_pow(
- _n_test_samp,
+ _n_samp_test,
_feats[0]->test_value_ptr(2 * offset, for_comp),
1.0,
0.0,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/six_pow/sixth_power.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/six_pow/sixth_power.hpp
index 2342a5c5bad713d5262dc5741a9165fe3c816f98..c0d85f625c4f4e5b7b606eae3b5feffc86d29e6f 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/six_pow/sixth_power.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/six_pow/sixth_power.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/sixth_power.hpp
- * @brief Class describing the sixth power operator
+ * @brief Class describing the sixth power operator
*
* This class represents the unary operator -> (A)^6
*
@@ -53,8 +53,7 @@ public:
SixPowNode();
/**
- * @brief Constructor
- * @details Constructs the Node from an array of features to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -63,8 +62,7 @@ public:
// DocString: six_pow_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -74,8 +72,8 @@ public:
SixPowNode(const node_ptr feat, const unsigned long int feat_ind, const double l_bound, const double u_bound);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
virtual node_ptr hard_copy()const;
@@ -129,7 +127,7 @@ public:
// DocString: six_pow_node_rung
/**
- * @brief return the rung of the feature (Height of the binary expression tree + 1)
+ * @brief return the rung of the feature (Height of the binary expression tree - 1)
*
*/
inline int rung(const int cur_rung=0) const {return _feats[0]->rung(cur_rung + 1);}
@@ -140,8 +138,7 @@ public:
virtual inline NODE_TYPE type() const {return NODE_TYPE::SIX_POW;}
/**
- * @brief Get the string character representation of the node for the postfix expression
- * @return the string representation of the node for the postfix expression
+ * @brief Get the term used in the postfix expression for this Node
*/
inline std::string get_postfix_term() const {return "sp";}
@@ -160,35 +157,34 @@ public:
}
/**
- * @brief update the dictionary used to check if an Add/Sub node is valid
+ * @brief update the dictionary used to check if an Add/Sub/AbsDiff node is valid
*
* @param add_sub_leaves the dictionary used to check if an Add/Sub node is valid
- * @param pl_mn if for an addition node: 1 if for a subtraction node: -1
+ * @param pl_mn 1 for addition and -1 for subtraction
* @param expected_abs_tot The expected absolute sum of all values in add_sub_leaves
*/
void update_add_sub_leaves(std::map<std::string, int>& add_sub_leaves, const int pl_mn, int& expected_abs_tot) const;
/**
- * @brief update the dictionary used to check if a mult/div node is valid
+ * @brief update the dictionary used to check if an Mult/Div node is valid
*
* @param div_mult_leaves the dictionary used to check if an Mult/Div node is valid
- * @param fact Factor used to update leaves (power of the feature)
+ * @param fact amount to increment the element (a primary features) of the dictionary by
* @param expected_abs_tot The expected absolute sum of all values in div_mult_leaves
- *
*/
void update_div_mult_leaves(std::map<std::string, double>& div_mult_leaves, const double fact, double& expected_abs_tot) const;
#ifdef PARAMETERIZE
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
virtual std::vector<double> parameters() const {return {};}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
virtual void get_parameters(std::shared_ptr<NLOptimizer> optimizer){return;}
@@ -201,36 +197,36 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
virtual void set_parameters(const double* params){return;}
/**
* @brief Set the value of all training samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief Set the value of all test samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_test_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief A human readable equation representing the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return feature expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return A human readable equation representing the feature
*/
inline std::string expr(const double* params, const int depth=1) const
{
@@ -244,9 +240,9 @@ public:
/**
* @brief Get the valid LaTeX expression that represents the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return the latexified expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return Get the valid LaTeX expression that represents the feature
*/
inline std::string get_latex_expr(const double* params, const int depth=1) const
{
@@ -260,8 +256,8 @@ public:
/**
* @brief Get the string that corresponds to the code needed to evaluate the node in matlab
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
* @return The matlab code for the feature
*/
inline std::string matlab_fxn_expr(const double* params, const int depth=1) const
@@ -274,27 +270,26 @@ public:
}
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Set the upper and lower bounds for the scale and bias term of this Node and its children
*
- * @param lb pointer to the lower bounds data
- * @param ub pointer to the upper bounds data
- * @param from_parent How many parameters are between the start of this node's parameters and its parent
- * @param depth the current depth of the node on the binary expression tree
+ * @param lb A pointer to the location where the lower bounds for the scale and bias term of this Node is set
+ * @param ub A pointer to the location where the upper bounds for the scale and bias term of this Node is set
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void set_bounds(double* lb, double* ub, const int depth=1) const;
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Initialize the scale and bias terms for this Node and its children
*
- * @param params pointer to the parameters vector
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void initialize_params(double* params, const int depth = 1) const;
/**
* @brief Calculates the derivative of an operation with respect to the parameters for a given sample
*
- * @param params pointer to the parameters
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param dfdp pointer to where the feature derivative pointers are located
*/
inline void param_derivative(const double* params, double* dfdp) const
@@ -310,9 +305,9 @@ public:
*
* @param feat_list list of features already generated
* @param feat feature to attempt to take the sixth power of
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
*/
void generateSixPowNode(
std::vector<node_ptr>& feat_list,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sq/parameterized_square.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sq/parameterized_square.cpp
index 1c0b4be23fee10a1e4e229835838db3b2bed2917..8bfe1d78d0ff58bad413247ce4f5f6ad71c291f0 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sq/parameterized_square.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sq/parameterized_square.cpp
@@ -178,7 +178,7 @@ void SqNode::set_test_value(const double* params, int offset, const bool for_com
}
allowed_op_funcs::sq(
- _n_test_samp,
+ _n_samp_test,
vp_0,
params[0],
params[1],
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sq/parameterized_square.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sq/parameterized_square.hpp
index 5f8c0e5c2ebcceead7ba962b4520dd982fcf7ec6..ee609023ef10b048107b93528f5d280f3dd87f77 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sq/parameterized_square.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sq/parameterized_square.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/sq/parameterized_square.hpp
- * @brief Class describing the parameterized square operator
+ * @brief Class describing the parameterized square operator
*
* This class represents the parameterized unary operator -> (alpha * A + a)^2
*
@@ -55,7 +55,7 @@ protected:
using SqNode::get_latex_expr;
using SqNode::matlab_fxn_expr;
- std::vector<double> _params; //!< The parameters vector
+ std::vector<double> _params; //!< The vector containing the scale and bias terms for the Node
public:
/**
@@ -65,14 +65,13 @@ public:
SqParamNode();
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
* @param l_bound (double) Minimum absolute value allowed for the feature.
* @param u_bound (double) Maximum absolute value allowed for the feature.
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
SqParamNode(
const node_ptr feat,
@@ -83,19 +82,17 @@ public:
);
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
SqParamNode(const node_ptr feat, const unsigned long int feat_ind, std::shared_ptr<NLOptimizer> optimizer);
// DocString: sq_param_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -105,8 +102,8 @@ public:
SqParamNode(const node_ptr feat, const unsigned long int feat_ind, double const l_bound=1e-50, const double u_bound=1e50);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
node_ptr hard_copy()const;
@@ -132,10 +129,10 @@ public:
* @brief Get the pointer to the feature's training data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
- * @return pointer to the feature's test values
+ * @return A pointer to the feature's training data
*/
inline double* value_ptr(int offset=-1, const bool for_comp=false) const {return value_ptr(_params.data(), offset, for_comp);}
@@ -143,7 +140,7 @@ public:
* @brief Get the pointer to the feature's test data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
* @return pointer to the feature's test values
@@ -176,15 +173,15 @@ public:
inline std::string matlab_fxn_expr() const {return matlab_fxn_expr(_params.data());}
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
inline std::vector<double> parameters() const {return _params;}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
void get_parameters(std::shared_ptr<NLOptimizer> optimizer);
@@ -205,7 +202,7 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
inline void set_parameters(const double* params)
{
@@ -229,9 +226,9 @@ public:
*
* @param feat_list list of features already generated
* @param feat feature to attempt to take the square of
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
* @param optimizer The optimizer used to find the parameters of the node
*/
void generateSqParamNode(
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sq/square.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sq/square.cpp
index 5ffa0b365a33e0d6b5bfcaf9ce69e30fea786e15..3ed4aa33bcc582397788bdb96d5169d7d13989e6 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sq/square.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sq/square.cpp
@@ -142,7 +142,7 @@ void SqNode::set_test_value(int offset, const bool for_comp) const
{
offset += (offset == -1);
allowed_op_funcs::sq(
- _n_test_samp,
+ _n_samp_test,
_feats[0]->test_value_ptr(2 * offset, for_comp),
1.0,
0.0,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sq/square.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sq/square.hpp
index 09ba6495ab18bd3056bc5ff6907429e3ecc00c70..011bea6fee17e50b851e4074fd3a941a7c5b4dbf 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sq/square.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sq/square.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/square.hpp
- * @brief Class describing the square operator
+ * @brief Class describing the square operator
*
* This class represents the unary operator -> (A)^2
*
@@ -52,8 +52,7 @@ public:
SqNode();
/**
- * @brief Constructor
- * @details Constructs the Node from an array of features to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -62,8 +61,7 @@ public:
// DocString: sq_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -73,8 +71,8 @@ public:
SqNode(const node_ptr feat, const unsigned long int feat_ind, const double l_bound, const double u_bound);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
virtual node_ptr hard_copy()const;
@@ -128,7 +126,7 @@ public:
// DocString: sq_node_rung
/**
- * @brief return the rung of the feature (Height of the binary expression tree + 1)
+ * @brief return the rung of the feature (Height of the binary expression tree - 1)
*
*/
inline int rung(const int cur_rung=0) const {return _feats[0]->rung(cur_rung + 1);}
@@ -139,8 +137,7 @@ public:
virtual inline NODE_TYPE type() const {return NODE_TYPE::SQ;}
/**
- * @brief Get the string character representation of the node for the postfix expression
- * @return the string representation of the node for the postfix expression
+ * @brief Get the term used in the postfix expression for this Node
*/
inline std::string get_postfix_term() const {return "sq";}
@@ -159,35 +156,34 @@ public:
}
/**
- * @brief update the dictionary used to check if an Add/Sub node is valid
+ * @brief update the dictionary used to check if an Add/Sub/AbsDiff node is valid
*
* @param add_sub_leaves the dictionary used to check if an Add/Sub node is valid
- * @param pl_mn if for an addition node: 1 if for a subtraction node: -1
+ * @param pl_mn 1 for addition and -1 for subtraction
* @param expected_abs_tot The expected absolute sum of all values in add_sub_leaves
*/
void update_add_sub_leaves(std::map<std::string, int>& add_sub_leaves, const int pl_mn, int& expected_abs_tot) const;
/**
- * @brief update the dictionary used to check if a mult/div node is valid
+ * @brief update the dictionary used to check if an Mult/Div node is valid
*
* @param div_mult_leaves the dictionary used to check if an Mult/Div node is valid
- * @param fact Factor used to update leaves (power of the feature)
+ * @param fact amount to increment the element (a primary features) of the dictionary by
* @param expected_abs_tot The expected absolute sum of all values in div_mult_leaves
- *
*/
void update_div_mult_leaves(std::map<std::string, double>& div_mult_leaves, const double fact, double& expected_abs_tot) const;
#ifdef PARAMETERIZE
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
virtual std::vector<double> parameters() const {return {};}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
virtual void get_parameters(std::shared_ptr<NLOptimizer> optimizer){return;}
@@ -200,36 +196,36 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
virtual void set_parameters(const double* params){return;}
/**
* @brief Set the value of all training samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief Set the value of all test samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_test_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief A human readable equation representing the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return feature expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return A human readable equation representing the feature
*/
inline std::string expr(const double* params, const int depth=1) const
{
@@ -243,9 +239,9 @@ public:
/**
* @brief Get the valid LaTeX expression that represents the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return the latexified expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return Get the valid LaTeX expression that represents the feature
*/
inline std::string get_latex_expr(const double* params, const int depth=1) const
{
@@ -259,8 +255,8 @@ public:
/**
* @brief Get the string that corresponds to the code needed to evaluate the node in matlab
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
* @return The matlab code for the feature
*/
inline std::string matlab_fxn_expr(const double* params, const int depth=1) const
@@ -273,27 +269,26 @@ public:
}
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Set the upper and lower bounds for the scale and bias term of this Node and its children
*
- * @param lb pointer to the lower bounds data
- * @param ub pointer to the upper bounds data
- * @param from_parent How many parameters are between the start of this node's parameters and its parent
- * @param depth the current depth of the node on the binary expression tree
+ * @param lb A pointer to the location where the lower bounds for the scale and bias term of this Node is set
+ * @param ub A pointer to the location where the upper bounds for the scale and bias term of this Node is set
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void set_bounds(double* lb, double* ub, const int depth=1) const;
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Initialize the scale and bias terms for this Node and its children
*
- * @param params pointer to the parameters vector
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void initialize_params(double* params, const int depth = 1) const;
/**
* @brief Calculates the derivative of an operation with respect to the parameters for a given sample
*
- * @param params pointer to the parameters
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param dfdp pointer to where the feature derivative pointers are located
*/
inline void param_derivative(const double* params, double* dfdp) const
@@ -309,9 +304,9 @@ public:
*
* @param feat_list list of features already generated
* @param feat feature to attempt to take the square of
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
*/
void generateSqNode(
std::vector<node_ptr>& feat_list,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sqrt/parameterized_square_root.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sqrt/parameterized_square_root.cpp
index 0a2bb3f48437b46369d9f5b5ebf0a1f9075a19d4..8f7f1fca02be820f23f2e531599a304d8a201a60 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sqrt/parameterized_square_root.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sqrt/parameterized_square_root.cpp
@@ -204,7 +204,7 @@ void SqrtNode::set_test_value(const double* params, int offset, const bool for_c
}
allowed_op_funcs::sqrt(
- _n_test_samp,
+ _n_samp_test,
vp_0,
params[0],
params[1],
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sqrt/parameterized_square_root.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sqrt/parameterized_square_root.hpp
index b31f596c337e75b3cfe73d8e899537bac4eeee05..429bdc24c8369d0b2cbfd7d5433007be1a41d60a 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sqrt/parameterized_square_root.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sqrt/parameterized_square_root.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/sqrt/parameterized_square_root.hpp
- * @brief Class describing the parameterized square root operator
+ * @brief Class describing the parameterized square root operator
*
* This class represents the parameterized unary operator -> sqrt(alpha * A + a)
*
@@ -56,7 +56,7 @@ protected:
using SqrtNode::get_latex_expr;
using SqrtNode::matlab_fxn_expr;
- std::vector<double> _params; //!< The parameters vector
+ std::vector<double> _params; //!< The vector containing the scale and bias terms for the Node
double _sign_alpha; //!< 1 if alpha is positive, -1 if alpha is negative
public:
@@ -67,14 +67,13 @@ public:
SqrtParamNode();
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
* @param l_bound (double) Minimum absolute value allowed for the feature.
* @param u_bound (double) Maximum absolute value allowed for the feature.
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
SqrtParamNode(
const node_ptr feat,
@@ -85,19 +84,17 @@ public:
);
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
SqrtParamNode(const node_ptr feat, const unsigned long int feat_ind, std::shared_ptr<NLOptimizer> optimizer);
// DocString: sqrt_param_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -107,8 +104,8 @@ public:
SqrtParamNode(const node_ptr feat, const unsigned long int feat_ind, double const l_bound=1e-50, const double u_bound=1e50);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
node_ptr hard_copy()const;
@@ -134,10 +131,10 @@ public:
* @brief Get the pointer to the feature's training data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
- * @return pointer to the feature's test values
+ * @return A pointer to the feature's training data
*/
inline double* value_ptr(int offset=-1, const bool for_comp=false) const {return value_ptr(_params.data(), offset, for_comp);}
@@ -145,7 +142,7 @@ public:
* @brief Get the pointer to the feature's test data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
* @return pointer to the feature's test values
@@ -178,15 +175,15 @@ public:
inline std::string matlab_fxn_expr() const {return matlab_fxn_expr(_params.data());}
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
inline std::vector<double> parameters() const {return _params;}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
void get_parameters(std::shared_ptr<NLOptimizer> optimizer);
@@ -207,7 +204,7 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
inline void set_parameters(const double* params)
{
@@ -215,20 +212,19 @@ public:
}
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Set the upper and lower bounds for the scale and bias term of this Node and its children
*
- * @param lb pointer to the lower bounds data
- * @param ub pointer to the upper bounds data
- * @param from_parent How many parameters are between the start of this node's parameters and its parent
- * @param depth the current depth of the node on the binary expression tree
+ * @param lb A pointer to the location where the lower bounds for the scale and bias term of this Node is set
+ * @param ub A pointer to the location where the upper bounds for the scale and bias term of this Node is set
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void set_bounds(double* lb, double* ub, const int depth=1) const;
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Initialize the scale and bias terms for this Node and its children
*
- * @param params pointer to the parameters vector
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void initialize_params(double* params, const int depth = 1) const;
@@ -249,9 +245,9 @@ public:
*
* @param feat_list list of features already generated
* @param feat feature to attempt to take the square root of
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
* @param optimizer The optimizer used to find the parameters of the node
*/
void generateSqrtParamNode(
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sqrt/square_root.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sqrt/square_root.cpp
index 9c076cee87dcb2d8aaf5c9c9d0fd67bb09f4a9be..03da01cd0d81bb27e23778a532be59af3a0cbf3c 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sqrt/square_root.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sqrt/square_root.cpp
@@ -151,7 +151,7 @@ void SqrtNode::set_test_value(int offset, const bool for_comp) const
{
offset += (offset == -1);
allowed_op_funcs::sqrt(
- _n_test_samp,
+ _n_samp_test,
_feats[0]->test_value_ptr(2 * offset, for_comp),
1.0,
0.0,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sqrt/square_root.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sqrt/square_root.hpp
index a130bd9668f4acbe11650a3c4b88e8580fcb16e7..10a747e329f2560f266af9ee7b31fbbce3af790d 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sqrt/square_root.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sqrt/square_root.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/square_root.hpp
- * @brief Class describing the square root operator
+ * @brief Class describing the square root operator
*
* This class represents the unary operator -> (A)^(1/2)
*
@@ -53,8 +53,7 @@ public:
SqrtNode();
/**
- * @brief Constructor
- * @details Constructs the Node from an array of features to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -63,8 +62,7 @@ public:
// DocString: sqrt_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat (Node) shared_ptr of the feature to operate on (A)
* @param feat_ind (int) Index of the new feature
@@ -74,8 +72,8 @@ public:
SqrtNode(const node_ptr feat, const unsigned long int feat_ind, const double l_bound, const double u_bound);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
virtual node_ptr hard_copy()const;
@@ -129,7 +127,7 @@ public:
// DocString: sqrt_node_rung
/**
- * @brief return the rung of the feature (Height of the binary expression tree + 1)
+ * @brief return the rung of the feature (Height of the binary expression tree - 1)
*
*/
inline int rung(const int cur_rung=0) const {return _feats[0]->rung(cur_rung + 1);}
@@ -140,8 +138,7 @@ public:
virtual inline NODE_TYPE type() const {return NODE_TYPE::SQRT;}
/**
- * @brief Get the string character representation of the node for the postfix expression
- * @return the string representation of the node for the postfix expression
+ * @brief Get the term used in the postfix expression for this Node
*/
inline std::string get_postfix_term() const {return "sqrt";}
@@ -160,35 +157,34 @@ public:
}
/**
- * @brief update the dictionary used to check if an Add/Sub node is valid
+ * @brief update the dictionary used to check if an Add/Sub/AbsDiff node is valid
*
* @param add_sub_leaves the dictionary used to check if an Add/Sub node is valid
- * @param pl_mn if for an addition node: 1 if for a subtraction node: -1
+ * @param pl_mn 1 for addition and -1 for subtraction
* @param expected_abs_tot The expected absolute sum of all values in add_sub_leaves
*/
void update_add_sub_leaves(std::map<std::string, int>& add_sub_leaves, const int pl_mn, int& expected_abs_tot) const;
/**
- * @brief update the dictionary used to check if a mult/div node is valid
+ * @brief update the dictionary used to check if an Mult/Div node is valid
*
* @param div_mult_leaves the dictionary used to check if an Mult/Div node is valid
- * @param fact Factor used to update leaves (power of the feature)
+ * @param fact amount to increment the element (a primary features) of the dictionary by
* @param expected_abs_tot The expected absolute sum of all values in div_mult_leaves
- *
*/
void update_div_mult_leaves(std::map<std::string, double>& div_mult_leaves, const double fact, double& expected_abs_tot) const;
#ifdef PARAMETERIZE
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
virtual std::vector<double> parameters() const {return {};}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
virtual void get_parameters(std::shared_ptr<NLOptimizer> optimizer){return;}
@@ -201,35 +197,35 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
virtual void set_parameters(const double* params){return;}
/**
* @brief Set the value of all training samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief Set the value of all test samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_test_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief A human readable equation representing the feature
*
- * @param params parameter values for non-linear operations
- * @return feature expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @return A human readable equation representing the feature
*/
inline std::string expr(const double* params, const int depth=1) const
{
@@ -244,9 +240,9 @@ public:
/**
* @brief Get the valid LaTeX expression that represents the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return the latexified expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return Get the valid LaTeX expression that represents the feature
*/
inline std::string get_latex_expr(const double* params, const int depth=1) const
{
@@ -261,8 +257,8 @@ public:
/**
* @brief Get the string that corresponds to the code needed to evaluate the node in matlab
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
* @return The matlab code for the feature
*/
inline std::string matlab_fxn_expr(const double* params, const int depth=1) const
@@ -276,27 +272,26 @@ public:
}
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Set the upper and lower bounds for the scale and bias term of this Node and its children
*
- * @param lb pointer to the lower bounds data
- * @param ub pointer to the upper bounds data
- * @param from_parent How many parameters are between the start of this node's parameters and its parent
- * @param depth the current depth of the node on the binary expression tree
+ * @param lb A pointer to the location where the lower bounds for the scale and bias term of this Node is set
+ * @param ub A pointer to the location where the upper bounds for the scale and bias term of this Node is set
+ * @param depth How far down a given Node is from the root OperatorNode
*/
virtual void set_bounds(double* lb, double* ub, const int depth=1) const;
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Initialize the scale and bias terms for this Node and its children
*
- * @param params pointer to the parameters vector
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
*/
virtual void initialize_params(double* params, const int depth = 1) const;
/**
* @brief Calculates the derivative of an operation with respect to the parameters for a given sample
*
- * @param params pointer to the parameters
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param dfdp pointer to where the feature derivative pointers are located
*/
inline void param_derivative(const double* params, double* dfdp) const
@@ -312,9 +307,9 @@ public:
*
* @param feat_list list of features already generated
* @param feat feature to attempt to take the square root of
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
*/
void generateSqrtNode(
std::vector<node_ptr>& feat_list,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sub/parameterized_subtract.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sub/parameterized_subtract.cpp
index 9e5ab703b73a62f8ccac443ce3469acc84cce9e7..c94afde4bd897bab835cdf91a3c7190ae01645e1 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sub/parameterized_subtract.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sub/parameterized_subtract.cpp
@@ -172,7 +172,7 @@ void SubNode::set_test_value(const double* params, int offset, const bool for_co
vp_1 = _feats[1]->test_value_ptr(2 * offset + 1, for_comp);
}
allowed_op_funcs::sub(
- _n_test_samp,
+ _n_samp_test,
vp_0,
vp_1,
params[0],
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sub/parameterized_subtract.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sub/parameterized_subtract.hpp
index 04cc6f2bf47f55b8c3094531a0974f61004edb8a..216cbac11ee8ba2e00030d80e7e5871926b1febd 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sub/parameterized_subtract.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sub/parameterized_subtract.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/sub/parameterized_subtract.hpp
- * @brief Class describing the parameterized addition operator
+ * @brief Class describing the parameterized addition operator
*
* This class represents the parameterized unary operator -> A - (alpha * B + a)
*
@@ -65,15 +65,14 @@ public:
SubParamNode();
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
- * @param feat_1 shared_ptr of the feature to operate on (A)
- * @param feat_2 shared_ptr of the feature to operate on (B)
+ * @param feat_1 (Node) shared_ptr of the feature to operate on (A)
+ * @param feat_2 (Node) shared_ptr of the feature to operate on (B)
* @param feat_ind (int) Index of the new feature
* @param l_bound (double) Minimum absolute value allowed for the feature.
* @param u_bound (double) Maximum absolute value allowed for the feature.
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
SubParamNode(
const node_ptr feat_1,
@@ -85,23 +84,21 @@ public:
);
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
- * @param feat_1 shared_ptr of the feature to operate on (A)
- * @param feat_2 shared_ptr of the feature to operate on (B)
+ * @param feat_1 (Node) shared_ptr of the feature to operate on (A)
+ * @param feat_2 (Node) shared_ptr of the feature to operate on (B)
* @param feat_ind (int) Index of the new feature
- * @param optimizer The optimizer to find the optimal parameters for the features
+ * @param optimizer (NLOptimizer) The optimizer to find the optimal parameters for the features
*/
SubParamNode(const node_ptr feat_1, const node_ptr feat_2, const unsigned long int feat_ind, std::shared_ptr<NLOptimizer> optimizer);
// DocString: sub_param_node_init
/**
* @brief Constructor
- * @details Constructs the Node from node pointer of the feature to operate on
*
- * @param feat_1 shared_ptr of the feature to operate on (A)
- * @param feat_2 shared_ptr of the feature to operate on (B)
+ * @param feat_1 (Node) shared_ptr of the feature to operate on (A)
+ * @param feat_2 (Node) shared_ptr of the feature to operate on (B)
* @param feat_ind (int) Index of the new feature
* @param l_bound (double) Minimum absolute value allowed for the feature.
* @param u_bound (double) Maximum absolute value allowed for the feature.
@@ -109,8 +106,8 @@ public:
SubParamNode(const node_ptr feat_1, const node_ptr feat_2, const unsigned long int feat_ind, double const l_bound=1e-50, const double u_bound=1e50);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
node_ptr hard_copy()const;
@@ -136,10 +133,10 @@ public:
* @brief Get the pointer to the feature's training data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
- * @return pointer to the feature's test values
+ * @return A pointer to the feature's training data
*/
inline double* value_ptr(int offset=-1, const bool for_comp=false) const {return value_ptr(_params.data(), offset, for_comp);}
@@ -147,7 +144,7 @@ public:
* @brief Get the pointer to the feature's test data
* @details If the feature is not already stored in memory, then calculate the feature and return the pointer to the data
*
- * @param offset(int) the integer value to offset the location in the temporary storage array
+ * @param offset (int) the integer value to offset the location in the temporary storage array
* @param for_comp (bool) If true then the evaluation is used for comparing features
*
* @return pointer to the feature's test values
@@ -180,15 +177,15 @@ public:
inline std::string matlab_fxn_expr() const {return matlab_fxn_expr(_params.data());}
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
inline std::vector<double> parameters() const {return _params;}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
void get_parameters(std::shared_ptr<NLOptimizer> optimizer);
@@ -209,7 +206,7 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
inline void set_parameters(const double* params)
{
@@ -234,9 +231,9 @@ public:
* @param feat_list list of features already generated
* @param feat_1 The (A) feature
* @param feat_2 The (B) feature
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
* @param optimizer The optimizer used to find the parameters of the node
*/
void generateSubParamNode(
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sub/subtract.cpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sub/subtract.cpp
index 2a39757c771bc190eed7476ef96b9e2fbc41d104..e9a97a34dfda440e90057cfbb717ca7edf4ffbce 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sub/subtract.cpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sub/subtract.cpp
@@ -181,7 +181,7 @@ void SubNode::set_test_value(int offset, const bool for_comp) const
{
offset += (offset == -1);
allowed_op_funcs::sub(
- _n_test_samp,
+ _n_samp_test,
_feats[0]->test_value_ptr(2 * offset, for_comp),
_feats[1]->test_value_ptr(2 * offset + 1, for_comp),
1.0,
diff --git a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sub/subtract.hpp b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sub/subtract.hpp
index e0a1ce5a2fc96fec4dbf6988e9e5513c89f814d8..90315e047b55f8b430e4386740e692fa7c52a6ad 100644
--- a/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sub/subtract.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_operator_nodes/sub/subtract.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_operator_nodes/subtract.hpp
- * @brief Class describing the subtraction operator
+ * @brief Class describing the subtraction operator
*
* This class represents the binary operator -> A - B
*
@@ -54,8 +54,7 @@ public:
SubNode();
/**
- * @brief Constructor
- * @details Constructs the Node from an array of features to operate on
+ * @brief Constructor excluding bounds on the maximum absolute value of the Node
*
* @param feat_1 (Node) shared_ptr of the first feature to operate on (A)
* @param feat_2 (Node) shared_ptr of the second feature to operate on (B)
@@ -65,8 +64,7 @@ public:
// DocString: sub_node_init
/**
- * @brief Constructor
- * @details Constructs the Node from node pointer of the features to operate on
+ * @brief Constructor including bounds on the maximum absolute value of the Node
*
* @param feat_1 (Node) shared_ptr of the first feature to operate on (A)
* @param feat_2 (Node) shared_ptr of the second feature to operate on (B)
@@ -77,8 +75,8 @@ public:
SubNode(const node_ptr feat_1, const node_ptr feat_2, const unsigned long int feat_ind, const double l_bound, const double u_bound);
/**
- * @brief Makes a hard copy node (all items in _feats are a new shared_ptr)
- * @return A shared_ptr of a copy of this node with all members hard copied
+ * @brief Makes a hard copy node (All members of the Node are independent of the original one)
+ * @return A shared_ptr to the copied node
*/
virtual node_ptr hard_copy()const;
@@ -134,7 +132,7 @@ public:
// DocString: sub_node_rung
/**
- * @brief return the rung of the feature (Height of the binary expression tree + 1)
+ * @brief return the rung of the feature (Height of the binary expression tree - 1)
*
*/
inline int rung(const int cur_rung=0) const {return std::max(_feats[0]->rung(cur_rung + 1), _feats[1]->rung(cur_rung + 1));}
@@ -145,8 +143,7 @@ public:
virtual inline NODE_TYPE type() const {return NODE_TYPE::SUB;}
/**
- * @brief Get the string character representation of the node for the postfix expression
- * @return the string representation of the node for the postfix expression
+ * @brief Get the term used in the postfix expression for this Node
*/
inline std::string get_postfix_term() const {return "sub";}
@@ -172,35 +169,34 @@ public:
void check_feats();
/**
- * @brief update the dictionary used to check if an Add/Sub node is valid
+ * @brief update the dictionary used to check if an Add/Sub/AbsDiff node is valid
*
* @param add_sub_leaves the dictionary used to check if an Add/Sub node is valid
- * @param pl_mn if for an addition node: 1 if for a subtraction node: -1
+ * @param pl_mn 1 for addition and -1 for subtraction
* @param expected_abs_tot The expected absolute sum of all values in add_sub_leaves
*/
void update_add_sub_leaves(std::map<std::string, int>& add_sub_leaves, const int pl_mn, int& expected_abs_tot) const;
/**
- * @brief update the dictionary used to check if a mult/div node is valid
+ * @brief update the dictionary used to check if an Mult/Div node is valid
*
* @param div_mult_leaves the dictionary used to check if an Mult/Div node is valid
- * @param fact Factor used to update leaves (power of the feature)
+ * @param fact amount to increment the element (a primary features) of the dictionary by
* @param expected_abs_tot The expected absolute sum of all values in div_mult_leaves
- *
*/
void update_div_mult_leaves(std::map<std::string, double>& div_mult_leaves, const double fact, double& expected_abs_tot) const;
#ifdef PARAMETERIZE
/**
- * @brief The parameters used for introducing more non linearity in the operators
+ * @brief The parameters used for including individual scale and bias terms to each operator in the Node
*/
virtual std::vector<double> parameters() const {return {};}
/**
- * @brief Solve the non-linear optimization to set the parameters
- * @details Fits the data points from _feats->value_ptr and prop to get the parameters for the feature
+ * @brief Optimize the scale and bias terms for each operation in the Node.
+ * @details Use optimizer to find the scale and bias terms that minimizes the associated loss function
*
- * @param optimizer The optimizer used to find the best scale and shift parameters
+ * @param optimizer The optimizer used to evaluate the loss function for each optimization and find the optimal parameters
*/
virtual void get_parameters(std::shared_ptr<NLOptimizer> optimizer){return;}
@@ -213,36 +209,36 @@ public:
/**
* @brief Set the non-linear parameters
- * @param params The new parameters of the node
+ * @param params The new scale and bias terms of this node
*/
virtual void set_parameters(const double* params){return;}
/**
* @brief Set the value of all training samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief Set the value of all test samples for the feature inside the central data storage array
*
- * @param params pointer to the non-linear parameter values
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param offset (int) Where the current node is in the binary expression tree relative to other nodes at the same depth
* @param for_comp (bool) If true then the evaluation is used for comparing features
- * @param depth the max depth of paramterization (in the binary expression tree)
+ * @param depth (int) How far down a given Node is from the root OperatorNode
*/
void set_test_value(const double* params, int offset=-1, const bool for_comp=false, const int depth=1) const;
/**
* @brief A human readable equation representing the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return feature expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return A human readable equation representing the feature
*/
inline std::string expr(const double* params, const int depth=1) const
{
@@ -257,9 +253,9 @@ public:
/**
* @brief Get the valid LaTeX expression that represents the feature
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
- * @return the latexified expression
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
+ * @return Get the valid LaTeX expression that represents the feature
*/
inline std::string get_latex_expr(const double* params, const int depth=1) const
{
@@ -274,8 +270,8 @@ public:
/**
* @brief Get the string that corresponds to the code needed to evaluate the node in matlab
*
- * @param params parameter values for non-linear operations
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
* @return The matlab code for the feature
*/
inline std::string matlab_fxn_expr(const double* params, const int depth=1) const
@@ -289,27 +285,26 @@ public:
}
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Set the upper and lower bounds for the scale and bias term of this Node and its children
*
- * @param lb pointer to the lower bounds data
- * @param ub pointer to the upper bounds data
- * @param from_parent How many parameters are between the start of this node's parameters and its parent
- * @param depth the current depth of the node on the binary expression tree
+ * @param lb A pointer to the location where the lower bounds for the scale and bias term of this Node is set
+ * @param ub A pointer to the location where the upper bounds for the scale and bias term of this Node is set
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void set_bounds(double* lb, double* ub, const int depth=1) const;
/**
- * @brief Set the bounds for the nl parameterization
+ * @brief Initialize the scale and bias terms for this Node and its children
*
- * @param params pointer to the parameters vector
- * @param depth the current depth of the node on the binary expression tree
+ * @param params A pointer to the bias and scale terms for this Node and its children
+ * @param depth How far down a given Node is from the root OperatorNode
*/
void initialize_params(double* params, const int depth = 1) const;
/**
* @brief Calculates the derivative of an operation with respect to the parameters for a given sample
*
- * @param params pointer to the parameters
+ * @param params A pointer to the bias and scale terms for this Node and its children
* @param dfdp pointer to where the feature derivative pointers are located
*/
inline void param_derivative(const double* params, double* dfdp) const {std::fill_n(dfdp, _n_samp, -1.0);}
@@ -322,9 +317,9 @@ public:
* @param feat_list list of features already generated
* @param feat_1 The (A) feature
* @param feat_2 The (B) feature
- * @param feat_ind The index of the last generated feature
- * @param l_bound lower bound for the maximum feature value
- * @param u_bound upper bound for the maximum feature value
+ * @param feat_ind Index of the new feature
+ * @param l_bound Minimum absolute value allowed for the feature.
+ * @param u_bound Maximum absolute value allowed for the feature.
*/
void generateSubNode(std::vector<node_ptr>& feat_list, const node_ptr feat_1, const node_ptr feat_2, unsigned long int& feat_ind, const double l_bound, const double u_bound);
diff --git a/src/feature_creation/node/operator_nodes/allowed_ops.hpp b/src/feature_creation/node/operator_nodes/allowed_ops.hpp
index 1415ea7f1a9e8916b42edc47eef8b2c29f78c0f3..fab3fff6aa6c3c43866d82d8438aad3077ab6161 100644
--- a/src/feature_creation/node/operator_nodes/allowed_ops.hpp
+++ b/src/feature_creation/node/operator_nodes/allowed_ops.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/allowed_ops.hpp
- * @brief Definition of the map that converts the string representation of an operation to the correct OperatorNode constructor
+ * @brief Definition of the map that converts the string representation of an operation to the correct OperatorNode constructor
*
* @author Thomas A. R. Purcell (tpurcell)
* @bug No known bugs.
diff --git a/src/feature_creation/node/operator_nodes/functions.hpp b/src/feature_creation/node/operator_nodes/functions.hpp
index fc394bb47d26181a9586cda6a3f5f720d4e89a8c..46a55d4064cc9f981eefcdc6f020271aaba6263d 100644
--- a/src/feature_creation/node/operator_nodes/functions.hpp
+++ b/src/feature_creation/node/operator_nodes/functions.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/operator_nodes/functions.hpp
- * @brief Definition of all allowed functions OperatorNodes can have
+ * @brief Definition of all allowed functions OperatorNodes can have
*
* @author Thomas A. R. Purcell (tpurcell)
* @bug No known bugs.
diff --git a/src/feature_creation/node/utils.hpp b/src/feature_creation/node/utils.hpp
index 4ee003e83746c9faf5342c09fb0b2ab96f2ca401..032b56524e2dda1a0142d445c024ea9ea18d28ed 100644
--- a/src/feature_creation/node/utils.hpp
+++ b/src/feature_creation/node/utils.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/utils.hpp
- * @brief utility functions to build node_ptrs from strings
+ * @brief utility functions to build node_ptrs from strings
*
* @author Thomas A. R. Purcell (tpurcell)
* @bug No known bugs.
@@ -28,12 +28,12 @@
namespace str2node
{
/**
- * @brief Convert a postfix expression into a node_ptr
+ * @brief Convert a postfix expression of a feature into a node_ptr
* @details Creates a stack to iteratively generate the feature represented by the expression
*
* @param postfix_expr The postfix expression of the feature node
- * @param phi_0 The initial feature set
- * @param feat_ind The desired feature index
+ * @param phi_0 The primary feature space
+ * @param feat_ind The index of the new feature
* @return The feature node described by the postfix expression
*/
node_ptr postfix2node(const std::string postfix_expr, const std::vector<node_ptr>& phi_0, unsigned long int& feat_ind);
@@ -50,7 +50,7 @@ node_ptr postfix2node(const std::string postfix_expr, const std::vector<node_ptr
std::vector<node_ptr> phi_selected_from_file(const std::string filename, const std::vector<node_ptr> phi_0);
/**
- * @brief Convert a text file into the feature space
+ * @brief Convert a text file containing postfix expressions of features into the feature space
* @details Read in the file to get the postfix expressions and regenerate the features using phi_0
*
* @param filename The name of the file storing all the features
diff --git a/src/feature_creation/node/value_storage/nodes_value_containers.hpp b/src/feature_creation/node/value_storage/nodes_value_containers.hpp
index 07131e6e92bfc364ac2bb98d83186d05f373c07b..31a3b990e5f159cf9d69ce083b7d6bd63f303b1c 100644
--- a/src/feature_creation/node/value_storage/nodes_value_containers.hpp
+++ b/src/feature_creation/node/value_storage/nodes_value_containers.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/node/value_storage/node_value_containers.hpp
- * @brief Centeral storage area for all descriptor matrix data
+ * @brief Centeral storage area for all descriptor matrix data
*
* This file contains definitions where all descriptor/feature data is stored permanently or in temporary storage arrays
* All Nodes's (except ModelNodes) value_ptr/test_value_ptrs point to data here.
@@ -42,40 +42,45 @@ namespace py = boost::python;
namespace node_value_arrs
{
- extern std::vector<double> TEMP_STORAGE_ARR; //!< Array to temporarily store feature training data
- extern std::vector<double> VALUES_ARR; //!< Training data of the stored features
- extern std::vector<double> TEMP_STORAGE_TEST_ARR; //!< Array to temporarily store feature test data
- extern std::vector<double> TEST_VALUES_ARR; //!< Value of the stored features test data
- extern std::vector<double> D_MATRIX; //!< The descriptor matrix
- extern std::vector<double> PARAM_STORAGE_ARR; //!< Array to store parameterized feature values
- extern std::vector<double> PARAM_STORAGE_TEST_ARR; //!< Array to store parameterized feature values for the test set
-
- extern std::vector<int> TEMP_STORAGE_REG; //!< Register to see which feature is stored in each slot for the training data
- extern std::vector<int> TEMP_STORAGE_TEST_REG; //!< Register to see which feature is stored in each slot for the test data
+ extern std::vector<double> VALUES_ARR; //!< The central storage location for the values of the training data for each feature with a rung <= N_RUNGS_STORED
+ extern std::vector<double> TEST_VALUES_ARR; //!< The central storage location for the values of the test data for each feature with a rung <= N_RUNGS_STORED
+
+ extern std::vector<double> TEMP_STORAGE_ARR; //!< The vector used to temporarily store the values of each feature with a rung > N_RUNGS_STORED (These are calculated on the fly when the values are required)
+ extern std::vector<double> TEMP_STORAGE_TEST_ARR; //!< The vector used to temporarily store the values of each feature with a rung > N_RUNGS_STORED (These are calculated on the fly when the values are required)
+ extern std::vector<int> TEMP_STORAGE_REG; //!< Register that maps the slots in TEMP_STORAGE_ARR to the index of the feature whose data is currently stored there (reset by setting all elements of this vector to -1)
+ extern std::vector<int> TEMP_STORAGE_TEST_REG; //!< Register that maps the slots in TEMP_STORAGE_TEST_ARR to the index of the feature whose data is currently stored there (reset by setting all elements of this vector to -1)
+
+ extern std::vector<double> PARAM_STORAGE_ARR; //!< The vector used to temporarily store the values of each feature that has free-parameters (These are calculated on the fly when the values are required)
+ extern std::vector<double> PARAM_STORAGE_TEST_ARR; //!< The vector used to temporarily store the values of each feature that has free-parameters (These are calculated on the fly when the values are required)
+
+ extern std::vector<double> D_MATRIX; //!< The descriptor matrix (Central storage for the selected feature space)
+
extern std::vector<int> TASK_SZ_TRAIN; //!< Number of training samples per task
extern std::vector<int> TASK_SZ_TEST; //!< Number of test sample per task
- extern int N_SELECTED; //!< Number of features selected
- extern int N_SAMPLES; //!< Number of training samples for each feature
- extern int N_SAMPLES_TEST; //!< Number of test samples for each feature
- extern int N_PRIMARY_FEATURES; //!< Number of primary features with stored values
+ extern int N_SELECTED; //!< Number of selected features
+
+ extern int N_SAMPLES; //!< Number of training samples for each feature (Sum of all elements in TASK_SZ_TRAIN)
+ extern int N_SAMPLES_TEST; //!< Number of test samples for each feature (Sum of all elements in TASK_SZ_TEST)
+
+ extern int N_PRIMARY_FEATURES; //!< Number of primary features
extern int N_STORE_FEATURES; //!< Number of features with stored values
- extern int N_RUNGS_STORED; //!< Number of rungs with values stored
- extern int MAX_N_THREADS; //!< Get the maximum number of threads possible
- extern int N_OP_SLOTS; //!< The number of possible operator slots
- extern int N_PARAM_OP_SLOTS; //!< The maximum number of possible operator slots
+ extern int N_RUNGS_STORED; //!< Maximum rung for permanently storing a features value
extern int MAX_RUNG; //!< The maximum rung for all features
+ extern int MAX_N_THREADS; //!< Get the maximum number of threads possible for a calculation
+ extern int N_OP_SLOTS; //!< The number of possible nodes of the binary expression tree that maybe calculated on the fly
+ extern int N_PARAM_OP_SLOTS; //!< The number of possible non-leaf nodes of the binary expression tree
+
/**
- * @brief Initialize the node value arrays
- * @details Using the size of the initial feature space constructor the storage arrays
+ * @brief Initialize all central storage vectors/descriptive variables
*
- * @param n_samples Number of training samples for each feature
- * @param n_samples_test Number of test samples for each feature
- * @param n_primary_feat Number of primary features
- * @param max_rung Largest rung of a feature
+ * @param n_samples The number of training samples for each feature (Sum of all elements in TASK_SZ_TRAIN)
+ * @param n_samples_test The number of test samples for each feature (Sum of all elements in TASK_SZ_TEST)
+ * @param n_primary_feat The number of primary features
+ * @param max_rung The maximum rung for all features
* @param set_test_task_sz If True reset the task_sz vectors
- * @param use_params If True set up parameter storage
+ * @param use_params If True set up parameterized feature storage as well
*/
void initialize_values_arr(
const int n_samples,
@@ -88,13 +93,12 @@ namespace node_value_arrs
// DocString: node_vals_init_no_ts
/**
- * @brief Initialize the node value arrays
- * @details Using the size of the initial feature space constructor the storage arrays
+ * @brief Initialize all central storage vectors/descriptive variables
*
- * @param n_samples Number of training samples for each feature
- * @param n_samples_test Number of test samples for each feature
- * @param n_primary_feat Number of primary features
- * @param max_rung Largest rung of a feature
+ * @param n_samples The number of training samples for each feature (Sum of all elements in TASK_SZ_TRAIN)
+ * @param n_samples_test The number of test samples for each feature (Sum of all elements in TASK_SZ_TEST)
+ * @param n_primary_feat The number of primary features
+ * @param max_rung The maximum rung for all features
*/
inline void initialize_values_arr(
const int n_samples,
@@ -107,14 +111,13 @@ namespace node_value_arrs
}
/**
- * @brief Initialize the node value arrays
- * @details Using the size of the initial feature space constructor the storage arrays
+ * @brief Initialize all central storage vectors/descriptive variables
*
- * @param task_sz_train Number of training samples per task
- * @param task_sz_test Number of test sample per task
- * @param n_primary_feat Number of primary features
- * @param max_rung Largest rung of a feature
- * @param use_params If True set up parameter storage
+ * @param task_sz_train The number of training samples per task
+ * @param task_sz_test The number of test sample per task
+ * @param n_primary_feat The number of primary features
+ * @param max_rung The maximum rung for all features
+ * @param use_params If True set up parameterized feature storage as well
*/
void initialize_values_arr(
const std::vector<int> task_sz_train,
@@ -125,31 +128,29 @@ namespace node_value_arrs
);
/**
- * @brief Resize the node value arrays
- * @details Using the updated number of features increase the size of the storage arrays
+ * @brief Resize the central storage array given a new number of features and the current rung of the features
*
- * @param n_dims Number of dimensions to store
- * @param n_feat number of features to store
- * @param use_temp If true keep the temporary_storage
+ * @param n_rung The current rung of the generated features
+ * @param n_feat The new number of features to store
*/
void resize_values_arr(const int n_dims, const int n_feat);
+ // DocString: node_vals_init_param
/**
* @brief Initialize the parameter storage array
*/
+
void initialize_param_storage();
- // DocString: node_vlas_init_d_mat
+ // DocString: node_vals_init_d_mat
/**
* @brief Initialize the descriptor matrix
- * @details Initialize an empty descriptor matrix
*
*/
void initialize_d_matrix_arr();
/**
- * @brief Resize the descriptor matrix
- * @details Using the number of newly selected features increase the size of the descriptor matrix
+ * @brief Resize the descriptor matrix for the new number of selected features
*
* @param n_select Number of features to select
*/
@@ -173,7 +174,7 @@ namespace node_value_arrs
* @brief Get the operator slot associated with a given rung/offset
*
* @param rung Rung of the feature
- * @param offset Offset used to prevent overwrites
+ * @param offset Offset used to prevent overwrites (determines where in the binary expression tree this operator is)
* @param for_comp If true get a slot dedicated to comparing features
*
* @return The operator slot to use
@@ -187,7 +188,7 @@ namespace node_value_arrs
* @brief Get the parameter operator slot associated with a given rung/offset
*
* @param rung Rung of the feature
- * @param offset Offset used to prevent overwrites
+ * @param offset Offset used to prevent overwrites (determines where in the binary expression tree this operator is)
* @param for_comp If true get a slot dedicated to comparing features
*
* @return The operator slot to use
@@ -232,7 +233,7 @@ namespace node_value_arrs
*
* @param ind The Node's arr_ind
* @param rung Rung of the feature
- * @param offset Offset used to prevent overwrites
+ * @param offset Offset used to prevent overwrites (determines where in the binary expression tree this operator is)
* @param for_comp If true get a slot dedicated to comparing features
*
* @return The register element for a given feature index and offset
@@ -251,7 +252,7 @@ namespace node_value_arrs
*
* @param ind The Node's arr_ind
* @param rung Rung of the feature
- * @param offset Offset used to prevent overwrites
+ * @param offset Offset used to prevent overwrites (determines where in the binary expression tree this operator is)
* @param for_comp If true get a slot dedicated to comparing features
*
* @return The register element for a given feature index and offset
@@ -268,7 +269,7 @@ namespace node_value_arrs
/**
* @brief Access element of the permanent training data storage array
*
- * @param feature_ind The _arr_ind of Node to get the training data of
+ * @param feature_ind The _feat_ind of Node to get the training data of
*
* @return pointer to the Node's training data
*/
@@ -277,7 +278,7 @@ namespace node_value_arrs
/**
* @brief Access element of the permanent test data storage array
*
- * @param feature_ind The _arr_ind of Node to get the test data of
+ * @param feature_ind The _feat_ind of Node to get the test data of
*
* @return pointer to the Node's test data
*/
@@ -305,7 +306,7 @@ namespace node_value_arrs
* @brief Access the param storage array
*
* @param rung Rung of the feature
- * @param offset(int) the offset for temporary storage access
+ * @param offset Offset used to prevent overwrites (determines where in the binary expression tree this operator is)
* @param for_comp If true get a slot dedicated to comparing features
* @return pointer to the correct element of the PARAM_STORAGE_ARR
*/
@@ -324,7 +325,7 @@ namespace node_value_arrs
* @brief Access the param storage array for the test set
*
* @param rung Rung of the feature
- * @param offset(int) the offset for temporary storage access
+ * @param offset Offset used to prevent overwrites (determines where in the binary expression tree this operator is)
* @param for_comp If true get a slot dedicated to comparing features
* @return pointer to the correct element of the PARAM_STORAGE_ARR
*/
@@ -345,7 +346,7 @@ namespace node_value_arrs
* @param arr_ind Nodes _arr_ind
* @param feat_ind Nodes _feat_ind
* @param rung Rung of the feature
- * @param offset(int) the offset for temporary storage access
+ * @param offset Offset used to prevent overwrites (determines where in the binary expression tree this operator is)
* @param for_comp If true get a slot dedicated to comparing features
*
* @return The value pointer
@@ -364,7 +365,7 @@ namespace node_value_arrs
* @param arr_ind Nodes _arr_ind
* @param feat_ind Nodes _feat_ind
* @param rung Rung of the feature
- * @param offset(int) the offset for temporary storage access
+ * @param offset Offset used to prevent overwrites (determines where in the binary expression tree this operator is)
* @param for_comp If true get a slot dedicated to comparing features
*
* @return The value pointer
@@ -413,10 +414,10 @@ namespace node_value_arrs
* @brief Initialize the node value arrays
* @details Using the size of the initial feature space constructor the storage arrays
*
- * @param task_sz_train Number of training samples per task
- * @param task_sz_test Number of test sample per task
- * @param n_primary_feat Number of primary features
- * @param max_rung Largest rung of a feature
+ * @param task_sz_train (list) The number of training samples per task
+ * @param task_sz_test (list) The number of test sample per task
+ * @param n_primary_feat (int) The number of primary features
+ * @param max_rung (int) The maximum rung for all features
*/
inline void initialize_values_arr(
py::list task_sz_train,
@@ -439,10 +440,10 @@ namespace node_value_arrs
* @brief Initialize the node value arrays
* @details Using the size of the initial feature space constructor the storage arrays
*
- * @param task_sz_train Number of training samples per task
- * @param task_sz_test Number of test sample per task
- * @param n_primary_feat Number of primary features
- * @param max_rung Largest rung of a feature
+ * @param task_sz_train (np.ndarray) The number of training samples per task
+ * @param task_sz_test (np.ndarray) The number of test sample per task
+ * @param n_primary_feat (int) The number of primary features
+ * @param max_rung (int) The maximum rung for all features
*/
inline void initialize_values_arr(
np::ndarray task_sz_train,
diff --git a/src/feature_creation/units/Unit.hpp b/src/feature_creation/units/Unit.hpp
index 1f5459b6d3695f96b0ba85df8a48861e7ba168e9..bae27a85893791b2e7046d25cedeca72c69b189d 100644
--- a/src/feature_creation/units/Unit.hpp
+++ b/src/feature_creation/units/Unit.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/unit/Unit.hpp
- * @brief A representation of the units for each feature
+ * @brief A representation of the units for each feature
*
* A dictionary representation of the unit for a features (key = unit, value = power)
*
@@ -51,7 +51,7 @@ private:
ar & _dct;
}
protected:
- std::map<std::string, double> _dct; //!< the dictionary describing the units <unit string, unit power>
+ std::map<std::string, double> _dct; //!< The dictionary representation of the unit <unit string, unit power>
public:
// DocString: unit_base_const
@@ -71,7 +71,7 @@ public:
// DocString: unit_init_str
/**
- * @brief Construct the unit based on a string
+ * @brief Construct the unit based on a string representation of the unit
* @details Take a string representation of a unit, and build the Unit
*
* @param unit_str The string to build the unit
diff --git a/src/inputs/InputParser.cpp b/src/inputs/InputParser.cpp
index 831074639186f108e4f2c7bff0ebf4e8d5cea064..4d894742f81cbe54271b6162ae0aafe379421542 100644
--- a/src/inputs/InputParser.cpp
+++ b/src/inputs/InputParser.cpp
@@ -28,7 +28,7 @@ InputParser::InputParser(pt::ptree IP, std::string fn, std::shared_ptr<MPI_Inter
_n_dim(IP.get<int>("desc_dim")),
_n_sis_select(IP.get<int>("n_sis_select")),
_max_rung(IP.get<int>("max_rung")),
- _max_store_rung(IP.get<int>("n_rung_store", _max_rung - 1)),
+ _n_rung_store(IP.get<int>("n_rung_store", _max_rung - 1)),
_n_rung_generate(IP.get<int>("n_rung_generate", 0)),
_n_samp(0),
_n_residuals(IP.get<int>("n_residual", 1)),
@@ -254,7 +254,7 @@ void InputParser::generate_feature_space(
int cur_line = 0;
int n_train_samp = 0;
- int n_test_samp = 0;
+ int n_samp_test = 0;
while (std::getline(data_stream, line))
{
@@ -273,7 +273,7 @@ void InputParser::generate_feature_space(
// This data point is in the validation dataset
// Find where the row should be in the data structure
- n_test_samp = 0;
+ n_samp_test = 0;
for(auto& task : tasks)
{
int task_ind = std::find(task.second.begin(), task.second.end(), cur_line) - task.second.begin();
@@ -281,7 +281,7 @@ void InputParser::generate_feature_space(
{
if(std::any_of(_leave_out_inds.begin(), _leave_out_inds.end(), [&ii, &task](int ind){return ind == task.second[ii];}))
{
- ++n_test_samp;
+ ++n_samp_test;
}
}
@@ -296,7 +296,7 @@ void InputParser::generate_feature_space(
{
if(ii - 1 != taskind)
{
- test_data[ii - 1][n_test_samp] = std::stod(split_line[ii]);
+ test_data[ii - 1][n_samp_test] = std::stod(split_line[ii]);
}
}
}
@@ -415,7 +415,7 @@ void InputParser::generate_feature_space(
_calc_type,
_max_rung,
_n_sis_select,
- _max_store_rung,
+ _n_rung_store,
_n_rung_generate,
_cross_cor_max,
_l_bound,
@@ -433,7 +433,7 @@ void InputParser::generate_feature_space(
_calc_type,
_max_rung,
_n_sis_select,
- _max_store_rung,
+ _n_rung_store,
_n_rung_generate,
_cross_cor_max,
_l_bound,
diff --git a/src/inputs/InputParser.hpp b/src/inputs/InputParser.hpp
index b49901465e7f9b98833aa3574e46e0c7aca3d29c..f24f08083e9afc23165bad722ef2f2e995d75fe0 100644
--- a/src/inputs/InputParser.hpp
+++ b/src/inputs/InputParser.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file inputs/InputParser.hpp
- * @brief Define the class that parses the input file
+ * @brief Define the class that parses the input file
*
* This class parses and stores the parameters in the input file
*
@@ -46,46 +46,47 @@ class InputParser
{
public:
- const std::vector<std::string> _param_opset; //!< Map of parameterization operator set (set of operators and non-linear parameters used for a non-linear least squares fit to property)
- std::vector<std::string> _opset; //!< List of all operators for feature combination
- std::vector<double> _prop_train; //!< The values of the property in the training set
- std::vector<double> _prop_test; //!< The values of the property in the test set
+ const std::vector<std::string> _param_opset; //!< Vector containing all allowed operators strings for operators with free parameters
+ std::vector<std::string> _opset; //!< Vector containing all allowed operators strings
+ std::vector<double> _prop_train; //!< The value of the property to evaluate the loss function against for the training set
+ std::vector<double> _prop_test; //!< The value of the property to evaluate the loss function against for the test set
- std::vector<int> _leave_out_inds; //!< Indexes from the initial data file to use as a test set
- std::vector<int> _task_sizes_train; //!< Number of samples in each task's training data
- std::vector<int> _task_sizes_test; //!< //!< Number of samples in each task's test data
+ std::vector<int> _leave_out_inds; //!< List of indexes from the initial data file in the test set
+ std::vector<int> _task_sizes_train; //!< Number of training samples per task
+ std::vector<int> _task_sizes_test; //!< Number of testing samples per task
- Unit _prop_unit; //!< The Unit for the property
+ Unit _prop_unit; //!< The Unit of the property
const std::string _filename; //!< Name of the input file
const std::string _data_file; //!< Name of the data file
const std::string _prop_key; //!< Key used to find the property column in the data file
- std::string _prop_label; //!< label used to describe the property
+ std::string _prop_label; //!< The label of the property
const std::string _task_key; //!< Key used to find the task column in the data file
- const std::string _calc_type; //!< Type of projection operator to use
+ const std::string _calc_type; //!< The type of LossFunction to use when projecting the features onto a property
std::shared_ptr<FeatureSpace> _feat_space; //!< shared_ptr to the FeatureSpace generated from the data file and the input file
double _cross_cor_max; //!< Maximum cross-correlation used for selecting features
- double _l_bound; //!< Minimum absolute value allowed for the feature.
- double _u_bound; //!< Maximum absolute value allowed for the feature.
+ double _l_bound; //!< The lower bound for the maximum absolute value of the features
+ double _u_bound; //!< The upper bound for the maximum absolute value of the features
- const int _n_dim; //!< Number of dimensions to calculate
+ const int _n_dim; //!< The maximum number of features allowed in the linear model
const int _max_rung; //!< Maximum rung for the feature creation
- const int _max_store_rung; //!< Total rungs stored in memory
- const int _n_rung_generate; //!< Total number of rungs to generate on the fly
- const int _n_sis_select; //!< //!< Number of features to select for each dimensions
- int _n_samp; //!< //!< Number of samples (training data)
+ const int _n_rung_store; //!< The number of rungs to calculate and store the value of the features for all samples
+ const int _n_rung_generate; //!< Either 0 or 1, and is the number of rungs to generate on the fly during SIS
+ const int _n_sis_select; //!< Number of features to select during each SIS iteration
+ int _n_samp; //!< Number of samples in the training set
const int _n_residuals; //!< Number of residuals to pass to the next sis model
- const int _n_models_store; //!< Number of models to store
- const int _max_param_depth; //!< Max depth to parameterize a feature (default=_max_rung)
+ const int _n_models_store; //!< The number of models to output to files
+ const int _max_param_depth; //!< The maximum depth in the binary expression tree to set non-linear optimization
const int _nlopt_seed; //!< The seed used for the nlOpt library
- const bool _fix_intercept; //!< If true force intercept to be 0.0
+ const bool _fix_intercept; //!< If true the bias term is fixed at 0
const bool _global_param_opt; //!< True if global optimization is requested for non-linear optimization of parameters (Can break reproducibility)
- const bool _reparam_residual; //!< True if reparameterize features using the residuals
+ const bool _reparam_residual; //!< If True then reparameterize features using the residuals of each model
+
/**
- * @brief Constructor
+ * @brief Constructor of the InputParser
*
* @param IP Property tree generated from json file
* @param fn filename for the input file
@@ -94,13 +95,12 @@ public:
InputParser(pt::ptree IP, std::string fn, std::shared_ptr<MPI_Interface> comm);
/**
- * @brief Accessor function to _feat_space
+ * @brief The generated FeatureSpace
*/
inline std::shared_ptr<FeatureSpace> feat_space() const {return _feat_space;}
/**
* @brief Generate the feature space from input files and parameters
- * @details [long description]
*
* @param comm MPI communicator for the calculation
* @param headers column headers for all columns in the data file (expr of the Nodes)
@@ -112,8 +112,7 @@ public:
std::shared_ptr<MPI_Interface> comm,
std::vector<std::string> headers,
std::vector<Unit> units,
- std::map<std::string,
- std::vector<int>> tasks,
+ std::map<std::string, std::vector<int>> tasks,
int taskind
);
};
diff --git a/src/loss_function/LossFunction.hpp b/src/loss_function/LossFunction.hpp
index f2d67918b2893cc1e60433b7dcaa89452829c9c9..13f0a4034a1d25cb4d06b4c5812b94a6287d9cd1 100644
--- a/src/loss_function/LossFunction.hpp
+++ b/src/loss_function/LossFunction.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file loss_function/LossFunction.hpp
- * @brief A class used to calculate the projection score and l0-regularization error
+ * @brief A class used to calculate the projection score and l0-regularization error
*
* @author Thomas A. R. Purcell (tpurcell)
* @bug No known bugs.
@@ -36,7 +36,7 @@ protected:
std::vector<double> _error_train; //!< The error vector for the training set
std::vector<double> _error_test; //!< The error vector for the test set
- std::vector<double> _coefs; //!< Coefficients for the model
+ std::vector<double> _coefs; //!< Coefficients for the linear model
const std::vector<int> _task_sizes_train; //!< Number of training samples per task
const std::vector<int> _task_sizes_test; //!< Number of testing samples per task
@@ -44,10 +44,10 @@ protected:
const int _n_samp; //!< Number of samples in the training set
const int _n_samp_test; //!< Number of samples in the test set
int _n_project_prop; //!< Number of properties to project over
- int _n_feat;
- int _n_dim;
+ int _n_feat; //!< Number features in the linear model
+ int _n_dim; //!< Total number of constants to fit (scale and bias terms)
const int _n_task; //!< Number of tasks
- const bool _fix_intercept;
+ const bool _fix_intercept; //!< If true then the bias term is fixed at 0
public:
/**
@@ -142,50 +142,161 @@ public:
*/
virtual void reset_projection_prop(const std::vector<std::vector<model_node_ptr>>& models) = 0;
+ /**
+ * @brief The type of LossFunction
+ */
virtual LOSS_TYPE type() const = 0;
+
+ /**
+ * @brief Coefficients for the linear model
+ */
std::vector<double> coefs() const{return _coefs;};
+ /**
+ * @brief Number of training samples per task
+ */
inline std::vector<int> task_sizes_train() const {return _task_sizes_train;}
+
+ /**
+ * @brief Number of testing samples per task
+ */
inline std::vector<int> task_sizes_test() const {return _task_sizes_test;}
+ /**
+ * @brief The value of the property to evaluate the loss function against for the training set
+ */
inline const std::vector<double>& prop_train() const {return _prop_train;}
+
+ /**
+ * @brief The value of the property to evaluate the loss function against for the test set
+ */
inline const std::vector<double>& prop_test() const {return _prop_test;}
+ /**
+ * @brief The estimated value of the property to evaluate the loss function against for the training set
+ */
inline const std::vector<double>& prop_train_est() const {return _prop_train_est;}
+
+ /**
+ * @brief The estimated value of the property to evaluate the loss function against for the test set
+ */
inline const std::vector<double>& prop_test_est() const {return _prop_test_est;}
+ /**
+ * @brief The property vector used for projection
+ */
inline const std::vector<double>& prop_project() const {return _projection_prop;}
+ /**
+ * @brief Copy of the value of the property to evaluate the loss function against for the training set
+ */
inline const std::vector<double> prop_train_copy() const {return _prop_train;}
+
+ /**
+ * @brief Copy of the value of the property to evaluate the loss function against for the test set
+ */
inline const std::vector<double> prop_test_copy() const {return _prop_test;}
+ /**
+ * @brief Copy of the estimated value of the property to evaluate the loss function against for the training set
+ */
inline const std::vector<double> prop_train_est_copy() const {return _prop_train_est;}
+
+ /**
+ * @brief Copy of the estimated value of the property to evaluate the loss function against for the test set
+ */
inline const std::vector<double> prop_test_est_copy() const {return _prop_test_est;}
+ /**
+ * @brief Copy of the property vector used for projection
+ */
inline const std::vector<double> prop_project_copy() const {return _projection_prop;}
+ /**
+ * @brief Pointer to the head of the value of the property to evaluate the loss function against for the training set
+ */
inline const double* prop_pointer() const {return _prop_train.data();}
+
+ /**
+ * @brief Pointer to the head of the value of the property to evaluate the loss function against for the test set
+ */
inline const double* prop_test_pointer() const {return _prop_test.data();}
+ /**
+ * @brief Pointer to the head of the estimated value of the property to evaluate the loss function against for the training set
+ */
inline const double* prop_est_pointer() const {return _prop_train_est.data();}
+
+ /**
+ * @brief Pointer to the head of the estimated value of the property to evaluate the loss function against for the test set
+ */
inline const double* prop_test_est_pointer() const {return _prop_test_est.data();}
+ /**
+ * @brief Pointer to the head of the property vector used for projection
+ */
inline const double* prop_project_pointer() const {return _projection_prop.data();}
+ /**
+ * @brief The error vector for the training set
+ */
inline const std::vector<double>& error_train() const {return _error_train;}
+
+ /**
+ * @brief The error vector for the test set
+ */
inline const std::vector<double>& error_test() const {return _error_test;}
+ /**
+ * @brief Copy of the error vector for the training set
+ */
inline const std::vector<double> error_train_copy() const {return _error_train;}
+
+ /**
+ * @brief Copy of the error vector for the test set
+ */
inline const std::vector<double> error_test_copy() const {return _error_test;}
+ /**
+ * @brief If true, then the bias term is fixed at 0
+ */
inline bool fix_intercept() const {return _fix_intercept;}
+
+ /**
+ * @brief Number features in the linear model
+ */
inline int n_feat() const {return _n_feat;}
+
+ /**
+ * @brief Total number of constants to fit (scale and bias terms)
+ */
inline int n_dim() const {return _n_dim;}
+
+ /**
+ * @brief Number of samples in the training setCoefficients for the linear model
+ */
inline int n_samp() const {return _n_samp;}
+
+ /**
+ * @brief Number of samples in the test setCoefficients for the linear model
+ */
inline int n_samp_test() const {return _n_samp_test;}
+
+ /**
+ * @brief Number of tasks
+ */
inline int n_task() const {return _n_task;}
+
+ /**
+ * @brief Set the number of features and linear model constants to those for the new n_feat
+ *
+ * @param n_feat The updated number of features for the model
+ */
virtual inline void set_nfeat(int n_feat){_n_feat = n_feat; _n_dim = n_feat + (!_fix_intercept);}
+ /**
+ * @brief The number of classes in the calculation
+ */
virtual inline int n_class(){return 0;}
};
diff --git a/src/loss_function/LossFunctionConvexHull.cpp b/src/loss_function/LossFunctionConvexHull.cpp
index 432b475e9a3692fa746035e3400232a3c1618f53..12a7b743ce5516a01b26336cc95e226ddf2d358a 100644
--- a/src/loss_function/LossFunctionConvexHull.cpp
+++ b/src/loss_function/LossFunctionConvexHull.cpp
@@ -160,7 +160,7 @@ void LossFunctionConvexHull::setup_lp()
std::copy_n(n_samp_per_class.begin() + tt * _n_class, _n_class, samp_per_class.begin());
task_start += _task_sizes_train[tt];
- std::vector<int> n_test_samp_per_class(n_samp_per_class.size(), 0);
+ std::vector<int> n_samp_test_per_class(n_samp_per_class.size(), 0);
if(_task_sizes_test[tt] > 0)
{
inds.resize(_task_sizes_test[tt]);
@@ -174,13 +174,13 @@ void LossFunctionConvexHull::setup_lp()
_test_sample_inds_to_sorted_dmat_inds[inds[ii]] = ii + task_start_test;
if(_prop_test[inds[ii]] != _prop_test[inds[ii - 1]])
{
- n_test_samp_per_class[
+ n_samp_test_per_class[
rep_class[_prop_test[inds[ii - 1]]]
] = ii - cls_start;
cls_start = ii;
}
}
- n_test_samp_per_class[rep_class[_prop_test[inds.back()]]] = inds.size() - cls_start;
+ n_samp_test_per_class[rep_class[_prop_test[inds.back()]]] = inds.size() - cls_start;
task_start_test += _task_sizes_test[tt];
}
@@ -191,8 +191,8 @@ void LossFunctionConvexHull::setup_lp()
_n_feat,
std::accumulate(n_samp_per_class.begin(), n_samp_per_class.end(), 0),
_width,
- n_test_samp_per_class,
- std::accumulate(n_test_samp_per_class.begin(), n_test_samp_per_class.end(), 0)
+ n_samp_test_per_class,
+ std::accumulate(n_samp_test_per_class.begin(), n_samp_test_per_class.end(), 0)
);
_lp.push_back(lp);
}
diff --git a/src/loss_function/LossFunctionConvexHull.hpp b/src/loss_function/LossFunctionConvexHull.hpp
index 84932a98ae57a53d6df8c6855e21e20b41d95035..061acfe609f21b0da53fa39beca9d7d4e7b5a871 100644
--- a/src/loss_function/LossFunctionConvexHull.hpp
+++ b/src/loss_function/LossFunctionConvexHull.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file loss_function/LossFunctionConvexHull.hpp
- * @brief A class used to calculate the projection score and l0-regularization error
+ * @brief A class used to calculate the projection score and l0-regularization error
*
* @author Thomas A. R. Purcell (tpurcell)
* @bug No known bugs.
@@ -32,14 +32,14 @@
class LossFunctionConvexHull: public LossFunction
{
protected:
- std::vector<ConvexHull1D> _convex_hull;
- std::vector<LPWrapper> _lp;
- std::vector<double> _scores;
+ std::vector<ConvexHull1D> _convex_hull; //!< Vector to the convex hull objects for projection
+ std::vector<LPWrapper> _lp; //!< Vector for calculating the LP-optimization
+ std::vector<double> _scores; //!< The scores for each of the projection properties
std::map<int, int> _sample_inds_to_sorted_dmat_inds; //!< map from input sample inds to the SORTED_D_MATRIX_INDS
std::map<int, int> _test_sample_inds_to_sorted_dmat_inds; //!< map from input sample inds to the SORTED_D_MATRIX_INDS
- const double _width;
- int _n_class;
+ const double _width; //!< The width used as the tolerance for the LP optimization
+ int _n_class; //!< Number of classes in the property
public:
@@ -90,8 +90,7 @@ public:
LossFunctionConvexHull& operator= (LossFunctionConvexHull&& o) = default;
/**
- * @brief Setup the _lp objects
- * @details This is used for calculating the LossFunction
+ * @brief Initialize the LPWrappers used for calculating the loss function
*/
void setup_lp();
@@ -133,17 +132,27 @@ public:
virtual double test_loss(const std::vector<model_node_ptr>& feats);
/**
- * @brief Set the model list from the previous generation
- * @details [long description]
+ * @brief Reset the the property used for projection
*
- * @param model_list [description]
+ * @param model_list The models used to update the projection property
*/
void reset_projection_prop(const std::vector<std::vector<model_node_ptr>>& models);
+ /**
+ * @brief The type of LossFunction
+ */
inline LOSS_TYPE type() const {return LOSS_TYPE::CONVEX_HULL;}
+ /**
+ * @brief Set the number of features and linear model constants to those for the new n_feat
+ *
+ * @param n_feat The updated number of features for the model
+ */
void set_nfeat(int n_feat);
+ /**
+ * @brief The number of classes in the calculation
+ */
inline int n_class(){return _n_class;}
};
diff --git a/src/loss_function/LossFunctionLogPearsonRMSE.hpp b/src/loss_function/LossFunctionLogPearsonRMSE.hpp
index c5f111c85e1ec066160f3cc96a215856c42ac695..1d1d6f4cf59d1bba4dc97d04fed5f6d397d92bce 100644
--- a/src/loss_function/LossFunctionLogPearsonRMSE.hpp
+++ b/src/loss_function/LossFunctionLogPearsonRMSE.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file loss_function/LossFunctionLogPearsonRMSE.hpp
- * @brief A class used to calculate the projection score and l0-regularization error
+ * @brief A class used to calculate the projection score and l0-regularization error
*
* @author Thomas A. R. Purcell (tpurcell)
* @bug No known bugs.
@@ -26,7 +26,7 @@
class LossFunctionLogPearsonRMSE: public LossFunctionPearsonRMSE
{
protected:
- std::vector<double> _log_feat;
+ std::vector<double> _log_feat; //!< A feature to store the log of the Features' data
public:
@@ -86,43 +86,53 @@ public:
double project(const node_ptr& feat);
/**
- * @brief Set the _a matrix
+ * @brief Set the A matrix used for solving the least squares regression
*
- * @param inds Indexes to access the central storage area
+ * @param inds A vector storing the _d_mat_ind of the features to pull data from the Descriptor Matrix storage array
+ * @param taskind The task for used for the least squares regression
+ * @param start The offset needed from the head of the feature's training data to where the task starts
*/
void set_a(const std::vector<int>& inds, int taskind, int start);
/**
- * @brief Set the _a matrix
+ * @brief Set the A matrix used for solving the least squares regression
*
* @param feats The features used to evaluate the loss function
+ * @param taskind The task for used for the least squares regression
+ * @param start The offset needed from the head of the feature's training data to where the task starts
*/
void set_a(const std::vector<model_node_ptr>& feats, int taskind, int start);
/**
- * @brief Set the error and return the RMSE
+ * @brief Set the error vector
*
* @param inds Indexes to access the central storage area
- * @return The RMSE of the model
+ * @param taskind The task for used for the least squares regression
+ * @param start The offset needed from the head of the feature's training data to where the task starts
*/
void set_prop_train_est(const std::vector<int>& inds, int taskind, int start);
/**
- * @brief Set the error and return the RMSE
+ * @brief Set the error
*
- * @param feats The features used to evaluate the loss function for the test set
- * @return The test RMSE of the feature
+ * @param feats The features used to evaluate the loss function
+ * @param taskind The task for used for the least squares regression
+ * @param start The offset needed from the head of the feature's training data to where the task starts
*/
void set_prop_train_est(const std::vector<model_node_ptr>& feats, int taskind, int start);
/**
- * @brief Set the error and return the RMSE
+ * @brief Set the test error and return the RMSE
*
- * @param feats The features used to evaluate the loss function
- * @return The RMSE of the model
+ * @param feats The features used to evaluate the loss function for the test data
+ * @param taskind The task for used for the least squares regression
+ * @param start The offset needed from the head of the feature's test data to where the task starts
*/
void set_prop_test_est(const std::vector<model_node_ptr>& feats, int taskind, int start);
+ /**
+ * @brief The type of LossFunction
+ */
inline LOSS_TYPE type() const {return LOSS_TYPE::LOG_PEARSON_RMSE;}
};
diff --git a/src/loss_function/LossFunctionPearsonRMSE.hpp b/src/loss_function/LossFunctionPearsonRMSE.hpp
index fb5340f3620302238cc86501e368e47f429640f4..dc35023ca27d70492aa33ec635bcf6c6c8cf242b 100644
--- a/src/loss_function/LossFunctionPearsonRMSE.hpp
+++ b/src/loss_function/LossFunctionPearsonRMSE.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file loss_function/LossFunctionPearsonRMSE.hpp
- * @brief A class used to calculate the projection score and l0-regularization error
+ * @brief A class used to calculate the projection score and l0-regularization error
*
* @author Thomas A. R. Purcell (tpurcell)
* @bug No known bugs.
@@ -134,59 +134,76 @@ public:
void set_opt_lwork();
/**
- * @brief Set the _a matrix
+ * @brief Set the A matrix used for solving the least squares regression
*
- * @param inds Indexes to access the central storage area
+ * @param inds A vector storing the _d_mat_ind of the features to pull data from the Descriptor Matrix storage array
+ * @param taskind The task for used for the least squares regression
+ * @param start The offset needed from the head of the feature's training data to where the task starts
*/
virtual void set_a(const std::vector<int>& inds, int taskind, int start);
/**
- * @brief Set the _a matrix
+ * @brief Set the A matrix used for solving the least squares regression
*
* @param feats The features used to evaluate the loss function
+ * @param taskind The task for used for the least squares regression
+ * @param start The offset needed from the head of the feature's training data to where the task starts
*/
virtual void set_a(const std::vector<model_node_ptr>& feats, int taskind, int start);
/**
- * @brief Set the error and return the RMSE
+ * @brief Set the error vector
*
* @param inds Indexes to access the central storage area
- * @return The RMSE of the model
+ * @param taskind The task for used for the least squares regression
+ * @param start The offset needed from the head of the feature's training data to where the task starts
*/
virtual void set_prop_train_est(const std::vector<int>& inds, int taskind, int start);
/**
- * @brief Set the error and return the RMSE
+ * @brief Set the error
*
* @param feats The features used to evaluate the loss function
- * @return The RMSE of the model
+ * @param taskind The task for used for the least squares regression
+ * @param start The offset needed from the head of the feature's training data to where the task starts
*/
virtual void set_prop_train_est(const std::vector<model_node_ptr>& feats, int taskind, int start);
- /**
- * @brief Set the error and return the RMSE
+ /**
+ * @brief Set the test error and return the RMSE
*
- * @param feats The features used to evaluate the loss function for the test set
- * @return The test RMSE of the feature
+ * @param feats The features used to evaluate the loss function for the test data
+ * @param taskind The task for used for the least squares regression
+ * @param start The offset needed from the head of the feature's test data to where the task starts
*/
virtual void set_prop_test_est(const std::vector<model_node_ptr>& feats, int taskind, int start);
/**
* @brief Perform least squares regression
*
+ * @param taskind The task for used for the least squares regression
+ * @param start The offset needed from the head of the feature's test data to where the task starts
* @return info The final info value from dgels
*/
int least_squares(int taskind, int start);
/**
- * @brief Set the model list from the previous generation
- * @details [long description]
+ * @brief Reset the the property used for projection
*
- * @param model_list [description]
+ * @param model_list The models used to update the projection property
*/
void reset_projection_prop(const std::vector<std::vector<model_node_ptr>>& models);
+ /**
+ * @brief The type of LossFunction
+ */
virtual inline LOSS_TYPE type() const {return LOSS_TYPE::PEARSON_RMSE;}
+
+ /**
+ * @brief Set the number of features and linear model constants to those for the new n_feat
+ *
+ * @param n_feat The updated number of features for the model
+ */
void set_nfeat(int n_feat);
};
diff --git a/src/loss_function/utils.hpp b/src/loss_function/utils.hpp
index 3359719e47d83bc3daab42ec6cb1890b01e31bdb..80da1d57edf93be100e426b6ab7506878b25a1fd 100644
--- a/src/loss_function/utils.hpp
+++ b/src/loss_function/utils.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file loss_function/utils.hpp
- * @brief utilities for generating loss_functions
+ * @brief utilities for generating loss_functions
*
* @author Thomas A. R. Purcell (tpurcell)
* @bug No known bugs.
@@ -27,6 +27,17 @@
namespace loss_function_util
{
+/**
+ * Create a shared_ptr to a new LossFunction
+ *
+ * @param type The type of LossFunction to create
+ * @param prop_train The value of the property to evaluate the loss function against for the training set
+ * @param prop_test The value of the property to evaluate the loss function against for the test set
+ * @param task_sizes_train Number of training samples per task
+ * @param task_sizes_test Number of testing samples per task
+ * @param fix_intercept If true then the bias term is fixed at 0
+ * @return A shared_ptr to the newly created LossFunction
+ */
std::shared_ptr<LossFunction> get_loss_function(
std::string type,
std::vector<double> prop_train,
@@ -36,6 +47,12 @@ std::shared_ptr<LossFunction> get_loss_function(
bool fix_intercept = false
);
+/**
+ * @brief Make a copy of a LossFunction and return a shared_ptr to it
+ *
+ * @param loss A shared_ptr to the LossFunction to be copied
+ * @return The shared_ptr to the copied LossFunction
+ */
std::shared_ptr<LossFunction> copy(std::shared_ptr<LossFunction> loss);
};
diff --git a/src/mpi_interface/MPI_Interface.hpp b/src/mpi_interface/MPI_Interface.hpp
index c7f8f2a852616e2bf58192f529632b856b4c6c7c..9f607f1a7eaec332a3947c06091fe8ca29d42f89 100644
--- a/src/mpi_interface/MPI_Interface.hpp
+++ b/src/mpi_interface/MPI_Interface.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file mpi_interface/MPI_Interface.hpp
- * @brief Define MPI interface for the calculation
+ * @brief Define MPI interface for the calculation
*
* This class modifies the boost mpi::communicator for easier communication
*
@@ -27,7 +27,6 @@
namespace mpi = boost::mpi;
-// Augment the boost mpi communicator class with some other useful data
/**
* @brief Augment the boost MPI communicator class with some other useful data
* @details MPI communicator used to transfer data throughout the cell.
diff --git a/src/mpi_interface/MPI_ops.hpp b/src/mpi_interface/MPI_ops.hpp
index 5144ff184be30e3ddfcd9f6c26ff5320366d129a..d5fd50b4a44c4321507150776e40c428f21e0039 100644
--- a/src/mpi_interface/MPI_ops.hpp
+++ b/src/mpi_interface/MPI_ops.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file mpi_interface/MPI_Ops.hpp
- * @brief Define MPI reduce all operator to facilitate a distributed sorting algorithm
+ * @brief Define MPI reduce all operator to facilitate a distributed sorting algorithm
*
* This is based off of a project from Meisam Farzalipour Tabriz at the MPCDF (mpi_topk)
*
diff --git a/src/nl_opt/NLOptWrapper.hpp b/src/nl_opt/NLOptWrapper.hpp
index ad7738a7a8604b876b54913b119016e0c42f8de1..8a99a773933e2de00cefe2d2bd7cd4e5b495f4e7 100644
--- a/src/nl_opt/NLOptWrapper.hpp
+++ b/src/nl_opt/NLOptWrapper.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file feature_creation/parameterization/NLOptWrapper.hpp
- * @brief Wrapper and functions to use libnlopt to find parameterizations for features
+ * @brief Wrapper and functions to use libnlopt to find parameterizations for features
*
* @author Thomas A. R. Purcell (tpurcell)
* @bug No known bugs.
@@ -29,24 +29,24 @@ class NLOptimizer
protected:
const nlopt::func _objective; //!< Objective function to use
- std::vector<double> _a; //!< vector to store the A matrix for dgels
- std::vector<double> _params; //!< The working parameters vector
- std::vector<double> _lb_init; //!< The lower bound for the global optimization
- std::vector<double> _ub_init; //!< The upper bound for the global optimization
- std::vector<double> _lb_final; //!< The lower bound for the local optimization
- std::vector<double> _ub_final; //!< The upper bound for the local optimization
- const std::vector<double> _prop; //!< The property to fit the functions against
- std::vector<double> _prop_copy; //!< Copy of the property to keep for dgels
- std::vector<double> _work; //!< work array for dgels
- const std::vector<double> _zeros; //!< array of zeros to fill parameters
- const std::vector<int> _task_sizes; //!< number of samples in each task
+ std::vector<double> _a; //!< The A matrix used for calculating the least squares solution of the problem
+ std::vector<double> _params; //!< The scale and bias terms for the OperatorNode
+ std::vector<double> _lb_init; //!< The lower bound for the scale and bias terms during the global optimization
+ std::vector<double> _ub_init; //!< The upper bound for the scale and bias terms during the global optimization
+ std::vector<double> _lb_final; //!< The lower bound for the scale and bias terms during the local optimization
+ std::vector<double> _ub_final; //!< The upper bound for the scale and bias terms during the local optimization
+ const std::vector<double> _prop; //!< The value of the property to evaluate the loss function against for the training set
+ std::vector<double> _prop_copy; //!< A copy of the value of the property to evaluate the loss function against for the training set
+ std::vector<double> _work; //!< Work vector for dgels
+ const std::vector<double> _zeros; //!< A vector filled with zeros to fill parameters
+ const std::vector<int> _task_sizes; //!< Number of training samples per task
std::vector<bool> _free_param; //!< Elements are true if parameter is free
- const int _n_samp; //!< total number of samples
+ const int _n_samp; //!< Total number of samples
const int _n_rung; //!< Maximum rung of the features
const int _n_prop; //!< Number of property vectors to optimize against
int _max_params; //!< Maximum number of possible parameters
- int _max_param_depth; //!< parameterize features to all depths of the tree
+ int _max_param_depth; //!< The maximum depth in the binary expression tree to set non-linear optimization
nlopt::algorithm _local_opt_alg; //!< Algorithm used for local optimization
@@ -54,12 +54,13 @@ public:
/**
* @brief Constructor
*
- * @param task_sizes number of samples in each task
- * @param prop The property to fit the functions against
+ * @param task_sizes Number of training samples per task
+ * @param prop The value of the property to evaluate the loss function against for the training set
* @param n_rung Maximum rung of the features
- * @param objective The objective function to use
- * @param max_param_depth maximum depth of the binary expression tress to parameterize from the root
+ * @param objective The objective function to use for the optimization
+ * @param max_param_depth The maximum depth in the binary expression tree to set non-linear optimization
* @param reset_max_param_depth If true reset the maximum parameter depth
+ * @param local_opt_alg Algorithm used for local optimization
*/
NLOptimizer(
const std::vector<int>& task_sizes,
@@ -75,54 +76,52 @@ public:
/**
* @brief uses nlopt to optimize the parameters of a feature
*
- * @param feat pointer to the node to optimize
- * @param use_grad If true use a gradient based optimizer (LD_VAR2) for local optimization
- * Otherwise use a simplex based approach (LN_SBPLX) if possible
+ * @param feat Pointer to the node to optimize
*/
double optimize_feature_params(Node* feat);
/**
- * @brief Accessor function to the task sizes
+ * @brief The number of training samples per task
*/
inline const std::vector<int>& task_sizes() const {return _task_sizes;}
/**
- * @brief Accessor function to the the property vector
+ * @brief The value of the property to evaluate the loss function against for the training set
*/
inline const std::vector<double>& prop() const {return _prop;}
/**
- * @brief Accessor function to the pointer of the work vector's start
+ * @brief Work vector for dgels
*/
inline double* work(){return _work.data();}
/**
- * @brief Accessor function to the number of samples
+ * @brief Total number of samples
*/
inline int n_samp() const {return _n_samp;}
/**
- * @brief Accessor function to the maximum number of rungs
+ * @brief Maximum rung of the features
*/
inline int n_rung() const {return _n_rung;}
/**
- * @brief Accessor function to the number of property vectors
+ * @brief Number of property vectors to optimize against
*/
inline int n_prop() const {return _n_prop;}
/**
- * @brief Accessor function to the maximum number of possible parameters
+ * @brief Maximum number of possible parameters
*/
inline int max_params() const {return _max_params;}
/**
- * @brief Accessor function to the maximum parameter depth
+ * @brief The maximum depth in the binary expression tree to set non-linear optimization
*/
inline int max_param_depth() const {return _max_param_depth;}
/**
- * @brief Accessor function to the default local optimization algorithm
+ * @brief Algorithm used for local optimization
*/
inline nlopt::algorithm local_opt_alg () const {return _local_opt_alg;}
@@ -135,19 +134,28 @@ public:
inline bool free_param(int pp){return _free_param[pp];}
/**
- * @brief Accessor function to the convex hull optimizers (if the projection is a classification problem)
+ * @brief Access one of the convex hull objects for projection (If using the Classifier)
+ *
+ * @param ind The index of the _convex_hull vector to access
*/
virtual std::shared_ptr<ConvexHull1D> convex_hull(int ind) const = 0;
+
/**
- * @brief Accessor function to calculate the feature gradient
+ * @brief Access an element of parameter gradient for non-linear optimization
+ *
+ * @param ind The index of the gradient vector to access
*/
virtual double* feature_gradient(int ind) = 0;
+
/**
- * @brief Accessor function to the pointer for the residuals
+ * @brief Access an element of the residuals for non-linear optimization
+ *
+ * @param ind The index of the gradient vector to access
*/
virtual double* residuals(int ind) = 0;
+
/**
- * @brief Accessor function to cauchy scaling factor for regression problems
+ * @brief The Cauchy scaling factor for regression problems
*/
virtual double cauchy_scaling()const = 0;
@@ -160,16 +168,17 @@ public:
class NLOptimizerClassification: public NLOptimizer
{
protected:
- std::vector<std::shared_ptr<ConvexHull1D>> _convex_hull; //!< Object to perform classification
+ std::vector<std::shared_ptr<ConvexHull1D>> _convex_hull; //!< A vector containing the convex hull objects for projection (If using the Classifier)
public:
/**
* @brief Constructor
*
- * @param task_sizes number of samples in each task
- * @param prop The property to fit the functions against
+ * @param task_sizes Number of training samples per task
+ * @param prop The value of the property to evaluate the loss function against for the training set
* @param n_rung Maximum rung of the features
- * @param max_param_depth maximum depth of the binary expression tress to parameterize from the root
+ * @param max_param_depth The maximum depth in the binary expression tree to set non-linear optimization
* @param reset_max_param_depth If true reset the maximum parameter depth
+ * @param local_opt_alg Algorithm used for local optimization
*/
NLOptimizerClassification(
const std::vector<int>& task_sizes,
@@ -179,23 +188,30 @@ public:
bool reset_max_param_depth=false,
const nlopt::algorithm local_opt_alg=nlopt::LN_SBPLX
);
+
/**
- * @brief Accessor function to the convex hull optimizers (if the projection is a classification problem)
+ * @brief Access one of the convex hull objects for projection (If using the Classifier)
+ *
+ * @param ind The index of the _convex_hull vector to access
*/
inline std::shared_ptr<ConvexHull1D> convex_hull(int ind) const {return _convex_hull[ind];}
/**
- * @brief Accessor function to the feature gradient (always nullptr)
+ * @brief Access an element of parameter gradient for non-linear optimization
+ *
+ * @param ind The index of the gradient vector to access
*/
inline double* feature_gradient(int ind){return nullptr;}
/**
- * @brief Accessor function to the residuals (always nullptr)
+ * @brief Access an element of the residuals for non-linear optimization
+ *
+ * @param ind The index of the gradient vector to access
*/
inline double* residuals(int ind){return nullptr;}
/**
- * @brief Accessor function to the cauchy scaling factor (always nullptr)
+ * @brief The Cauchy scaling factor for regression problems
*/
inline double cauchy_scaling() const {return 0.0;}
@@ -212,16 +228,19 @@ protected:
std::vector<double> _residuals; //!< storage space for the residuals
const double _cauchy_scaling; //!< Scaling factor for calculating the cauchy loss function
public:
+
/**
* @brief Constructor
*
- * @param task_sizes number of samples in each task
- * @param prop The property to fit the functions against
+ * @param task_sizes Number of training samples per task
+ * @param prop The value of the property to evaluate the loss function against for the training set
* @param n_rung Maximum rung of the features
- * @param max_param_depth maximum depth of the binary expression tress to parameterize from the root
- * @param cauchy_scaling the Cauchy scale factor used for calculating the residuals
+ * @param objective The objective function to use for the optimization
+ * @param max_param_depth The maximum depth in the binary expression tree to set non-linear optimization
+ * @param cauchy_scaling The Cauchy scale factor used for calculating the residuals
* @param log_reg True if the optimizer is for a log regressor
* @param reset_max_param_depth If true reset the maximum parameter depth
+ * @param local_opt_alg Algorithm used for local optimization
*/
NLOptimizerRegression(
const std::vector<int>& task_sizes,
@@ -235,22 +254,28 @@ public:
);
/**
- * @brief Accessor function to the convex hull optimizers (always nullptr)
+ * @brief Access one of the convex hull objects for projection (If using the Classifier)
+ *
+ * @param ind The index of the _convex_hull vector to access
*/
inline std::shared_ptr<ConvexHull1D> convex_hull(int ind) const{return nullptr;}
/**
- * @brief Accessor function to the feature gradient
+ * @brief Access an element of parameter gradient for non-linear optimization
+ *
+ * @param ind The index of the gradient vector to access
*/
inline double* feature_gradient(int ind){return &_feature_gradient[ind];}
/**
- * @brief Accessor function to the residuals
+ * @brief Access an element of the residuals for non-linear optimization
+ *
+ * @param ind The index of the gradient vector to access
*/
inline double* residuals(int ind){return &_residuals[ind];}
/**
- * @brief Accessor function to the cauchy scaling factor
+ * @brief The Cauchy scaling factor for regression problems
*/
inline double cauchy_scaling() const {return _cauchy_scaling;}
@@ -267,12 +292,13 @@ public:
/**
* @brief Constructor
*
- * @param task_sizes number of samples in each task
- * @param prop The property to fit the functions against
+ * @param task_sizes Number of training samples per task
+ * @param prop The value of the property to evaluate the loss function against for the training set
* @param n_rung Maximum rung of the features
- * @param max_param_depth maximum depth of the binary expression tress to parameterize from the root
- * @param cauchy_scaling the Cauchy scale factor used for calculating the residuals
+ * @param max_param_depth The maximum depth in the binary expression tree to set non-linear optimization
+ * @param cauchy_scaling The Cauchy scale factor used for calculating the residuals
* @param reset_max_param_depth If true reset the maximum parameter depth
+ * @param local_opt_alg Algorithm used for local optimization
*/
NLOptimizerLogRegression(
const std::vector<int>& task_sizes,
@@ -292,13 +318,13 @@ public:
namespace nlopt_wrapper
{
-extern int MAX_PARAM_DEPTH; //!< The maximum parameter depth for the problem
-extern int NLOPT_SEED; //!< The seed used to initialize the nlopt psuedorandom number seed
-extern bool USE_GLOBAL; //!< If true then use global optimization when optimizing features
+extern int MAX_PARAM_DEPTH; //!< The maximum depth in the binary expression tree to set non-linear optimization
+extern int NLOPT_SEED; //!< The seed used for the nlOpt library
+extern bool USE_GLOBAL; //!< True if global optimization is requested for non-linear optimization of parameters (Can break reproducibility)
typedef struct
{
- const double* _prop; //!< pointer to the property vector
+ const double* _prop; //!< A pointer to the value of the property to evaluate the loss function against for the training set
Node* _feat; //!< Node pointer of the feature to parameterize
NLOptimizer* _optimizer; //!< Data structure to store information for the optimization
} feat_data;
@@ -340,12 +366,12 @@ static double objective_log_reg(unsigned int n, const double* p, double* grad, v
* @brief Get an optimizer for the desired task
*
* @param project_type The type of projection operator to optimize the features for
- * @param task_sizes number of samples in each task
- * @param prop The property to fit the functions against
+ * @param task_sizes Number of training samples per task
+ * @param prop The value of the property to evaluate the loss function against for the training set
* @param n_rung Maximum rung of the features
- * @param max_param_depth maximum depth of the binary expression tress to parameterize from the root
- * @param cauchy_scaling scaling factor used for the Cauchy loss function
- * @param reset_max_param_depth If true reset the maximum parameter depth
+ * @param max_param_depth The maximum depth in the binary expression tree to set non-linear optimization
+ * @param cauchy_scaling The Cauchy scale factor used for calculating the residuals
+ * @param max_param_depth The maximum depth in the binary expression tree to set non-linear optimization
*
* @return The correct optimizer
*/
@@ -364,12 +390,12 @@ std::shared_ptr<NLOptimizer> get_optimizer(
/**
* @brief Get an optimizer for the desired task
*
- * @param task_sizes number of samples in each task
- * @param prop The property to fit the functions against
+ * @param task_sizes Number of training samples per task
+ * @param prop The value of the property to evaluate the loss function against for the training set
* @param n_rung Maximum rung of the features
- * @param max_param_depth maximum depth of the binary expression tress to parameterize from the root
- * @param cauchy_scaling scaling factor used for the Cauchy loss function
- * @param use_global If true use global optimization
+ * @param max_param_depth The maximum depth in the binary expression tree to set non-linear optimization
+ * @param cauchy_scaling The Cauchy scale factor used for calculating the residuals
+ * @param use_global True if global optimization is requested for non-linear optimization of parameters (Can break reproducibility)
*
* @return The correct optimizer
*/
@@ -400,12 +426,12 @@ inline NLOptimizerRegression get_reg_optimizer(
/**
* @brief Get an optimizer for the desired task
*
- * @param task_sizes number of samples in each task
- * @param prop The property to fit the functions against
+ * @param task_sizes Number of training samples per task
+ * @param prop The value of the property to evaluate the loss function against for the training set
* @param n_rung Maximum rung of the features
- * @param max_param_depth maximum depth of the binary expression tress to parameterize from the root
- * @param cauchy_scaling scaling factor used for the Cauchy loss function
- * @param use_global If true use global optimization
+ * @param max_param_depth The maximum depth in the binary expression tree to set non-linear optimization
+ * @param cauchy_scaling The Cauchy scale factor used for calculating the residuals
+ * @param use_global True if global optimization is requested for non-linear optimization of parameters (Can break reproducibility)
*
* @return The correct optimizer
*/
@@ -436,12 +462,12 @@ inline NLOptimizerRegression get_reg_optimizer(
/**
* @brief Get an optimizer for the desired task
*
- * @param task_sizes number of samples in each task
- * @param prop The property to fit the functions against
+ * @param task_sizes Number of training samples per task
+ * @param prop The value of the property to evaluate the loss function against for the training set
* @param n_rung Maximum rung of the features
- * @param max_param_depth maximum depth of the binary expression tress to parameterize from the root
- * @param cauchy_scaling scaling factor used for the Cauchy loss function
- * @param use_global If true use global optimization
+ * @param max_param_depth The maximum depth in the binary expression tree to set non-linear optimization
+ * @param cauchy_scaling The Cauchy scale factor used for calculating the residuals
+ * @param use_global True if global optimization is requested for non-linear optimization of parameters (Can break reproducibility)
*
* @return The correct optimizer
*/
@@ -472,12 +498,12 @@ inline NLOptimizerRegression get_reg_optimizer(
/**
* @brief Get an optimizer for the desired task
*
- * @param task_sizes number of samples in each task
- * @param prop The property to fit the functions against
+ * @param task_sizes Number of training samples per task
+ * @param prop The value of the property to evaluate the loss function against for the training set
* @param n_rung Maximum rung of the features
- * @param max_param_depth maximum depth of the binary expression tress to parameterize from the root
- * @param cauchy_scaling scaling factor used for the Cauchy loss function
- * @param use_global If true use global optimization
+ * @param max_param_depth The maximum depth in the binary expression tree to set non-linear optimization
+ * @param cauchy_scaling The Cauchy scale factor used for calculating the residuals
+ * @param use_global True if global optimization is requested for non-linear optimization of parameters (Can break reproducibility)
*
* @return The correct optimizer
*/
@@ -508,12 +534,12 @@ inline NLOptimizerRegression get_reg_optimizer(
/**
* @brief Get an optimizer for the desired task
*
- * @param task_sizes number of samples in each task
- * @param prop The property to fit the functions against
+ * @param task_sizes Number of training samples per task
+ * @param prop The value of the property to evaluate the loss function against for the training set
* @param n_rung Maximum rung of the features
- * @param max_param_depth maximum depth of the binary expression tress to parameterize from the root
- * @param scaling factor used for the Cauchy loss function
- * @param use_global If true use global optimization
+ * @param max_param_depth The maximum depth in the binary expression tree to set non-linear optimization
+ * @param cauchy_scaling The Cauchy scale factor used for calculating the residuals
+ * @param use_global True if global optimization is requested for non-linear optimization of parameters (Can break reproducibility)
*
* @return The correct optimizer
*/
@@ -543,12 +569,12 @@ inline NLOptimizerLogRegression get_log_reg_optimizer(
/**
* @brief Get an optimizer for the desired task
*
- * @param task_sizes number of samples in each task
- * @param prop The property to fit the functions against
+ * @param task_sizes Number of training samples per task
+ * @param prop The value of the property to evaluate the loss function against for the training set
* @param n_rung Maximum rung of the features
- * @param max_param_depth maximum depth of the binary expression tress to parameterize from the root
- * @param scaling factor used for the Cauchy loss function
- * @param use_global If true use global optimization
+ * @param max_param_depth The maximum depth in the binary expression tree to set non-linear optimization
+ * @param cauchy_scaling The Cauchy scale factor used for calculating the residuals
+ * @param use_global True if global optimization is requested for non-linear optimization of parameters (Can break reproducibility)
*
* @return The correct optimizer
*/
@@ -578,12 +604,12 @@ inline NLOptimizerLogRegression get_log_reg_optimizer(
/**
* @brief Get an optimizer for the desired task
*
- * @param task_sizes number of samples in each task
- * @param prop The property to fit the functions against
+ * @param task_sizes Number of training samples per task
+ * @param prop The value of the property to evaluate the loss function against for the training set
* @param n_rung Maximum rung of the features
- * @param max_param_depth maximum depth of the binary expression tress to parameterize from the root
- * @param scaling factor used for the Cauchy loss function
- * @param use_global If true use global optimization
+ * @param max_param_depth The maximum depth in the binary expression tree to set non-linear optimization
+ * @param cauchy_scaling The Cauchy scale factor used for calculating the residuals
+ * @param use_global True if global optimization is requested for non-linear optimization of parameters (Can break reproducibility)
*
* @return The correct optimizer
*/
@@ -613,12 +639,12 @@ inline NLOptimizerLogRegression get_log_reg_optimizer(
/**
* @brief Get an optimizer for the desired task
*
- * @param task_sizes number of samples in each task
- * @param prop The property to fit the functions against
+ * @param task_sizes Number of training samples per task
+ * @param prop The value of the property to evaluate the loss function against for the training set
* @param n_rung Maximum rung of the features
- * @param max_param_depth maximum depth of the binary expression tress to parameterize from the root
- * @param scaling factor used for the Cauchy loss function
- * @param use_global If true use global optimization
+ * @param max_param_depth The maximum depth in the binary expression tree to set non-linear optimization
+ * @param cauchy_scaling The Cauchy scale factor used for calculating the residuals
+ * @param use_global True if global optimization is requested for non-linear optimization of parameters (Can break reproducibility)
*
* @return The correct optimizer
*/
@@ -648,11 +674,11 @@ inline NLOptimizerLogRegression get_log_reg_optimizer(
/**
* @brief Get an optimizer for the desired task
*
- * @param task_sizes number of samples in each task
- * @param prop The property to fit the functions against
+ * @param task_sizes Number of training samples per task
+ * @param prop The value of the property to evaluate the loss function against for the training set
* @param n_rung Maximum rung of the features
- * @param max_param_depth maximum depth of the binary expression tress to parameterize from the root
- * @param use_global If true use global optimization
+ * @param max_param_depth The maximum depth in the binary expression tree to set non-linear optimization
+ * @param use_global True if global optimization is requested for non-linear optimization of parameters (Can break reproducibility)
*
* @return The correct optimizer
*/
@@ -680,11 +706,11 @@ inline NLOptimizerClassification get_class_optimizer(
/**
* @brief Get an optimizer for the desired task
*
- * @param task_sizes number of samples in each task
- * @param prop The property to fit the functions against
+ * @param task_sizes Number of training samples per task
+ * @param prop The value of the property to evaluate the loss function against for the training set
* @param n_rung Maximum rung of the features
- * @param max_param_depth maximum depth of the binary expression tress to parameterize from the root
- * @param use_global If true use global optimization
+ * @param max_param_depth The maximum depth in the binary expression tree to set non-linear optimization
+ * @param use_global True if global optimization is requested for non-linear optimization of parameters (Can break reproducibility)
*
* @return The correct optimizer
*/
@@ -712,11 +738,11 @@ inline NLOptimizerClassification get_class_optimizer(
/**
* @brief Get an optimizer for the desired task
*
- * @param task_sizes number of samples in each task
- * @param prop The property to fit the functions against
+ * @param task_sizes Number of training samples per task
+ * @param prop The value of the property to evaluate the loss function against for the training set
* @param n_rung Maximum rung of the features
- * @param max_param_depth maximum depth of the binary expression tress to parameterize from the root
- * @param use_global If true use global optimization
+ * @param max_param_depth The maximum depth in the binary expression tree to set non-linear optimization
+ * @param use_global True if global optimization is requested for non-linear optimization of parameters (Can break reproducibility)
*
* @return The correct optimizer
*/
@@ -744,11 +770,11 @@ inline NLOptimizerClassification get_class_optimizer(
/**
* @brief Get an optimizer for the desired task
*
- * @param task_sizes number of samples in each task
- * @param prop The property to fit the functions against
+ * @param task_sizes Number of training samples per task
+ * @param prop The value of the property to evaluate the loss function against for the training set
* @param n_rung Maximum rung of the features
- * @param max_param_depth maximum depth of the binary expression tress to parameterize from the root
- * @param use_global If true use global optimization
+ * @param max_param_depth The maximum depth in the binary expression tree to set non-linear optimization
+ * @param use_global True if global optimization is requested for non-linear optimization of parameters (Can break reproducibility)
*
* @return The correct optimizer
*/
diff --git a/src/python/py_binding_cpp_def/bindings_docstring_keyed.cpp b/src/python/py_binding_cpp_def/bindings_docstring_keyed.cpp
index bc07614d092b61d825d5cca2608c7f84e790af44..f65a77a8b7d66cef0ca5bfcfc7d86df0da71b727 100644
--- a/src/python/py_binding_cpp_def/bindings_docstring_keyed.cpp
+++ b/src/python/py_binding_cpp_def/bindings_docstring_keyed.cpp
@@ -164,9 +164,9 @@ void sisso::feature_creation::registerFeatureSpace()
arg("allowed_param_ops"),
arg("prop"),
arg("project_type")="regression",
- arg("max_phi")=1,
+ arg("max_rung")=1,
arg("n_sis_select")=1,
- arg("max_store_rung")=-1,
+ arg("n_rung_store")=-1,
arg("n_rung_generate")=0,
arg("cross_corr_max")=1.0,
arg("min_abs_feat_val")=1e-50,
@@ -186,9 +186,9 @@ void sisso::feature_creation::registerFeatureSpace()
arg("allowed_param_ops"),
arg("prop"),
arg("project_type")="regression",
- arg("max_phi")=1,
+ arg("max_rung")=1,
arg("n_sis_select")=1,
- arg("max_store_rung")=-1,
+ arg("n_rung_store")=-1,
arg("n_rung_generate")=0,
arg("cross_corr_max")=1.0,
arg("min_abs_feat_val")=1e-50,
@@ -240,11 +240,11 @@ void sisso::feature_creation::registerFeatureSpace()
.add_property("scores", &FeatureSpace::scores_py, "@DocString_feat_space_scores_py@")
.add_property("task_sizes", &FeatureSpace::task_sizes_py, "@DocString_feat_space_task_sizes_py@")
.add_property("allowed_ops", &FeatureSpace::allowed_ops_py, "@DocString_feat_space_allowed_ops_py@")
- .add_property("start_gen", &FeatureSpace::start_gen_py, "@DocString_feat_space_start_gen_py@")
+ .add_property("start_rung", &FeatureSpace::start_rung_py, "@DocString_feat_space_start_rung_py@")
.add_property("feature_space_file", &FeatureSpace::feature_space_file, "@DocString_feat_space_feature_space_file@")
.add_property("l_bound", &FeatureSpace::l_bound, "@DocString_feat_space_l_bound@")
.add_property("u_bound", &FeatureSpace::u_bound, "@DocString_feat_space_u_bound@")
- .add_property("max_phi", &FeatureSpace::max_phi, "@DocString_feat_space_max_phi@")
+ .add_property("max_rung", &FeatureSpace::max_rung, "@DocString_feat_space_max_rung@")
.add_property("n_sis_select", &FeatureSpace::n_sis_select, "@DocString_feat_space_n_sis_select@")
.add_property("n_samp", &FeatureSpace::n_samp, "@DocString_feat_space_n_samp@")
.add_property("n_feat", &FeatureSpace::n_feat, "@DocString_feat_space_n_feat@")
@@ -261,9 +261,9 @@ void sisso::feature_creation::registerFeatureSpace()
arg("allowed_ops"),
arg("prop"),
arg("project_type")="regression",
- arg("max_phi")=1,
+ arg("max_rung")=1,
arg("n_sis_select")=1,
- arg("max_store_rung")=-1,
+ arg("n_rung_store")=-1,
arg("n_rung_generate")=0,
arg("cross_corr_max")=1.0,
arg("min_abs_feat_val")=1e-50,
@@ -281,9 +281,9 @@ void sisso::feature_creation::registerFeatureSpace()
arg("allowed_ops"),
arg("prop"),
arg("project_type")="regression",
- arg("max_phi")=1,
+ arg("max_rung")=1,
arg("n_sis_select")=1,
- arg("max_store_rung")=-1,
+ arg("n_rung_store")=-1,
arg("n_rung_generate")=0,
arg("cross_corr_max")=1.0,
arg("min_abs_feat_val")=1e-50,
@@ -334,11 +334,11 @@ void sisso::feature_creation::registerFeatureSpace()
.add_property("scores", &FeatureSpace::scores_py, "@DocString_feat_space_scores_py@")
.add_property("task_sizes", &FeatureSpace::task_sizes_py, "@DocString_feat_space_task_sizes_py@")
.add_property("allowed_ops", &FeatureSpace::allowed_ops_py, "@DocString_feat_space_allowed_ops_py@")
- .add_property("start_gen", &FeatureSpace::start_gen_py, "@DocString_feat_space_start_gen_py@")
+ .add_property("start_rung", &FeatureSpace::start_rung_py, "@DocString_feat_space_start_rung_py@")
.add_property("feature_space_file", &FeatureSpace::feature_space_file, "@DocString_feat_space_feature_space_file@")
.add_property("l_bound", &FeatureSpace::l_bound, "@DocString_feat_space_l_bound@")
.add_property("u_bound", &FeatureSpace::u_bound, "@DocString_feat_space_u_bound@")
- .add_property("max_phi", &FeatureSpace::max_phi, "@DocString_feat_space_max_phi@")
+ .add_property("max_rung", &FeatureSpace::max_rung, "@DocString_feat_space_max_rung@")
.add_property("n_sis_select", &FeatureSpace::n_sis_select, "@DocString_feat_space_n_sis_select@")
.add_property("n_samp", &FeatureSpace::n_samp, "@DocString_feat_space_n_samp@")
.add_property("n_feat", &FeatureSpace::n_feat, "@DocString_feat_space_n_feat@")
@@ -395,7 +395,7 @@ void sisso::feature_creation::registerUnit()
class_<sisso::feature_creation::node::NodeWrap, boost::noncopyable>("Node", no_init)
.add_property("n_samp", &Node::n_samp, "@DocString_node_n_samp@")
- .add_property("n_test_samp", &Node::n_test_samp, "@DocString_node_n_test_samp@")
+ .add_property("n_samp_test", &Node::n_samp_test, "@DocString_node_n_samp_test@")
.add_property("feat_ind", &Node::feat_ind, "@DocString_node_feat_ind@")
.add_property("arr_ind", &Node::arr_ind, "@DocString_node_arr_ind@")
.add_property("selected", &Node::selected, &Node::set_selected, "@DocString_node_selected@")
@@ -425,7 +425,7 @@ void sisso::feature_creation::registerUnit()
void (Node::*reindex_2)(unsigned long int, unsigned long int) = &Node::reindex;
class_<sisso::feature_creation::node::NodeWrap, boost::noncopyable>("Node", no_init)
.add_property("n_samp", &Node::n_samp, "@DocString_node_n_samp@")
- .add_property("n_test_samp", &Node::n_test_samp, "@DocString_node_n_test_samp@")
+ .add_property("n_samp_test", &Node::n_samp_test, "@DocString_node_n_samp_test@")
.add_property("feat_ind", &Node::feat_ind, "@DocString_node_feat_ind@")
.add_property("arr_ind", &Node::arr_ind, "@DocString_node_arr_ind@")
.add_property("selected", &Node::selected, &Node::set_selected, "@DocString_node_set_selected@")
@@ -494,13 +494,13 @@ void sisso::feature_creation::node::registerModelNode()
class_<ModelNode, bases<FeatureNode>>(
"ModelNode",
init<int, int, std::string, std::string, std::string, std::string, py::list, py::list, py::list, Unit>(
- (arg("self"), arg("feat_ind"), arg("rung"), arg("expr"), arg("latex_expr"), arg("expr_postfix"), arg("matlab_fxn_expr"), arg("value"), arg("test_value"), arg("x_in_expr_list"), arg("unit")),
+ (arg("self"), arg("feat_ind"), arg("rung"), arg("expr"), arg("latex_expr"), arg("postfix_expr"), arg("matlab_fxn_expr"), arg("value"), arg("test_value"), arg("x_in_expr_list"), arg("unit")),
"@DocString_model_node_init_list@"
)
)
.def(
init<int, int, std::string, std::string, std::string, std::string, np::ndarray, np::ndarray, py::list, Unit>(
- (arg("self"), arg("feat_ind"), arg("rung"), arg("expr"), arg("latex_expr"), arg("expr_postfix"), arg("matlab_fxn_expr"), arg("value"), arg("test_value"), arg("x_in_expr_list"), arg("unit")),
+ (arg("self"), arg("feat_ind"), arg("rung"), arg("expr"), arg("latex_expr"), arg("postfix_expr"), arg("matlab_fxn_expr"), arg("value"), arg("test_value"), arg("x_in_expr_list"), arg("unit")),
"@DocString_model_node_init_arr@"
)
)
diff --git a/src/python/py_binding_cpp_def/bindings_docstring_keyed.hpp b/src/python/py_binding_cpp_def/bindings_docstring_keyed.hpp
index fcddf2d010a143f52281262901edfc7f58a17040..a927e79778186d86aa14f01e1e335b6633f2aefa 100644
--- a/src/python/py_binding_cpp_def/bindings_docstring_keyed.hpp
+++ b/src/python/py_binding_cpp_def/bindings_docstring_keyed.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file python/bindings.hpp
- * @brief Definitions to convert C++ classes into python classes
+ * @brief Definitions to convert C++ classes into python classes
*
* @author Thomas A. R. Purcell (tpurcell)
* @bug No known bugs.
@@ -536,7 +536,7 @@ namespace sisso
{
struct SISSO_DI_Wrap : SISSO_DI, py::wrapper<SISSO_DI>
{
- inline void l0_norm(){this->get_override("l0_norm")();}
+ inline void l0_regularization(){this->get_override("l0_regularization")();}
inline void fit(){this->get_override("fit")();}
};
diff --git a/src/python/py_binding_cpp_def/conversion_utils.hpp b/src/python/py_binding_cpp_def/conversion_utils.hpp
index 97ec51edcc3aa7b85e7a4779cdeeb76c3a074361..68087c9a50c9543b40305a02ccb8085f0581d7e0 100644
--- a/src/python/py_binding_cpp_def/conversion_utils.hpp
+++ b/src/python/py_binding_cpp_def/conversion_utils.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file python/py_binding_cpp_def/conversion_utils.hpp
- * @brief A set of functions to convert std::vectors to and from python objects
+ * @brief A set of functions to convert std::vectors to and from python objects
*
* @author Thomas A. R. Purcell (tpurcell)
* @bug No known bugs.
diff --git a/src/python/py_binding_cpp_def/feature_creation/FeatureNode.cpp b/src/python/py_binding_cpp_def/feature_creation/FeatureNode.cpp
index 274a29eb01d97542aff3606b0d24115dab2dc7e2..54950b8f247f5a629975dfb6cb4cd7777994e840 100644
--- a/src/python/py_binding_cpp_def/feature_creation/FeatureNode.cpp
+++ b/src/python/py_binding_cpp_def/feature_creation/FeatureNode.cpp
@@ -23,14 +23,14 @@ FeatureNode::FeatureNode(unsigned long int feat_ind, std::string expr, np::ndarr
// Automatically resize the storage arrays
if(node_value_arrs::N_STORE_FEATURES == 0)
{
- node_value_arrs::initialize_values_arr(_n_samp, _n_test_samp, 1, 0);
+ node_value_arrs::initialize_values_arr(_n_samp, _n_samp_test, 1, 0);
}
- else if((_n_samp != node_value_arrs::N_SAMPLES) || (_n_test_samp != node_value_arrs::N_SAMPLES_TEST))
+ else if((_n_samp != node_value_arrs::N_SAMPLES) || (_n_samp_test != node_value_arrs::N_SAMPLES_TEST))
{
throw std::logic_error(
"Number of samples in current feature is not the same as the others, (" +
std::to_string(_n_samp) +
- " and " + std::to_string(_n_test_samp) +
+ " and " + std::to_string(_n_samp_test) +
" vs. " +
std::to_string(node_value_arrs::N_SAMPLES) +
" and " +
@@ -57,15 +57,15 @@ FeatureNode::FeatureNode(unsigned long int feat_ind, std::string expr, py::list
// Automatically resize the storage arrays
if(node_value_arrs::N_STORE_FEATURES == 0)
{
- node_value_arrs::initialize_values_arr(_n_samp, _n_test_samp, 1, 0);
+ node_value_arrs::initialize_values_arr(_n_samp, _n_samp_test, 1, 0);
}
- else if((_n_samp != node_value_arrs::N_SAMPLES) || (_n_test_samp != node_value_arrs::N_SAMPLES_TEST))
+ else if((_n_samp != node_value_arrs::N_SAMPLES) || (_n_samp_test != node_value_arrs::N_SAMPLES_TEST))
{
throw std::logic_error(
"Number of samples in current feature is not the same as the others, (" +
std::to_string(_n_samp) +
" and " +
- std::to_string(_n_test_samp) +
+ std::to_string(_n_samp_test) +
" vs. " +
std::to_string(node_value_arrs::N_SAMPLES) +
" and " +
diff --git a/src/python/py_binding_cpp_def/feature_creation/FeatureSpace.cpp b/src/python/py_binding_cpp_def/feature_creation/FeatureSpace.cpp
index 42487f9a6e6608b44f3ebab44d6de240894d4e44..0888272a1f547b32d7645f54d0171ef6899e11ba 100644
--- a/src/python/py_binding_cpp_def/feature_creation/FeatureSpace.cpp
+++ b/src/python/py_binding_cpp_def/feature_creation/FeatureSpace.cpp
@@ -20,9 +20,9 @@ FeatureSpace::FeatureSpace(
py::list allowed_param_ops,
py::list prop,
std::string project_type,
- int max_phi,
+ int max_rung,
int n_sis_select,
- int max_store_rung,
+ int n_rung_store,
int n_rung_generate,
double cross_corr_max,
double min_abs_feat_val,
@@ -33,13 +33,13 @@ FeatureSpace::FeatureSpace(
_phi(python_conv_utils::shared_ptr_vec_from_list<Node, FeatureNode>(phi_0)),
_phi_0(_phi),
_end_no_params(1, 0),
- _start_gen_reparam(1, 0),
+ _start_rung_reparam(1, 0),
_allowed_param_ops(python_conv_utils::from_list<std::string>(allowed_param_ops)),
_allowed_ops(python_conv_utils::from_list<std::string>(allowed_ops)),
_prop(python_conv_utils::from_list<double>(prop)),
_scores(py::len(phi_0), 0.0),
_task_sizes(node_value_arrs::TASK_SZ_TRAIN),
- _start_gen(1, 0),
+ _start_rung(1, 0),
_feature_space_file("feature_space/selected_features.txt"),
_feature_space_summary_file("feature_space/SIS_summary.txt"),
_project_type(project_type),
@@ -47,10 +47,10 @@ FeatureSpace::FeatureSpace(
_cross_cor_max(cross_corr_max),
_l_bound(min_abs_feat_val),
_u_bound(max_abs_feat_val),
- _max_phi(max_phi),
+ _max_rung(max_rung),
_n_sis_select(n_sis_select),
_n_feat(py::len(phi_0)),
- _n_rung_store(max_store_rung),
+ _n_rung_store(n_rung_store),
_n_rung_generate(n_rung_generate),
_n_samp(_phi[0]->n_samp()),
_max_param_depth(max_param_depth),
@@ -65,9 +65,9 @@ FeatureSpace::FeatureSpace(
py::list allowed_param_ops,
np::ndarray prop,
std::string project_type,
- int max_phi,
+ int max_rung,
int n_sis_select,
- int max_store_rung,
+ int n_rung_store,
int n_rung_generate,
double cross_corr_max,
double min_abs_feat_val,
@@ -78,13 +78,13 @@ FeatureSpace::FeatureSpace(
_phi(python_conv_utils::shared_ptr_vec_from_list<Node, FeatureNode>(phi_0)),
_phi_0(_phi),
_end_no_params(1, 0),
- _start_gen_reparam(1, 0),
+ _start_rung_reparam(1, 0),
_allowed_param_ops(python_conv_utils::from_list<std::string>(allowed_param_ops)),
_allowed_ops(python_conv_utils::from_list<std::string>(allowed_ops)),
_prop(python_conv_utils::from_ndarray<double>(prop)),
_scores(py::len(phi_0), 0.0),
_task_sizes(node_value_arrs::TASK_SZ_TRAIN),
- _start_gen(1, 0),
+ _start_rung(1, 0),
_feature_space_file("feature_space/selected_features.txt"),
_feature_space_summary_file("feature_space/SIS_summary.txt"),
_project_type(project_type),
@@ -92,10 +92,10 @@ FeatureSpace::FeatureSpace(
_cross_cor_max(cross_corr_max),
_l_bound(min_abs_feat_val),
_u_bound(max_abs_feat_val),
- _max_phi(max_phi),
+ _max_rung(max_rung),
_n_sis_select(n_sis_select),
_n_feat(py::len(phi_0)),
- _n_rung_store(max_store_rung),
+ _n_rung_store(n_rung_store),
_n_rung_generate(n_rung_generate),
_n_samp(_phi[0]->n_samp()),
_max_param_depth(max_param_depth),
@@ -109,9 +109,9 @@ FeatureSpace::FeatureSpace(
py::list allowed_ops,
py::list prop,
std::string project_type,
- int max_phi,
+ int max_rung,
int n_sis_select,
- int max_store_rung,
+ int n_rung_store,
int n_rung_generate,
double cross_corr_max,
double min_abs_feat_val,
@@ -123,7 +123,7 @@ FeatureSpace::FeatureSpace(
_prop(python_conv_utils::from_list<double>(prop)),
_scores(py::len(phi_0), 0.0),
_task_sizes(node_value_arrs::TASK_SZ_TRAIN),
- _start_gen(1, 0),
+ _start_rung(1, 0),
_feature_space_file("feature_space/selected_features.txt"),
_feature_space_summary_file("feature_space/SIS_summary.txt"),
_project_type(project_type),
@@ -131,10 +131,10 @@ FeatureSpace::FeatureSpace(
_cross_cor_max(cross_corr_max),
_l_bound(min_abs_feat_val),
_u_bound(max_abs_feat_val),
- _max_phi(max_phi),
+ _max_rung(max_rung),
_n_sis_select(n_sis_select),
_n_feat(py::len(phi_0)),
- _n_rung_store(max_store_rung),
+ _n_rung_store(n_rung_store),
_n_rung_generate(n_rung_generate),
_n_samp(_phi[0]->n_samp()),
_max_param_depth(0),
@@ -148,9 +148,9 @@ FeatureSpace::FeatureSpace(
py::list allowed_ops,
np::ndarray prop,
std::string project_type,
- int max_phi,
+ int max_rung,
int n_sis_select,
- int max_store_rung,
+ int n_rung_store,
int n_rung_generate,
double cross_corr_max,
double min_abs_feat_val,
@@ -162,7 +162,7 @@ FeatureSpace::FeatureSpace(
_prop(python_conv_utils::from_ndarray<double>(prop)),
_scores(py::len(phi_0), 0.0),
_task_sizes(node_value_arrs::TASK_SZ_TRAIN),
- _start_gen(1, 0),
+ _start_rung(1, 0),
_feature_space_file("feature_space/selected_features.txt"),
_feature_space_summary_file("feature_space/SIS_summary.txt"),
_project_type(project_type),
@@ -170,10 +170,10 @@ FeatureSpace::FeatureSpace(
_cross_cor_max(cross_corr_max),
_l_bound(min_abs_feat_val),
_u_bound(max_abs_feat_val),
- _max_phi(max_phi),
+ _max_rung(max_rung),
_n_sis_select(n_sis_select),
_n_feat(py::len(phi_0)),
- _n_rung_store(max_store_rung),
+ _n_rung_store(n_rung_store),
_n_rung_generate(n_rung_generate),
_n_samp(_phi[0]->n_samp()),
_max_param_depth(0),
@@ -194,7 +194,7 @@ FeatureSpace::FeatureSpace(
_phi_0(python_conv_utils::shared_ptr_vec_from_list<Node, FeatureNode>(phi_0)),
_prop(python_conv_utils::from_ndarray<double>(prop)),
_scores(py::len(phi_0), 0.0),
- _start_gen(1, 0),
+ _start_rung(1, 0),
_task_sizes(python_conv_utils::from_list<int>(task_sizes)),
_feature_space_file("feature_space/selected_features.txt"),
_feature_space_summary_file("feature_space/SIS_summary.txt"),
@@ -252,8 +252,8 @@ FeatureSpace::FeatureSpace(
}
std::vector<int> rung_inds = util_funcs::argsort<int>(rungs);
- _max_phi = *std::max_element(rungs.begin(), rungs.end());
- node_value_arrs::MAX_RUNG = _max_phi;
+ _max_rung = *std::max_element(rungs.begin(), rungs.end());
+ node_value_arrs::MAX_RUNG = _max_rung;
_phi[0] = phi_temp[rung_inds[0]];
for(int ff = 1; ff < _n_feat; ++ff)
@@ -261,13 +261,13 @@ FeatureSpace::FeatureSpace(
_phi[ff] = phi_temp[rung_inds[ff]];
if(_phi[ff]->rung() != _phi[ff - 1]->rung())
{
- _start_gen.push_back(ff);
+ _start_rung.push_back(ff);
}
}
#ifdef PARAMETERIZE
- _start_gen_reparam = {0};
+ _start_rung_reparam = {0};
_end_no_params = {0};
- for(int rr = 1; rr < _max_phi; ++rr)
+ for(int rr = 1; rr < _max_rung; ++rr)
{
nlopt_wrapper::MAX_PARAM_DEPTH = rr;
bool is_correct = true;
@@ -292,33 +292,33 @@ FeatureSpace::FeatureSpace(
}
node_value_arrs::initialize_param_storage();
- for(int rr = 1; rr <= _max_phi; ++rr)
+ for(int rr = 1; rr <= _max_rung; ++rr)
{
- int n_feat_in_rung = (rr < _max_phi ?_start_gen[rr + 1] : _phi.size()) - _start_gen[rr];
+ int n_feat_in_rung = (rr < _max_rung ?_start_rung[rr + 1] : _phi.size()) - _start_rung[rr];
// Reorder features based on the number of parameters they have (none goes first)
std::vector<int> feat_n_params(n_feat_in_rung);
std::transform(
- _phi.begin() + _start_gen[rr],
- _phi.begin() + _start_gen[rr] + n_feat_in_rung,
+ _phi.begin() + _start_rung[rr],
+ _phi.begin() + _start_rung[rr] + n_feat_in_rung,
feat_n_params.begin(),
[](node_ptr feat){return feat->n_params();}
);
std::vector<int> inds = util_funcs::argsort<int>(feat_n_params);
std::vector<node_ptr>phi_copy(n_feat_in_rung);
- std::copy_n(_phi.begin() + _start_gen.back(), n_feat_in_rung, phi_copy.begin());
+ std::copy_n(_phi.begin() + _start_rung.back(), n_feat_in_rung, phi_copy.begin());
std::transform(
inds.begin(),
inds.end(),
- _phi.begin() + _start_gen[rr],
+ _phi.begin() + _start_rung[rr],
[&phi_copy](int ind){return phi_copy[ind];}
);
// Reindex the features to match the sorted order
- int cur_ind = _start_gen.back();
+ int cur_ind = _start_rung.back();
std::for_each(
- _phi.begin() + _start_gen.back(),
- _phi.begin() + _start_gen.back() + n_feat_in_rung,
+ _phi.begin() + _start_rung.back(),
+ _phi.begin() + _start_rung.back() + n_feat_in_rung,
[&cur_ind](node_ptr feat){feat->reindex(cur_ind); ++cur_ind;}
);
@@ -343,7 +343,7 @@ FeatureSpace::FeatureSpace(
_phi_0(python_conv_utils::shared_ptr_vec_from_list<Node, FeatureNode>(phi_0)),
_prop(python_conv_utils::from_list<double>(prop)),
_scores(py::len(phi_0), 0.0),
- _start_gen(1, 0),
+ _start_rung(1, 0),
_task_sizes(python_conv_utils::from_list<int>(task_sizes)),
_feature_space_file("feature_space/selected_features.txt"),
_feature_space_summary_file("feature_space/SIS_summary.txt"),
@@ -400,8 +400,8 @@ FeatureSpace::FeatureSpace(
}
std::vector<int> rung_inds = util_funcs::argsort<int>(rungs);
- _max_phi = *std::max_element(rungs.begin(), rungs.end());
- node_value_arrs::MAX_RUNG = _max_phi;
+ _max_rung = *std::max_element(rungs.begin(), rungs.end());
+ node_value_arrs::MAX_RUNG = _max_rung;
_phi[0] = phi_temp[rung_inds[0]];
for(int ff = 1; ff < _n_feat; ++ff)
@@ -409,13 +409,13 @@ FeatureSpace::FeatureSpace(
_phi[ff] = phi_temp[rung_inds[ff]];
if(_phi[ff]->rung() != _phi[ff - 1]->rung())
{
- _start_gen.push_back(ff);
+ _start_rung.push_back(ff);
}
}
#ifdef PARAMETERIZE
- _start_gen_reparam = {0};
+ _start_rung_reparam = {0};
_end_no_params = {0};
- for(int rr = 1; rr < _max_phi; ++rr)
+ for(int rr = 1; rr < _max_rung; ++rr)
{
nlopt_wrapper::MAX_PARAM_DEPTH = rr;
bool is_correct = true;
@@ -440,33 +440,33 @@ FeatureSpace::FeatureSpace(
}
node_value_arrs::initialize_param_storage();
- for(int rr = 1; rr <= _max_phi; ++rr)
+ for(int rr = 1; rr <= _max_rung; ++rr)
{
- int n_feat_in_rung = (rr < _max_phi ?_start_gen[rr + 1] : _phi.size()) - _start_gen[rr];
+ int n_feat_in_rung = (rr < _max_rung ?_start_rung[rr + 1] : _phi.size()) - _start_rung[rr];
// Reorder features based on the number of parameters they have (none goes first)
std::vector<int> feat_n_params(n_feat_in_rung);
std::transform(
- _phi.begin() + _start_gen[rr],
- _phi.begin() + _start_gen[rr] + n_feat_in_rung,
+ _phi.begin() + _start_rung[rr],
+ _phi.begin() + _start_rung[rr] + n_feat_in_rung,
feat_n_params.begin(),
[](node_ptr feat){return feat->n_params();}
);
std::vector<int> inds = util_funcs::argsort<int>(feat_n_params);
std::vector<node_ptr>phi_copy(n_feat_in_rung);
- std::copy_n(_phi.begin() + _start_gen.back(), n_feat_in_rung, phi_copy.begin());
+ std::copy_n(_phi.begin() + _start_rung.back(), n_feat_in_rung, phi_copy.begin());
std::transform(
inds.begin(),
inds.end(),
- _phi.begin() + _start_gen[rr],
+ _phi.begin() + _start_rung[rr],
[&phi_copy](int ind){return phi_copy[ind];}
);
// Reindex the features to match the sorted order
- int cur_ind = _start_gen.back();
+ int cur_ind = _start_rung.back();
std::for_each(
- _phi.begin() + _start_gen.back(),
- _phi.begin() + _start_gen.back() + n_feat_in_rung,
+ _phi.begin() + _start_rung.back(),
+ _phi.begin() + _start_rung.back() + n_feat_in_rung,
[&cur_ind](node_ptr feat){feat->reindex(cur_ind); ++cur_ind;}
);
diff --git a/src/python/py_binding_cpp_def/feature_creation/ModelNode.cpp b/src/python/py_binding_cpp_def/feature_creation/ModelNode.cpp
index 5077c2ee01ac5e2a5df316905f2d054fdcdb94ad..531036d67479546397f36e42b7118ccf209f2521 100644
--- a/src/python/py_binding_cpp_def/feature_creation/ModelNode.cpp
+++ b/src/python/py_binding_cpp_def/feature_creation/ModelNode.cpp
@@ -18,7 +18,7 @@ ModelNode::ModelNode(
const unsigned long int rung,
const std::string expr,
const std::string latex_expr,
- const std::string expr_postfix,
+ const std::string postfix_expr,
const std::string matlab_fxn_expr,
const np::ndarray value,
const np::ndarray test_value,
@@ -30,7 +30,7 @@ ModelNode(
rung,
expr,
latex_expr,
- expr_postfix,
+ postfix_expr,
matlab_fxn_expr,
python_conv_utils::from_ndarray<double>(value),
python_conv_utils::from_ndarray<double>(test_value),
@@ -44,7 +44,7 @@ ModelNode::ModelNode(
const unsigned long int rung,
const std::string expr,
const std::string latex_expr,
- const std::string expr_postfix,
+ const std::string postfix_expr,
const std::string matlab_fxn_expr,
const py::list value,
const py::list test_value,
@@ -56,7 +56,7 @@ ModelNode(
rung,
expr,
latex_expr,
- expr_postfix,
+ postfix_expr,
matlab_fxn_expr,
python_conv_utils::from_list<double>(value),
python_conv_utils::from_list<double>(test_value),
diff --git a/src/python/py_binding_cpp_def/feature_creation/node_utils.hpp b/src/python/py_binding_cpp_def/feature_creation/node_utils.hpp
index f0eacc42ea1afab075b30dc4698f5a3e46cfb8d4..bf78fc2afbbf68544180cc6190f703537b1e6899 100644
--- a/src/python/py_binding_cpp_def/feature_creation/node_utils.hpp
+++ b/src/python/py_binding_cpp_def/feature_creation/node_utils.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file python/feature_creation/node_utils.hpp
- * @brief python binding functions for node utilities
+ * @brief python binding functions for node utilities
*
* @author Thomas A. R. Purcell (tpurcell)
* @bug No known bugs.
diff --git a/src/utils/compare_features.hpp b/src/utils/compare_features.hpp
index db95c3aec967ef6f5e010bc4a2ddfdd27a726f7e..5672ad262385059e349283e5de438052a1a8e9e2 100644
--- a/src/utils/compare_features.hpp
+++ b/src/utils/compare_features.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file utils/compare_features.hpp
- * @brief A set of functions to project features onto a second vector
+ * @brief A set of functions to project features onto a second vector
*
* @author Thomas A. R. Purcell (tpurcell)
* @bug No known bugs.
@@ -28,18 +28,18 @@ typedef std::tuple<node_ptr, double> node_sc_pair;
namespace comp_feats
{
- extern std::vector<double> CORR_CHECK; //!< vector for storing cross_corelation values
+ extern std::vector<double> CORR_CHECK; //!< Vector for storing cross_corelation values
extern std::vector<double> RANK; //!< Global variable used to store the rank variables for Spearman correlation
extern std::vector<int> INDEX; //!< Global variable used to store the sorting indexes for Spearman correlation
/**
* @brief Set the is_valid and is_valid_feat_list functions for SIS
*
- * @param project_type Type of projection to perform
- * @param max_corr Maximum cross-correlation value
- * @param n_samp number of samples in the training set
- * @param is_valid The _is_valid function from the feature_space
- * @param is_valid_feat_list The _is_valid_feat_list function from the feature_space
+ * @param project_type The type of LossFunction to use when projecting the features onto a property
+ * @param max_corr Maximum cross-correlation used for selecting features
+ * @param n_samp Number of samples in the training set
+ * @param is_valid Function used to determine of a feature is too correlated to previously selected features
+ * @param is_valid_feat_list Function used to determine of a feature is too correlated to previously selected features within a given list
*/
void set_is_valid_fxn(
const std::string project_type,
@@ -52,13 +52,13 @@ namespace comp_feats
/**
* @brief Checks the feature to see if it is still valid against previously selected features (using the Pearson correlation coefficient with a maximum cross-correlation of 1.0)
*
- * @param val_ptr pointer to value array of the current feature
- * @param n_samp number of samples in the training set
- * @param cross_cor_max The maximum cross correlation of the featues
- * @param scores_sel the scores of the previously selected features
+ * @param val_ptr Pointer to value array of the current feature
+ * @param n_samp Number of samples in the training set
+ * @param cross_cor_max Maximum cross-correlation used for selecting features
+ * @param scores_sel The projection scores for the selected feature
* @param cur_score The score of the current candidate feature
- * @param end_sel index of the feature to stop checking
- * @param start_sel index of the feature to stop checking
+ * @param end_sel The index of the feature to stop checking
+ * @param start_sel The index of the feature to start checking
*
* @return True if the feature is still valid
*/
@@ -75,18 +75,18 @@ namespace comp_feats
/**
* @brief Checks the feature to see if it is still valid against previously selected features (using the Pearson correlation coefficient with a maximum cross-correlation of 1.0)
*
- * @param val_ptr pointer to value array of the current feature
- * @param n_samp number of samples in the training set
- * @param cross_cor_max The maximum cross correlation of the featues
- * @param selected list of previously selected features
- * @param scores_sel the scores of the previously selected features
+ * @param val_ptr Pointer to value array of the current feature
+ * @param n_samp Number of samples in the training set
+ * @param cross_cor_max Maximum cross-correlation used for selecting features
+ * @param selected The list of previously selected features
+ * @param scores_sel The projection scores for the selected feature
* @param cur_score The score of the current candidate feature
*
* @return True if the feature is still valid
*/
bool valid_feature_against_selected_pearson_max_corr_1_feat_list(
- const double*
- const val_ptr, int n_samp,
+ const double* val_ptr,
+ const int n_samp,
const double cross_cor_max,
const std::vector<node_ptr>& selected,
const std::vector<double>& scores_sel,
@@ -96,11 +96,10 @@ namespace comp_feats
/**
* @brief Checks the feature to see if it is still valid against previously selected features (using the Pearson correlation coefficient with a maximum cross-correlation of 1.0)
*
- * @param val_ptr pointer to value array of the current feature
- * @param n_samp number of samples per feature
- * @param cross_cor_max maximum cross correlation for the featues
- * @param start pointer to the first feature to compare against
- * @param end pointer to the final feature to compare against
+ * @param val_ptr Pointer to value array of the current feature
+ * @param n_samp Number of samples in the training set
+ * @param cross_cor_max Maximum cross-correlation used for selecting features
+ * @param out_vec The output vector for the MPI reduce all operator
* @param cur_score The score of the current candidate feature
*
* @return True if the feature is still valid
@@ -116,13 +115,13 @@ namespace comp_feats
/**
* @brief Checks the feature to see if it is still valid against previously selected features (using the Pearson correlation coefficient)
*
- * @param val_ptr pointer to value array of the current feature
- * @param n_samp number of samples in the training set
- * @param cross_cor_max The maximum cross correlation of the featues
- * @param scores_sel the scores of the previously selected features (not used, but passed to keep consistency)
- * @param cur_score The score of the current candidate feature (not used, but passed to keep consistency)
- * @param end_sel index of the feature to stop checking
- * @param start_sel index of the feature to stop checking
+ * @param val_ptr Pointer to value array of the current feature
+ * @param n_samp Number of samples in the training set
+ * @param cross_cor_max Maximum cross-correlation used for selecting features
+ * @param scores_sel The projection scores for the selected feature
+ * @param cur_score The score of the current candidate feature
+ * @param end_sel The index of the feature to stop checking
+ * @param start_sel The index of the feature to start checking
*
* @return True if the feature is still valid
*/
@@ -139,12 +138,12 @@ namespace comp_feats
/**
* @brief Checks the feature to see if it is still valid against previously selected features (using the Pearson correlation coefficient)
*
- * @param val_ptr pointer to value array of the current feature
- * @param n_samp number of samples in the training set
- * @param cross_cor_max The maximum cross correlation of the featues
- * @param selected list of previously selected features
- * @param scores_sel the scores of the previously selected features (not used, but passed to keep consistency)
- * @param cur_score The score of the current candidate feature (not used, but passed to keep consistency)
+ * @param val_ptr Pointer to value array of the current feature
+ * @param n_samp Number of samples in the training set
+ * @param cross_cor_max Maximum cross-correlation used for selecting features
+ * @param selected The list of previously selected features
+ * @param scores_sel The projection scores for the selected feature
+ * @param cur_score The score of the current candidate feature
*
* @return True if the feature is still valid
*/
@@ -160,12 +159,11 @@ namespace comp_feats
/**
* @brief Checks the feature to see if it is still valid against previously selected features (using the Pearson correlation coefficient)
*
- * @param val_ptr pointer to value array of the current feature
- * @param n_samp number of samples per feature
- * @param cross_cor_max maximum cross correlation for the featues
- * @param start pointer to the first feature to compare against
- * @param end pointer to the final feature to compare against
- * @param cur_score The score of the current candidate feature (not used, but passed to keep consistency)
+ * @param val_ptr Pointer to value array of the current feature
+ * @param n_samp Number of samples in the training set
+ * @param cross_cor_max Maximum cross-correlation used for selecting features
+ * @param out_vec The output vector for the MPI reduce all operator
+ * @param cur_score The score of the current candidate feature
*
* @return True if the feature is still valid
*/
@@ -180,13 +178,13 @@ namespace comp_feats
/**
* @brief Checks the feature to see if it is still valid against previously selected features (using the rank correlation coefficient with a maximum cross-correlation of 1.0)
*
- * @param val_ptr pointer to value array of the current feature
- * @param n_samp number of samples in the training set
- * @param cross_cor_max The maximum cross correlation of the featues
- * @param scores_sel the scores of the previously selected features
+ * @param val_ptr Pointer to value array of the current feature
+ * @param n_samp Number of samples in the training set
+ * @param cross_cor_max Maximum cross-correlation used for selecting features
+ * @param scores_sel The projection scores for the selected feature
* @param cur_score The score of the current candidate feature
- * @param end_sel index of the feature to stop checking
- * @param start_sel index of the feature to stop checking
+ * @param end_sel The index of the feature to stop checking
+ * @param start_sel The index of the feature to start checking
*
* @return True if the feature is still valid
*/
@@ -203,11 +201,11 @@ namespace comp_feats
/**
* @brief Checks the feature to see if it is still valid against previously selected features (using the rank correlation coefficient with a maximum cross-correlation of 1.0)
*
- * @param val_ptr pointer to value array of the current feature
- * @param n_samp number of samples in the training set
- * @param cross_cor_max The maximum cross correlation of the featues
- * @param selected list of previously selected features
- * @param scores_sel the scores of the previously selected features
+ * @param val_ptr Pointer to value array of the current feature
+ * @param n_samp Number of samples in the training set
+ * @param cross_cor_max Maximum cross-correlation used for selecting features
+ * @param selected The list of previously selected features
+ * @param scores_sel The projection scores for the selected feature
* @param cur_score The score of the current candidate feature
*
* @return True if the feature is still valid
@@ -224,11 +222,10 @@ namespace comp_feats
/**
* @brief Checks the feature to see if it is still valid against previously selected features (using the rank correlation coefficient with a maximum cross-correlation of 1.0)
*
- * @param val_ptr pointer to value array of the current feature
- * @param n_samp number of samples per feature
- * @param cross_cor_max maximum cross correlation for the featues
- * @param start pointer to the first feature to compare against
- * @param end pointer to the final feature to compare against
+ * @param val_ptr Pointer to value array of the current feature
+ * @param n_samp Number of samples in the training set
+ * @param cross_cor_max Maximum cross-correlation used for selecting features
+ * @param out_vec The output vector for the MPI reduce all operator
* @param cur_score The score of the current candidate feature
*
* @return True if the feature is still valid
@@ -244,13 +241,13 @@ namespace comp_feats
/**
* @brief Checks the feature to see if it is still valid against previously selected features (using the rank correlation coefficient)
*
- * @param val_ptr pointer to value array of the current feature
- * @param n_samp number of samples in the training set
- * @param cross_cor_max The maximum cross correlation of the featues
- * @param scores_sel the scores of the previously selected features (not used, but passed to keep consistency)
- * @param cur_score The score of the current candidate feature (not used, but passed to keep consistency)
- * @param end_sel index of the feature to stop checking
- * @param start_sel index of the feature to stop checking
+ * @param val_ptr Pointer to value array of the current feature
+ * @param n_samp Number of samples in the training set
+ * @param cross_cor_max Maximum cross-correlation used for selecting features
+ * @param scores_sel The projection scores for the selected feature
+ * @param cur_score The score of the current candidate feature
+ * @param end_sel The index of the feature to stop checking
+ * @param start_sel The index of the feature to start checking
*
* @return True if the feature is still valid
*/
@@ -267,12 +264,12 @@ namespace comp_feats
/**
* @brief Checks the feature to see if it is still valid against previously selected features (using the rank correlation coefficient)
*
- * @param val_ptr pointer to value array of the current feature
- * @param n_samp number of samples in the training set
- * @param cross_cor_max The maximum cross correlation of the featues
- * @param selected list of previously selected features
- * @param scores_sel the scores of the previously selected features (not used, but passed to keep consistency)
- * @param cur_score The score of the current candidate feature (not used, but passed to keep consistency)
+ * @param val_ptr Pointer to value array of the current feature
+ * @param n_samp Number of samples in the training set
+ * @param cross_cor_max Maximum cross-correlation used for selecting features
+ * @param selected The list of previously selected features
+ * @param scores_sel The projection scores for the selected feature
+ * @param cur_score The score of the current candidate feature
*
* @return True if the feature is still valid
*/
@@ -288,12 +285,11 @@ namespace comp_feats
/**
* @brief Checks the feature to see if it is still valid against previously selected features (using the rank correlation coefficient)
*
- * @param val_ptr pointer to value array of the current feature
- * @param n_samp number of samples per feature
- * @param cross_cor_max maximum cross correlation for the featues
- * @param start pointer to the first feature to compare against
- * @param end pointer to the final feature to compare against
- * @param cur_score The score of the current candidate feature (not used, but passed to keep consistency)
+ * @param val_ptr Pointer to value array of the current feature
+ * @param n_samp Number of samples in the training set
+ * @param cross_cor_max Maximum cross-correlation used for selecting features
+ * @param out_vec The output vector for the MPI reduce all operator
+ * @param cur_score The score of the current candidate feature
*
* @return True if the feature is still valid
*/
diff --git a/src/utils/enum.hpp b/src/utils/enum.hpp
index 105f454e9f19e784f8edc32bdcf5cb03631b3170..f158331c32b884071779c88f6cd8d2b195ecb1e6 100644
--- a/src/utils/enum.hpp
+++ b/src/utils/enum.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file utils/enum.hpp
- * @brief Define node type enum
+ * @brief Define node type enum
*
* @author Thomas A. R. Purcell (tpurcell)
* @bug No known bugs.
diff --git a/src/utils/math_funcs.hpp b/src/utils/math_funcs.hpp
index ef3947093d490b2a71a383768481dc3dcf0e879b..b2c8346c7aafc837ac8c4d6bceeff3ca59f97790 100644
--- a/src/utils/math_funcs.hpp
+++ b/src/utils/math_funcs.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file utils/math_funcs.hpp
- * @brief A set of functions to get standardized mathematical operations on
+ * @brief A set of functions to get standardized mathematical operations on
*
* @author Thomas A. R. Purcell (tpurcell)
* @bug No known bugs.
@@ -35,7 +35,7 @@ namespace util_funcs
* @brief Get the sign of a number
*
* @param number The number to get the sign of
- * @return -1 for negative, 0 for 0 and 1 for positive
+ * @return -1 for negative, 0 for 0, and 1 for positive
*/
template<typename T>
inline T sign(T number){return (number > T(0)) - (number < T(0));};
@@ -44,7 +44,7 @@ namespace util_funcs
* @brief Round a number to arbitrary precision
*
* @param num The number to round
- * @param n The precision
+ * @param n The number of decimal digits after the decimal point
*
* @return The rounded number
*/
@@ -58,7 +58,7 @@ namespace util_funcs
* @brief Round a number to arbitrary precision
*
* @param num The number to round
- * @param n The precision
+ * @param n The number of binary digits after the decimal point
*
* @return The rounded number
*/
@@ -80,9 +80,9 @@ namespace util_funcs
/**
* @brief Find the mean of of a vector
*
- * @param start pointer to the start of the vector
- * @param size the size of the vector
- * @return the mean of the vector
+ * @param start A pointer to the head of the vector
+ * @param size The size of the vector
+ * @return The mean of the vector
*/
template<typename T>
inline double mean(const T* start, int size){return static_cast<double>(std::accumulate(start, start + size, 0.0)) / static_cast<double>(size);};
@@ -91,7 +91,7 @@ namespace util_funcs
* @brief Find the standard deviation of a vector
*
* @param vec The vector to calculate the stand deviation of.
- * @return the standard deviation
+ * @return The standard deviation
*/
inline double stand_dev(const std::vector<double>& vec)
{
@@ -103,9 +103,9 @@ namespace util_funcs
/**
* @brief Find the standard deviation of a vector
*
- * @param start start pointer to the start of the vector
- * @param size size the size of the vector
- * @return The standard deviation of thwe vector
+ * @param start A pointer to the head of the vector
+ * @param size The size of the vector
+ * @return The standard deviation of the vector
*/
inline double stand_dev(const double* start, const int size)
{
@@ -129,12 +129,12 @@ namespace util_funcs
);
};
/**
- * @brief [brief description]
+ * @brief Find the standard deviation of a vector
- * @param start start pointer to the start of the vector
- * @param size size the size of the vector
+ * @param start A pointer to the head of the vector
+ * @param size The size of the vector
* @param vec_mean The mean of the vector
- * @return The standard deviation of thwe vector
+ * @return The standard deviation of the vector
*/
inline double stand_dev(const double* start, const int size, const double vec_mean)
{
@@ -144,7 +144,7 @@ namespace util_funcs
/**
* @brief Find the norm of a vector
*
- * @param vec The vector to calculate the stand deviation of.
+ * @param vec The vector to calculate the norm of.
* @return the norm of the vector
*/
inline double norm(const std::vector<double>& vec)
@@ -155,8 +155,8 @@ namespace util_funcs
/**
* @brief Find the norm of a vector
- * @param start start pointer to the start of the vector
- * @param size size the size of the vector
+ * @param start A pointer to the head of the vector
+ * @param size The size of the vector
* @return The norm of the vector
*/
inline double norm(const double* start, const int size)
@@ -167,9 +167,9 @@ namespace util_funcs
/**
* @brief The Pearson correlation for two vectors
*
- * @param a the pointer to the head of the first vector
- * @param b the pointer to the head of the second vector
- * @param size the size of the vector
+ * @param a The pointer to the head of the first vector
+ * @param b The pointer to the head of the second vector
+ * @param size The size of the vector
* @return The correlation coefficient between vector a and vector b
*/
inline double r(const double* a, const double* b, const int size)
@@ -187,11 +187,11 @@ namespace util_funcs
/**
* @brief The Pearson correlation for two vectors
*
- * @param a the pointer to the head of the first vector
- * @param b the pointer to the head of the second vector
- * @param size the size of the vector
- * @param mean_a the mean of the a vector
- * @param std_a the standard deviation of the a vector
+ * @param a The pointer to the head of the first vector
+ * @param b The pointer to the head of the second vector
+ * @param size The size of the vector
+ * @param mean_a The mean of the a vector
+ * @param std_a The standard deviation of the a vector
* @return The correlation coefficient between vector a and vector b
*/
inline double r(const double* a, const double* b, const int size, const double mean_a, const double std_a)
@@ -208,13 +208,13 @@ namespace util_funcs
/**
* @brief The Pearson correlation for two vectors
*
- * @param a the pointer to the head of the first vector
- * @param b the pointer to the head of the second vector
- * @param size the size of the vector
- * @param mean_a the mean of the a vector
- * @param std_a the standard deviation of the a vector
- * @param mean_b the mean of the b vector
- * @param std_b the standard deviation of the b vector
+ * @param a The pointer to the head of the first vector
+ * @param b The pointer to the head of the second vector
+ * @param size The size of the vector
+ * @param mean_a The mean of the a vector
+ * @param std_a The standard deviation of the a vector
+ * @param mean_b The mean of the b vector
+ * @param std_b The standard deviation of the b vector
* @return The correlation coefficient between vector a and vector b
*/
inline double r(const double* a, const double* b, const int size, const double mean_a, const double std_a, const double mean_b, const double std_b)
@@ -230,9 +230,9 @@ namespace util_funcs
/**
* @brief The Pearson correlation for two vectors
*
- * @param a the pointer to the head of the first vector
- * @param b the pointer to the head of the second vector
- * @param sizes the sizes of the tasks to calculate the correlation on
+ * @param a The pointer to the head of the first vector
+ * @param b The pointer to the head of the second vector
+ * @param sizes The sizes of the tasks to calculate the correlation on
* @return The average Pearson correlations
*/
double r(const double* a, const double* b, const std::vector<int>& sizes);
@@ -240,11 +240,11 @@ namespace util_funcs
/**
* @brief The Pearson correlation for two vectors
*
- * @param a the pointer to the head of the first vector
- * @param b the pointer to the head of the second vector
- * @param sizes the sizes of the tasks to calculate the correlation on
- * @param mean_a the mean of the a vector for each task
- * @param std_a the standard deviation of the a vector for each task
+ * @param a The pointer to the head of the first vector
+ * @param b The pointer to the head of the second vector
+ * @param sizes The sizes of the tasks to calculate the correlation on
+ * @param mean_a The mean of the a vector for each task
+ * @param std_a The standard deviation of the a vector for each task
* @return The average Pearson correlations
*/
double r(const double* a, const double* b, const std::vector<int>& sizes, const std::vector<double>& mean_a, const std::vector<double>& std_a);
@@ -252,13 +252,13 @@ namespace util_funcs
/**
* @brief The Pearson correlation for two vectors
*
- * @param a the pointer to the head of the first vector
- * @param b the pointer to the head of the second vector
- * @param sizes the sizes of the tasks to calculate the correlation on
- * @param mean_a the mean of the a vector for each task
- * @param std_a the standard deviation of the a vector for each task
- * @param mean_b the mean of the b vector for each task
- * @param std_b the standard deviation of the b vector for each task
+ * @param a The pointer to the head of the first vector
+ * @param b The pointer to the head of the second vector
+ * @param sizes The sizes of the tasks to calculate the correlation on
+ * @param mean_a The mean of the a vector for each task
+ * @param std_a The standard deviation of the a vector for each task
+ * @param mean_b The mean of the b vector for each task
+ * @param std_b The standard deviation of the b vector for each task
* @return The average Pearson correlations
*/
double r(
@@ -274,10 +274,10 @@ namespace util_funcs
/**
* @brief The Pearson correlation for two vectors
*
- * @param a the pointer to the head of the first vector
- * @param b the pointer to the head of the second vector
- * @param sz the start of vector that describes the sizes of the tasks to calculate the correlation on
- * @param n_tasks number of tasks to average over
+ * @param a The pointer to the head of the first vector
+ * @param b The pointer to the head of the second vector
+ * @param sz The start of vector that describes the sizes of the tasks to calculate the correlation on
+ * @param n_tasks The number of tasks to average over
* @return The average Pearson correlations
*/
double r(const double* a, const double* b, const int* sz, const int n_tasks);
@@ -285,12 +285,12 @@ namespace util_funcs
/**
* @brief The Pearson correlation for two vectors
*
- * @param a the pointer to the head of the first vector
- * @param b the pointer to the head of the second vector
- * @param sz the start of vector that describes the sizes of the tasks to calculate the correlation on
- * @param mean_a the mean of the a vector for each task
- * @param std_a the standard deviation of the a vector for each task
- * @param n_tasks number of tasks to average over
+ * @param a The pointer to the head of the first vector
+ * @param b The pointer to the head of the second vector
+ * @param sz The start of vector that describes the sizes of the tasks to calculate the correlation on
+ * @param mean_a The mean of the a vector for each task
+ * @param std_a The standard deviation of the a vector for each task
+ * @param n_tasks The number of tasks to average over
* @return The average Pearson correlations
*/
double r(const double* a, const double* b, const int* sz, const double* mean_a, const double* std_a, const int n_tasks);
@@ -298,14 +298,14 @@ namespace util_funcs
/**
* @brief The Pearson correlation for two vectors
*
- * @param a the pointer to the head of the first vector
- * @param b the pointer to the head of the second vector
- * @param sz the start of vector that describes the sizes of the tasks to calculate the correlation on
- * @param mean_a the mean of the a vector for each task
- * @param std_a the standard deviation of the b vector for each task
- * @param mean_b the mean of the a vector for each task
- * @param std_b the standard deviation of the b vector for each task
- * @param n_tasks number of tasks to average over
+ * @param a The pointer to the head of the first vector
+ * @param b The pointer to the head of the second vector
+ * @param sz The start of vector that describes the sizes of the tasks to calculate the correlation on
+ * @param mean_a The mean of the a vector for each task
+ * @param std_a The standard deviation of the b vector for each task
+ * @param mean_b The mean of the a vector for each task
+ * @param std_b The standard deviation of the b vector for each task
+ * @param n_tasks The number of tasks to average over
* @return The average Pearson correlations
*/
double r(const double* a, const double* b, const int* sz, const double* mean_a, const double* std_a, const double* mean_b, const double* std_b, const int n_tasks);
@@ -313,9 +313,9 @@ namespace util_funcs
/**
* @brief The Pearson correlation for two vectors
*
- * @param a the pointer to the head of the first vector
- * @param b the pointer to the head of the second vector
- * @param size the size of the vector
+ * @param a The pointer to the head of the first vector
+ * @param b The pointer to the head of the second vector
+ * @param size The size of the vector
* @return The coefficient of determination between vector a and vector b
*/
inline double r2(const double* a, const double* b, const int size)
@@ -337,11 +337,11 @@ namespace util_funcs
/**
* @brief The Pearson correlation for two vectors
*
- * @param a the pointer to the head of the first vector
- * @param b the pointer to the head of the second vector
- * @param size the size of the vector
- * @param mean_a the mean of the a vector
- * @param std_a the standard deviation of the a vector
+ * @param a The pointer to the head of the first vector
+ * @param b The pointer to the head of the second vector
+ * @param size The size of the vector
+ * @param mean_a The mean of the a vector
+ * @param std_a The standard deviation of the a vector
* @return The coefficient of determination between vector a and vector b
*/
inline double r2(const double* a, const double* b, const int size, const double mean_a, const double std_a)
@@ -362,13 +362,13 @@ namespace util_funcs
/**
* @brief The Pearson correlation for two vectors
*
- * @param a the pointer to the head of the first vector
- * @param b the pointer to the head of the second vector
- * @param size the size of the vector
- * @param mean_a the mean of the a vector
- * @param std_a the standard deviation of the a vector
- * @param mean_b the mean of the b vector
- * @param std_b the standard deviation of the b vector
+ * @param a The pointer to the head of the first vector
+ * @param b The pointer to the head of the second vector
+ * @param size The size of the vector
+ * @param mean_a The mean of the a vector
+ * @param std_a The standard deviation of the a vector
+ * @param mean_b The mean of the b vector
+ * @param std_b The standard deviation of the b vector
* @return The coefficient of determination between vector a and vector b
*/
inline double r2(const double* a, const double* b, const int size, const double mean_a, const double std_a, const double mean_b, const double std_b)
@@ -382,9 +382,9 @@ namespace util_funcs
/**
* @brief Calculate the average Coefficient of Determination between two vectors
*
- * @param a the pointer to the head of the first vector
- * @param b the pointer to the head of the second vector
- * @param sizes the sizes of the tasks to calculate the correlation on
+ * @param a The pointer to the head of the first vector
+ * @param b The pointer to the head of the second vector
+ * @param sizes The sizes of the tasks to calculate the correlation on
* @return The average Coefficient of Determination
*/
double r2(const double* a, const double* b, const std::vector<int>& sizes);
@@ -392,11 +392,11 @@ namespace util_funcs
/**
* @brief Calculate the average Coefficient of Determination between two vectors
*
- * @param a the pointer to the head of the first vector
- * @param b the pointer to the head of the second vector
- * @param sizes the sizes of the tasks to calculate the correlation on
- * @param mean_a the mean of the a vector for each task
- * @param std_a the standard deviation of the a vector for each task
+ * @param a The pointer to the head of the first vector
+ * @param b The pointer to the head of the second vector
+ * @param sizes The sizes of the tasks to calculate the correlation on
+ * @param mean_a The mean of the a vector for each task
+ * @param std_a The standard deviation of the a vector for each task
* @return The average Coefficient of Determination
*/
double r2(const double* a, const double* b, const std::vector<int>& sizes, const std::vector<double>& mean_a, const std::vector<double>& std_a);
@@ -404,13 +404,13 @@ namespace util_funcs
/**
* @brief Calculate the average Coefficient of Determination between two vectors
*
- * @param a the pointer to the head of the first vector
- * @param b the pointer to the head of the second vector
- * @param sizes the sizes of the tasks to calculate the correlation on
- * @param mean_a the mean of the a vector for each task
- * @param std_a the standard deviation of the a vector for each task
- * @param mean_b the mean of the b vector for each task
- * @param std_b the standard deviation of the b vector for each task
+ * @param a The pointer to the head of the first vector
+ * @param b The pointer to the head of the second vector
+ * @param sizes The sizes of the tasks to calculate the correlation on
+ * @param mean_a The mean of the a vector for each task
+ * @param std_a The standard deviation of the a vector for each task
+ * @param mean_b The mean of the b vector for each task
+ * @param std_b The standard deviation of the b vector for each task
* @return The average Coefficient of Determination
*/
double r2(
@@ -426,10 +426,10 @@ namespace util_funcs
/**
* @brief Calculate the average Coefficient of Determination between two vectors
*
- * @param a the pointer to the head of the first vector
- * @param b the pointer to the head of the second vector
- * @param sz the start of vector that describes the sizes of the tasks to calculate the correlation on
- * @param n_tasks number of tasks to average over
+ * @param a The pointer to the head of the first vector
+ * @param b The pointer to the head of the second vector
+ * @param sz The start of vector that describes the sizes of the tasks to calculate the correlation on
+ * @param n_tasks The number of tasks to average over
* @return The average Coefficient of Determination
*/
double r2(const double* a, const double* b, const int* sz, const int n_tasks);
@@ -437,12 +437,12 @@ namespace util_funcs
/**
* @brief Calculate the average Coefficient of Determination between two vectors
*
- * @param a the pointer to the head of the first vector
- * @param b the pointer to the head of the second vector
- * @param sz the start of vector that describes the sizes of the tasks to calculate the correlation on
- * @param mean_a the mean of the a vector for each task
- * @param std_a the standard deviation of the a vector for each task
- * @param n_tasks number of tasks to average over
+ * @param a The pointer to the head of the first vector
+ * @param b The pointer to the head of the second vector
+ * @param sz The start of vector that describes the sizes of the tasks to calculate the correlation on
+ * @param mean_a The mean of the a vector for each task
+ * @param std_a The standard deviation of the a vector for each task
+ * @param n_tasks The number of tasks to average over
* @return The average Coefficient of Determination
*/
double r2(const double* a, const double* b, const int* sz, const double* mean_a, const double* std_a, const int n_tasks);
@@ -450,14 +450,14 @@ namespace util_funcs
/**
* @brief Calculate the average Coefficient of Determination between two vectors
*
- * @param a the pointer to the head of the first vector
- * @param b the pointer to the head of the second vector
- * @param sz the start of vector that describes the sizes of the tasks to calculate the correlation on
- * @param mean_a the mean of the a vector for each task
- * @param std_a the standard deviation of the a vector for each task
- * @param mean_b the mean of the b vector for each task
- * @param std_b the standard deviation of the b vector for each task
- * @param n_tasks number of tasks to average over
+ * @param a The pointer to the head of the first vector
+ * @param b The pointer to the head of the second vector
+ * @param sz The start of vector that describes the sizes of the tasks to calculate the correlation on
+ * @param mean_a The mean of the a vector for each task
+ * @param std_a The standard deviation of the a vector for each task
+ * @param mean_b The mean of the b vector for each task
+ * @param std_b The standard deviation of the b vector for each task
+ * @param n_tasks The number of tasks to average over
* @return The average Coefficient of Determination
*/
double r2(const double* a, const double* b, const int* sz, const double* mean_a, const double* std_a, const double* mean_b, const double* std_b, const int n_tasks);
@@ -465,10 +465,10 @@ namespace util_funcs
/**
* @brief Calculate the Coefficient of Determination between two vectors (For the log transformed problem)
*
- * @param a the pointer to the head of the first vector
- * @param b the pointer to the head of the second vector
+ * @param a The pointer to the head of the first vector
+ * @param b The pointer to the head of the second vector
* @param log_a the pointer to the head of the vector used to store the log_transformed a value
- * @param size the size of the vector
+ * @param size The size of the vector
* @return The Coefficient of Determination
*/
double log_r2(const double* a, const double* b, double* log_a, const int size);
@@ -476,12 +476,12 @@ namespace util_funcs
/**
* @brief Calculate the Coefficient of Determination between two vectors (For the log transformed problem)
*
- * @param a the pointer to the head of the first vector
- * @param b the pointer to the head of the second vector
+ * @param a The pointer to the head of the first vector
+ * @param b The pointer to the head of the second vector
* @param log_a the pointer to the head of the vector used to store the log_transformed a value
- * @param size the size of the vector
- * @param mean_b the mean of the b vector for each task
- * @param std_b the standard deviation of the b vector for each task
+ * @param size The size of the vector
+ * @param mean_b The mean of the b vector for each task
+ * @param std_b The standard deviation of the b vector for each task
* @return The Coefficient of Determination
*/
double log_r2(const double* a, const double* b, double* log_a, const int size, const double mean_b, const double std_b);
@@ -489,10 +489,10 @@ namespace util_funcs
/**
* @brief Calculate the average Coefficient of Determination between two vectors (For the log transformed problem)
*
- * @param a the pointer to the head of the first vector
- * @param b the pointer to the head of the second vector
+ * @param a The pointer to the head of the first vector
+ * @param b The pointer to the head of the second vector
* @param log_a the pointer to the head of the vector used to store the log_transformed a value
- * @param sizes the sizes of the tasks to calculate the correlation on
+ * @param sizes The sizes of the tasks to calculate the correlation on
* @return The average Coefficient of Determination
*/
double log_r2(const double* a, const double* b, double* log_a, const std::vector<int>& sizes);
@@ -500,12 +500,12 @@ namespace util_funcs
/**
* @brief Calculate the average Coefficient of Determination between two vectors (For the log transformed problem)
*
- * @param a the pointer to the head of the first vector
- * @param b the pointer to the head of the second vector
+ * @param a The pointer to the head of the first vector
+ * @param b The pointer to the head of the second vector
* @param log_a the pointer to the head of the vector used to store the log_transformed a value
- * @param sizes the sizes of the tasks to calculate the correlation on
- * @param mean_b the mean of the b vector for each task
- * @param std_b the standard deviation of the b vector for each task
+ * @param sizes The sizes of the tasks to calculate the correlation on
+ * @param mean_b The mean of the b vector for each task
+ * @param std_b The standard deviation of the b vector for each task
* @return The average Coefficient of Determination
*/
double log_r2(const double* a, const double* b, double* log_a, const std::vector<int>& sizes, const std::vector<double>& mean_b, const std::vector<double>& std_b);
@@ -513,11 +513,11 @@ namespace util_funcs
/**
* @brief Calculate the average Coefficient of Determination between two vectors (For the log transformed problem)
*
- * @param a the pointer to the head of the first vector
- * @param b the pointer to the head of the second vector
+ * @param a The pointer to the head of the first vector
+ * @param b The pointer to the head of the second vector
* @param log_a the pointer to the head of the vector used to store the log_transformed a value
- * @param sz the start of vector that describes the sizes of the tasks to calculate the correlation on
- * @param n_tasks number of tasks to average over
+ * @param sz The start of vector that describes the sizes of the tasks to calculate the correlation on
+ * @param n_tasks The number of tasks to average over
* @return The average Coefficient of Determination
*/
double log_r2(const double* a, const double* b, double* log_a, const int* sz, const int n_tasks);
@@ -525,13 +525,13 @@ namespace util_funcs
/**
* @brief Calculate the average Coefficient of Determination between two vectors (For the log transformed problem)
*
- * @param a the pointer to the head of the first vector
- * @param b the pointer to the head of the second vector
+ * @param a The pointer to the head of the first vector
+ * @param b The pointer to the head of the second vector
* @param log_a the pointer to the head of the vector used to store the log_transformed a value
- * @param sz the start of vector that describes the sizes of the tasks to calculate the correlation on
- * @param mean_b the mean of the b vector for each task
- * @param std_b the standard deviation of the b vector for each task
- * @param n_tasks number of tasks to average over
+ * @param sz The start of vector that describes the sizes of the tasks to calculate the correlation on
+ * @param mean_b The mean of the b vector for each task
+ * @param std_b The standard deviation of the b vector for each task
+ * @param n_tasks The number of tasks to average over
* @return The average Coefficient of Determination
*/
double log_r2(const double* a, const double* b, double* log_a, const int* sz, const double* mean_b, const double* std_b, const int n_tasks);
@@ -549,12 +549,12 @@ namespace util_funcs
/**
* @brief Calculate the Spearman's rank correlation coefficient between two vectors
*
- * @param a the pointer to the head of the first vector
- * @param b the pointer to the head of the second vector
- * @param rank_a the pointer to the head of the vector used to rank of a
- * @param rank_b the pointer to the head of the vector used to rank of b
- * @param index the pointer used to store the sorted indexes
- * @param size the size of the vector
+ * @param a The pointer to the head of the first vector
+ * @param b The pointer to the head of the second vector
+ * @param rank_a The pointer to the head of the vector used to rank of a
+ * @param rank_b The pointer to the head of the vector used to rank of b
+ * @param index The pointer used to store the sorted indexes
+ * @param size The size of the vector
* @return The Coefficient of Determination
*/
double spearman_r(const double* a, const double* b, double* rank_a, double* rank_b, int* index, const int size);
@@ -562,12 +562,12 @@ namespace util_funcs
/**
* @brief Calculate the average Spearman's rank correlation coefficient between two vectors
*
- * @param a the pointer to the head of the first vector
- * @param b the pointer to the head of the second vector
- * @param rank_a the pointer to the head of the vector used to rank of a
- * @param rank_b the pointer to the head of the vector used to rank of b
- * @param index the pointer used to store the sorted indexes
- * @param sizes the sizes of the tasks to calculate the correlation on
+ * @param a The pointer to the head of the first vector
+ * @param b The pointer to the head of the second vector
+ * @param rank_a The pointer to the head of the vector used to rank of a
+ * @param rank_b The pointer to the head of the vector used to rank of b
+ * @param index The pointer used to store the sorted indexes
+ * @param sizes The sizes of the tasks to calculate the correlation on
* @return The average Coefficient of Determination
*/
double spearman_r(const double* a, const double* b, double* rank_a, double* rank_b, int* index, const std::vector<int>& sizes);
@@ -575,13 +575,13 @@ namespace util_funcs
/**
* @brief Calculate the average Spearman's rank correlation coefficient between two vectors
*
- * @param a the pointer to the head of the first vector
- * @param b the pointer to the head of the second vector
- * @param rank_a the pointer to the head of the vector used to rank of a
- * @param rank_b the pointer to the head of the vector used to rank of b
- * @param index the pointer used to store the sorted indexes
- * @param sz the start of vector that describes the sizes of the tasks to calculate the correlation on
- * @param n_tasks number of tasks to average over
+ * @param a The pointer to the head of the first vector
+ * @param b The pointer to the head of the second vector
+ * @param rank_a The pointer to the head of the vector used to rank of a
+ * @param rank_b The pointer to the head of the vector used to rank of b
+ * @param index The pointer used to store the sorted indexes
+ * @param sz The start of vector that describes the sizes of the tasks to calculate the correlation on
+ * @param n_tasks The number of tasks to average over
* @return The average Coefficient of Determination
*/
double spearman_r(const double* a, const double* b, double* rank_a, double* rank_b, int* index, const int* sz, const int n_tasks);
@@ -605,9 +605,9 @@ namespace util_funcs
/**
* @brief Sort a vector and return the indexes of the unsorted array that corresponds to the sorted one
*
- * @param begin the starting point for the sorting
- * @param end the end point for the sorting
- * @param vec vector to sort
+ * @param begin The starting point for the sorting
+ * @param end The end point for the sorting
+ * @param vec The vector to sort
* @return The indexes of the sorted array
*/
template<typename T>
@@ -619,9 +619,9 @@ namespace util_funcs
/**
* @brief Sort a vector and return the indexes of the unsorted array that corresponds to the sorted one
*
- * @param begin the starting point for the sorting
- * @param end the end point for the sorting
- * @param vec_begin start pointer to the vector to sort
+ * @param begin The starting point for the sorting
+ * @param end The end point for the sorting
+ * @param vec_begin A pointer to the head of the vector to sort
*/
template<typename T>
inline void argsort(int* begin, int* end, const T* vec_begin)
@@ -629,7 +629,7 @@ namespace util_funcs
std::sort(begin, end, [vec_begin](int i1, int i2){return vec_begin[i1] < vec_begin[i2];});
}
- /**const
+ /**
* @brief The maximum absolute value of the vector
*
* @param start The starting point for the comparison
@@ -658,9 +658,9 @@ namespace util_funcs
/**
* @brief Iterate the set of indices to get all combinations
*
- * @param inds indices to iterate over
- * @param size size of the indices
- * @param increment amount to increment the indices
+ * @param inds The indices to iterate over
+ * @param size The size of the indices
+ * @param increment The amount to increment the indices
* @return True if the all combinations have been explored
*/
bool iterate(std::vector<int>& inds, int size, int increment);
diff --git a/src/utils/mkl_interface.hpp b/src/utils/mkl_interface.hpp
index 5cf5a76592e0c78eb84e4b5dafa99f210ff346da..7248f1ddf037e488cd1581fcff75e4de4827ff3b 100644
--- a/src/utils/mkl_interface.hpp
+++ b/src/utils/mkl_interface.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file utils/mkl_interface.hpp
- * @brief wrapper functions to a mkl blas functionality
+ * @brief wrapper functions to a mkl blas functionality
*
* @author Thomas A. Purcell (tpurcell90)
* @bug No known bugs
diff --git a/src/utils/project.hpp b/src/utils/project.hpp
index 954a97156015f0b355b8121036c5819748789317..ba3cd752f4483c12539e1ba9dcf61b0adbdec6e8 100644
--- a/src/utils/project.hpp
+++ b/src/utils/project.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file utils/project.hpp
- * @brief A set of functions to project features onto a second vector
+ * @brief A set of functions to project features onto a second vector
*
* @author Thomas A. R. Purcell (tpurcell)
* @bug No known bugs.
@@ -56,7 +56,7 @@ namespace project_funcs
*
* @param loss The LossFunnction used for the projection
* @param feats The set of features to calculate
- * @param scores vector to output the scores to
+ * @param scores A pointer to the head of the vector to output the scores to
*/
void project_loss(std::shared_ptr<LossFunction> loss, const std::vector<node_ptr>& feats, double* scores);
@@ -65,7 +65,7 @@ namespace project_funcs
*
* @param loss The LossFunnction used for the projection
* @param feats The set of features to calculate
- * @param scores vector to output the scores to
+ * @param scores A pointer to the head of the vector to output the scores to
*/
void project_loss_no_omp(std::shared_ptr<LossFunction> loss, const std::vector<node_ptr>& feats, double* scores);
}
diff --git a/src/utils/string_utils.hpp b/src/utils/string_utils.hpp
index 652a9f27a45f858620d66d28f1dd57558be3983d..df48d193f9e0fbb57bb23a7c4a23e8949f4c97d7 100644
--- a/src/utils/string_utils.hpp
+++ b/src/utils/string_utils.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file utils/string_utils.hpp
- * @brief A set of functions to manipulate strings
+ * @brief A set of functions to manipulate strings
*
* @author Thomas A. R. Purcell (tpurcell)
* @bug No known bugs.
@@ -35,7 +35,7 @@ namespace str_utils
* @brief split a string and trim whitespace
*
* @param str String to split
- * @param split_tokens characters to split the string with
+ * @param split_tokens Characters to split the string with
* @return The vector with the string split along the split tokens
*/
std::vector<std::string> split_string_trim(const std::string str, const std::string split_tokens = ",;:");
@@ -44,7 +44,7 @@ namespace str_utils
* @brief Convert a string into a latex string
*
* @param str String to convert to a latex string
- * @return The latexified version of the string
+ * @return The LaTeXified version of the string
*/
std::string latexify(const std::string str);
diff --git a/src/utils/vector_utils.hpp b/src/utils/vector_utils.hpp
index 08c9f249595f7cd0488d040d6aa345b8a02bbb8b..7801e11df7e25599c134b44b43a95a2a347793fe 100644
--- a/src/utils/vector_utils.hpp
+++ b/src/utils/vector_utils.hpp
@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
/** @file utils/vector_utils.hpp
- * @brief A set of functions to manipulate vectors
+ * @brief A set of functions to manipulate vectors
*
* @author Thomas A. R. Purcell (tpurcell)
* @bug No known bugs.
@@ -25,6 +25,13 @@
namespace vector_utils
{
+/**
+ * @brief Return a vector of all unique elements of in_vec
+ *
+ * @param in_vec The input vector
+ * @tparam T The type of elements of T
+ * @return The vector of unique elements of in_vec
+ */
template<typename T>
std::vector<T> unique(const std::vector<T> in_vec)
{
diff --git a/tests/googletest/feature_creation/feature_space/test_feat_space.cc b/tests/googletest/feature_creation/feature_space/test_feat_space.cc
index be8e080a79b6f5401d44eafe2ac5a05dca182dc6..6f89b7fae353bdc2bc6ed25b0c89466743a41b1c 100644
--- a/tests/googletest/feature_creation/feature_space/test_feat_space.cc
+++ b/tests/googletest/feature_creation/feature_space/test_feat_space.cc
@@ -104,7 +104,7 @@ namespace
EXPECT_STREQ(feat_space.feature_space_file().c_str(), "feature_space/selected_features.txt");
EXPECT_EQ(feat_space.l_bound(), 1e-50);
EXPECT_EQ(feat_space.u_bound(), 1e50);
- EXPECT_EQ(feat_space.max_phi(), 2);
+ EXPECT_EQ(feat_space.max_rung(), 2);
EXPECT_EQ(feat_space.n_sis_select(), 10);
EXPECT_EQ(feat_space.n_samp(), 10);
EXPECT_EQ(feat_space.n_feat(), 154);
@@ -159,7 +159,7 @@ namespace
EXPECT_STREQ(feat_space.feature_space_file().c_str(), "feature_space/selected_features.txt");
EXPECT_EQ(feat_space.l_bound(), 1e-50);
EXPECT_EQ(feat_space.u_bound(), 1e50);
- EXPECT_EQ(feat_space.max_phi(), 2);
+ EXPECT_EQ(feat_space.max_rung(), 2);
EXPECT_EQ(feat_space.n_sis_select(), 10);
EXPECT_EQ(feat_space.n_samp(), 10);
EXPECT_EQ(feat_space.n_feat(), 16);
@@ -214,7 +214,7 @@ namespace
EXPECT_STREQ(feat_space.feature_space_file().c_str(), "feature_space/selected_features.txt");
EXPECT_EQ(feat_space.l_bound(), 1e-50);
EXPECT_EQ(feat_space.u_bound(), 1e50);
- EXPECT_EQ(feat_space.max_phi(), 2);
+ EXPECT_EQ(feat_space.max_rung(), 2);
EXPECT_EQ(feat_space.n_sis_select(), 10);
EXPECT_EQ(feat_space.n_samp(), 10);
EXPECT_EQ(feat_space.n_feat(), 154);
@@ -271,7 +271,7 @@ namespace
EXPECT_STREQ(feat_space.feature_space_file().c_str(), "feature_space/selected_features.txt");
EXPECT_EQ(feat_space.l_bound(), 1e-50);
EXPECT_EQ(feat_space.u_bound(), 1e50);
- EXPECT_EQ(feat_space.max_phi(), 2);
+ EXPECT_EQ(feat_space.max_rung(), 2);
EXPECT_EQ(feat_space.n_sis_select(), 10);
EXPECT_EQ(feat_space.n_samp(), 10);
EXPECT_EQ(feat_space.n_feat(), 154);
@@ -327,7 +327,7 @@ namespace
EXPECT_STREQ(feat_space.feature_space_file().c_str(), "feature_space/selected_features.txt");
EXPECT_EQ(feat_space.l_bound(), 1e-50);
EXPECT_EQ(feat_space.u_bound(), 1e50);
- EXPECT_EQ(feat_space.max_phi(), 0);
+ EXPECT_EQ(feat_space.max_rung(), 0);
EXPECT_EQ(feat_space.n_sis_select(), 1);
EXPECT_EQ(feat_space.n_samp(), 10);
EXPECT_EQ(feat_space.n_feat(), 3);