diff --git a/.gitignore b/.gitignore index 84a40ed61d242d3adf2f233ace8ca71e1b70bfb4..f59be5dd836b793b758bb7232117b2bbae2c727d 100644 --- a/.gitignore +++ b/.gitignore @@ -1,3 +1,5 @@ +.idea + # Byte-compiled / optimized / DLL files __pycache__/ *.py[cod] diff --git a/MANIFEST.in b/MANIFEST.in new file mode 100644 index 0000000000000000000000000000000000000000..e423c98c9223532b3fe494838fc109a69ecf83ee --- /dev/null +++ b/MANIFEST.in @@ -0,0 +1 @@ +include README.md LICENSE NOTICE diff --git a/assets/logo-mpg.png b/assets/tcmi/logo-mpg.png similarity index 100% rename from assets/logo-mpg.png rename to assets/tcmi/logo-mpg.png diff --git a/assets/logo-nomad.png b/assets/tcmi/logo-nomad.png similarity index 100% rename from assets/logo-nomad.png rename to assets/tcmi/logo-nomad.png diff --git a/data/2d_gaussian.csv b/data/tcmi/2d_gaussian.csv similarity index 100% rename from data/2d_gaussian.csv rename to data/tcmi/2d_gaussian.csv diff --git a/data/cache.dat b/data/tcmi/cache.dat similarity index 100% rename from data/cache.dat rename to data/tcmi/cache.dat diff --git a/data/concrete.csv b/data/tcmi/concrete.csv similarity index 100% rename from data/concrete.csv rename to data/tcmi/concrete.csv diff --git a/data/forestfires.csv b/data/tcmi/forestfires.csv similarity index 100% rename from data/forestfires.csv rename to data/tcmi/forestfires.csv diff --git a/data/friedman.csv b/data/tcmi/friedman.csv similarity index 100% rename from data/friedman.csv rename to data/tcmi/friedman.csv diff --git a/data/octet-binary-compound-semiconductors.csv b/data/tcmi/octet-binary-compound-semiconductors.csv similarity index 100% rename from data/octet-binary-compound-semiconductors.csv rename to data/tcmi/octet-binary-compound-semiconductors.csv diff --git a/metainfo.json b/metainfo.json new file mode 100644 index 0000000000000000000000000000000000000000..1927deab7ff1e69dfda75fd3bf46a5a43bb3a92a --- /dev/null +++ b/metainfo.json @@ -0,0 +1,45 @@ +{ + "authors": [ + "Regler, Benjamin", + "Scheffler, Matthias", + "Ghiringhelli, Luca M." + ], + "email": "regler@fhi-berlin.mpg.de", + "title": "Total cumulative mutual information", + "description": "This interactive notebook includes the original implementation of total cumulative mutual information (TCMI) to reproduce the main results presented in the publication.", + "url": "https://gitlab.mpcdf.mpg.de/nomad-lab/analytics-tcmi", + "link": "https://analytics-toolkit.nomad-coe.eu/hub/user-redirect/notebooks/tutorials/tcmi.ipynb", + "link_public": "https://analytics-toolkit.nomad-coe.eu/public/user-redirect/notebooks/tutorials/tcmi.ipynb", + "updated": "2020-01-14", + "flags": { + "featured": true, + "top_of_list": false + }, + "labels": { + "application_keyword": [ + "information theory", + "mutual information", + "cumulative entropy", + "feature selection" + ], + "application_section": [ + "Materials property prediction" + ], + "application_system": [ + "System" + ], + "category": [ + "Tutorial" + ], + "data_analytics_method": [ + "Clustering", "TCMI" + ], + "language": [ + "python", + "javascript" + ], + "platform": [ + "jupyter" + ] + } +} \ No newline at end of file diff --git a/setup.py b/setup.py new file mode 100644 index 0000000000000000000000000000000000000000..cd7dd9cea3480bd313cbbd49b2360acac1a2ee0e --- /dev/null +++ b/setup.py @@ -0,0 +1,32 @@ +# !/usr/bin/env python +# -*- coding: utf-8 -*- +""" +@package tcmi + +@copyright Copyright (c) 2018+ Fritz Haber Institute of the Max Planck Society, + Benjamin Regler <regler@fhi-berlin.mpg.de> +@license See LICENSE file for details. + +Licensed under the Apache License, Version 2.0 (the "License"). +You may not use this file except in compliance with the License. +""" + +import io +import json +import tcmi as pkg +from setuptools import setup, find_packages + +with io.open('metainfo.json', encoding='utf-8') as file: + metainfo = json.load(file) + +setup( + name=pkg.__name__, + version=pkg.__version__, + author=', '.join(metainfo['authors']), + author_email=metainfo['email'], + url=metainfo['url'], + description=metainfo['title'], + long_description=metainfo['description'], + packages=find_packages(), + install_requires=['numpy', 'scipy', 'pandas', 'scikit-learn', 'joblib'], +) \ No newline at end of file diff --git a/tcmi.ipynb b/tcmi.ipynb index fcc157cf0af217fe1cfc0d70efd3294d2e51a02c..232cd2d19501a766286157d32a4437f1f1d9306d 100644 --- a/tcmi.ipynb +++ b/tcmi.ipynb @@ -29,8 +29,8 @@ "</p>\n", " \n", "<div> \n", - " <img style=\"float: left;\" src=\"assets/logo-mpg.png\" width=\"200\"> \n", - " <img style=\"float: right;\" src=\"assets/logo-nomad.png\" width=\"250\">\n", + " <img style=\"float: left;\" src=\"assets/tcmi/logo-mpg.png\" width=\"200\"> \n", + " <img style=\"float: right;\" src=\"assets/tcmi/logo-nomad.png\" width=\"250\">\n", "</div>\n", "</p>\n", "</div>" @@ -72,11 +72,10 @@ }, { "cell_type": "code", - "execution_count": 1, + "execution_count": null, "metadata": { "ExecuteTime": { - "end_time": "2020-01-20T09:39:22.700649Z", - "start_time": "2020-01-20T09:39:22.698564Z" + "start_time": "2020-02-06T20:03:16.271Z" }, "init_cell": true }, @@ -97,11 +96,10 @@ }, { "cell_type": "code", - "execution_count": 2, + "execution_count": null, "metadata": { "ExecuteTime": { - "end_time": "2020-01-20T09:39:24.010953Z", - "start_time": "2020-01-20T09:39:22.704342Z" + "start_time": "2020-02-06T20:03:16.273Z" }, "hidden": true, "init_cell": true @@ -140,7 +138,7 @@ "# Main loop\n", "if __name__ == '__main__': \n", " # Provide cache\n", - " storage = Cache('data')\n", + " storage = Cache('data/tcmi')\n", " \n", " # Configure plot environment\n", " mpl.rc('font', family='sans', size=14)\n", @@ -194,11 +192,11 @@ }, { "cell_type": "code", - "execution_count": 5, + "execution_count": null, "metadata": { "ExecuteTime": { - "end_time": "2020-01-17T09:14:10.654185Z", - "start_time": "2020-01-17T09:14:10.651576Z" + "end_time": "2020-02-06T20:01:36.476498Z", + "start_time": "2020-02-06T20:01:36.469713Z" }, "hidden": true }, @@ -240,192 +238,15 @@ }, { "cell_type": "code", - "execution_count": 6, + "execution_count": null, "metadata": { "ExecuteTime": { - "end_time": "2020-01-17T09:14:10.688047Z", - "start_time": "2020-01-17T09:14:10.655785Z" + "end_time": "2020-02-06T20:01:36.578237Z", + "start_time": "2020-02-06T20:01:36.481310Z" }, "hidden": true }, - "outputs": [ - { - "data": { - "text/html": [ - "<div>\n", - "<style scoped>\n", - " .dataframe tbody tr th:only-of-type {\n", - " vertical-align: middle;\n", - " }\n", - "\n", - " .dataframe tbody tr th {\n", - " vertical-align: top;\n", - " }\n", - "\n", - " .dataframe thead th {\n", - " text-align: right;\n", - " }\n", - "</style>\n", - "<table border=\"1\" class=\"dataframe\">\n", - " <thead>\n", - " <tr style=\"text-align: right;\">\n", - " <th></th>\n", - " <th>rho2</th>\n", - " <th>adjusted_score</th>\n", - " <th>score</th>\n", - " <th>score0</th>\n", - " <th>cmi</th>\n", - " <th>mac</th>\n", - " <th>uds</th>\n", - " <th>mcde</th>\n", - " </tr>\n", - " </thead>\n", - " <tbody>\n", - " <tr>\n", - " <th>linear</th>\n", - " <td>1.0000</td>\n", - " <td>0.9743</td>\n", - " <td>1.0000</td>\n", - " <td>0.0257</td>\n", - " <td>1.0000</td>\n", - " <td>1.0000</td>\n", - " <td>0.6653</td>\n", - " <td>0.9979</td>\n", - " </tr>\n", - " <tr>\n", - " <th>exponential</th>\n", - " <td>1.0000</td>\n", - " <td>0.9743</td>\n", - " <td>1.0000</td>\n", - " <td>0.0257</td>\n", - " <td>0.4959</td>\n", - " <td>1.0000</td>\n", - " <td>0.6547</td>\n", - " <td>0.9977</td>\n", - " </tr>\n", - " <tr>\n", - " <th>step_2</th>\n", - " <td>0.9999</td>\n", - " <td>0.9588</td>\n", - " <td>0.9812</td>\n", - " <td>0.0224</td>\n", - " <td>1.0000</td>\n", - " <td>1.0000</td>\n", - " <td>0.6655</td>\n", - " <td>0.9978</td>\n", - " </tr>\n", - " <tr>\n", - " <th>step_4</th>\n", - " <td>0.9996</td>\n", - " <td>0.9269</td>\n", - " <td>0.9456</td>\n", - " <td>0.0187</td>\n", - " <td>1.0000</td>\n", - " <td>1.0000</td>\n", - " <td>0.6657</td>\n", - " <td>0.9977</td>\n", - " </tr>\n", - " <tr>\n", - " <th>step_8</th>\n", - " <td>0.9984</td>\n", - " <td>0.8673</td>\n", - " <td>0.8822</td>\n", - " <td>0.0149</td>\n", - " <td>1.0000</td>\n", - " <td>1.0000</td>\n", - " <td>0.6655</td>\n", - " <td>0.9976</td>\n", - " </tr>\n", - " <tr>\n", - " <th>random</th>\n", - " <td>0.0091</td>\n", - " <td>0.3285</td>\n", - " <td>0.6466</td>\n", - " <td>0.3181</td>\n", - " <td>0.0165</td>\n", - " <td>0.3376</td>\n", - " <td>0.0000</td>\n", - " <td>0.5354</td>\n", - " </tr>\n", - " <tr>\n", - " <th>sawtooth_8</th>\n", - " <td>0.0016</td>\n", - " <td>0.2326</td>\n", - " <td>0.3061</td>\n", - " <td>0.0735</td>\n", - " <td>0.0264</td>\n", - " <td>0.0273</td>\n", - " <td>0.0000</td>\n", - " <td>0.1391</td>\n", - " </tr>\n", - " <tr>\n", - " <th>sawtooth_4</th>\n", - " <td>0.0004</td>\n", - " <td>0.1744</td>\n", - " <td>0.2701</td>\n", - " <td>0.0957</td>\n", - " <td>0.0038</td>\n", - " <td>0.0144</td>\n", - " <td>0.0000</td>\n", - " <td>0.0876</td>\n", - " </tr>\n", - " <tr>\n", - " <th>sawtooth_2</th>\n", - " <td>0.0001</td>\n", - " <td>0.0903</td>\n", - " <td>0.1875</td>\n", - " <td>0.0972</td>\n", - " <td>0.0000</td>\n", - " <td>0.0000</td>\n", - " <td>0.0000</td>\n", - " <td>0.0347</td>\n", - " </tr>\n", - " <tr>\n", - " <th>zero</th>\n", - " <td>0.0000</td>\n", - " <td>0.0000</td>\n", - " <td>0.0000</td>\n", - " <td>0.0000</td>\n", - " <td>0.0000</td>\n", - " <td>0.0000</td>\n", - " <td>0.0000</td>\n", - " <td>0.9989</td>\n", - " </tr>\n", - " </tbody>\n", - "</table>\n", - "</div>" - ], - "text/plain": [ - " rho2 adjusted_score score score0 cmi mac uds \\\n", - "linear 1.0000 0.9743 1.0000 0.0257 1.0000 1.0000 0.6653 \n", - "exponential 1.0000 0.9743 1.0000 0.0257 0.4959 1.0000 0.6547 \n", - "step_2 0.9999 0.9588 0.9812 0.0224 1.0000 1.0000 0.6655 \n", - "step_4 0.9996 0.9269 0.9456 0.0187 1.0000 1.0000 0.6657 \n", - "step_8 0.9984 0.8673 0.8822 0.0149 1.0000 1.0000 0.6655 \n", - "random 0.0091 0.3285 0.6466 0.3181 0.0165 0.3376 0.0000 \n", - "sawtooth_8 0.0016 0.2326 0.3061 0.0735 0.0264 0.0273 0.0000 \n", - "sawtooth_4 0.0004 0.1744 0.2701 0.0957 0.0038 0.0144 0.0000 \n", - "sawtooth_2 0.0001 0.0903 0.1875 0.0972 0.0000 0.0000 0.0000 \n", - "zero 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 \n", - "\n", - " mcde \n", - "linear 0.9979 \n", - "exponential 0.9977 \n", - "step_2 0.9978 \n", - "step_4 0.9977 \n", - "step_8 0.9976 \n", - "random 0.5354 \n", - "sawtooth_8 0.1391 \n", - "sawtooth_4 0.0876 \n", - "sawtooth_2 0.0347 \n", - "zero 0.9989 " - ] - }, - "execution_count": 6, - "metadata": {}, - "output_type": "execute_result" - } - ], + "outputs": [], "source": [ "rng = np.random.RandomState(seed=seed)\n", "tests = {\n", @@ -511,113 +332,16 @@ }, { "cell_type": "code", - "execution_count": 7, + "execution_count": null, "metadata": { "ExecuteTime": { - "end_time": "2020-01-17T09:14:11.115344Z", - "start_time": "2020-01-17T09:14:10.689464Z" + "end_time": "2020-02-06T20:01:37.268567Z", + "start_time": "2020-02-06T20:01:36.584168Z" }, "hidden": true, "hide_input": false }, - "outputs": [ - { - "data": { - "image/png": "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\n", - "text/plain": [ - "<Figure size 400x400 with 1 Axes>" - ] - }, - "metadata": { - "needs_background": "light" - }, - "output_type": "display_data" - }, - { - "data": { - "text/html": [ - "<div>\n", - "<style scoped>\n", - " .dataframe tbody tr th:only-of-type {\n", - " vertical-align: middle;\n", - " }\n", - "\n", - " .dataframe tbody tr th {\n", - " vertical-align: top;\n", - " }\n", - "\n", - " .dataframe thead th {\n", - " text-align: right;\n", - " }\n", - "</style>\n", - "<table border=\"1\" class=\"dataframe\">\n", - " <thead>\n", - " <tr style=\"text-align: right;\">\n", - " <th></th>\n", - " <th>1.0</th>\n", - " <th>2.0</th>\n", - " <th>3.0</th>\n", - " <th>4.0</th>\n", - " <th>5.0</th>\n", - " <th>6.0</th>\n", - " <th>7.0</th>\n", - " <th>8.0</th>\n", - " <th>9.0</th>\n", - " <th>10.0</th>\n", - " <th>20.0</th>\n", - " <th>30.0</th>\n", - " <th>40.0</th>\n", - " <th>50.0</th>\n", - " <th>60.0</th>\n", - " <th>70.0</th>\n", - " <th>80.0</th>\n", - " <th>90.0</th>\n", - " <th>100.0</th>\n", - " </tr>\n", - " </thead>\n", - " <tbody>\n", - " <tr>\n", - " <th>0</th>\n", - " <td>1.0</td>\n", - " <td>0.5</td>\n", - " <td>0.424673</td>\n", - " <td>0.372276</td>\n", - " <td>0.333224</td>\n", - " <td>0.302749</td>\n", - " <td>0.278171</td>\n", - " <td>0.257845</td>\n", - " <td>0.240703</td>\n", - " <td>0.226015</td>\n", - " <td>0.144956</td>\n", - " <td>0.10951</td>\n", - " <td>0.089073</td>\n", - " <td>0.075596</td>\n", - " <td>0.065964</td>\n", - " <td>0.058698</td>\n", - " <td>0.053</td>\n", - " <td>0.048399</td>\n", - " <td>0.044597</td>\n", - " </tr>\n", - " </tbody>\n", - "</table>\n", - "</div>" - ], - "text/plain": [ - " 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 \\\n", - "0 1.0 0.5 0.424673 0.372276 0.333224 0.302749 0.278171 0.257845 \n", - "\n", - " 9.0 10.0 20.0 30.0 40.0 50.0 60.0 \\\n", - "0 0.240703 0.226015 0.144956 0.10951 0.089073 0.075596 0.065964 \n", - "\n", - " 70.0 80.0 90.0 100.0 \n", - "0 0.058698 0.053 0.048399 0.044597 " - ] - }, - "execution_count": 7, - "metadata": {}, - "output_type": "execute_result" - } - ], + "outputs": [], "source": [ "key = 'correction-term'\n", "steps = np.concatenate((np.linspace(1, 9, num=9), np.linspace(10, 100, num=10)))\n", @@ -698,51 +422,16 @@ }, { "cell_type": "code", - "execution_count": 8, + "execution_count": null, "metadata": { "ExecuteTime": { - "end_time": "2020-01-17T09:14:11.136801Z", - "start_time": "2020-01-17T09:14:11.116550Z" + "end_time": "2020-02-06T20:01:37.410841Z", + "start_time": "2020-02-06T20:01:37.278460Z" }, "hidden": true, "scrolled": true }, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "\n", - "Size: 10\n", - "- Dimension: 0 \n", - "- Dimension: 1 \n", - "- Dimension: 2 \n", - "- Dimension: 3 \n", - "- Dimension: 4 \n", - "\n", - "Size: 50\n", - "- Dimension: 0 \n", - "- Dimension: 1 \n", - "- Dimension: 2 \n", - "- Dimension: 3 \n", - "- Dimension: 4 \n", - "\n", - "Size: 100\n", - "- Dimension: 0 \n", - "- Dimension: 1 \n", - "- Dimension: 2 \n", - "- Dimension: 3 \n", - "- Dimension: 4 \n", - "\n", - "Size: 500\n", - "- Dimension: 0 \n", - "- Dimension: 1 \n", - "- Dimension: 2 \n", - "- Dimension: 3 \n", - "- Dimension: 4 \n" - ] - } - ], + "outputs": [], "source": [ "dtypes = [('dimension', np.int), ('total_score', np.float_),\n", " ('score', np.float_), ('score_corr', np.float_), ('score0', np.float_)]\n", @@ -792,28 +481,15 @@ }, { "cell_type": "code", - "execution_count": 9, + "execution_count": null, "metadata": { "ExecuteTime": { - "end_time": "2020-01-17T09:14:11.659328Z", - "start_time": "2020-01-17T09:14:11.138689Z" + "end_time": "2020-02-06T20:01:38.318289Z", + "start_time": "2020-02-06T20:01:37.421003Z" }, "hidden": true }, - "outputs": [ - { - "data": { - "image/png": "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\n", - "text/plain": [ - "<Figure size 1152x280.976 with 4 Axes>" - ] - }, - "metadata": { - "needs_background": "light" - }, - "output_type": "display_data" - } - ], + "outputs": [], "source": [ "##\n", "# Ploat data\n", @@ -950,31 +626,15 @@ }, { "cell_type": "code", - "execution_count": 10, + "execution_count": null, "metadata": { "ExecuteTime": { - "end_time": "2020-01-17T09:14:14.090857Z", - "start_time": "2020-01-17T09:14:11.660643Z" + "end_time": "2020-02-06T20:01:44.122288Z", + "start_time": "2020-02-06T20:01:38.339997Z" }, "hidden": true }, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "Subspace: {} -> [-inf +- 0.00]\n", - "x1 [Score: 0.22 +- 0.00]\n", - "x2 [Score: 0.22 +- 0.00]\n", - "x3 [Score: 0.22 +- 0.00]\n", - "Bound: -inf -> -inf\n", - "[2020-01-17 10:14:14 - 0 subspaces remaining]\n", - "\n", - "\n", - "Balance search tree\n" - ] - } - ], + "outputs": [], "source": [ "short_size = 50\n", "xx = 2 * np.random.random_sample(short_size) - 1\n", @@ -1010,11 +670,11 @@ }, { "cell_type": "code", - "execution_count": 11, + "execution_count": null, "metadata": { "ExecuteTime": { - "end_time": "2020-01-17T09:14:14.095658Z", - "start_time": "2020-01-17T09:14:14.092601Z" + "end_time": "2020-02-06T20:01:44.194903Z", + "start_time": "2020-02-06T20:01:44.138636Z" }, "hidden": true }, @@ -1044,27 +704,18 @@ }, { "cell_type": "code", - "execution_count": 12, + "execution_count": null, "metadata": { "ExecuteTime": { - "end_time": "2020-01-17T09:14:14.117423Z", - "start_time": "2020-01-17T09:14:14.096814Z" + "end_time": "2020-02-06T20:01:44.310396Z", + "start_time": "2020-02-06T20:01:44.203485Z" }, "hidden": true }, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "Empirical mean = array([-0.01529728, 0.0134394 ])\n", - "Empirical covariance = array([[0.98049267, 0.51549993], [0.51549993, 1.01985939]])\n" - ] - } - ], + "outputs": [], "source": [ "target = 'Gaussian'\n", - "data = pd.read_csv('data/2d_gaussian.csv', low_memory=False)\n", + "data = pd.read_csv('data/tcmi/2d_gaussian.csv', low_memory=False)\n", "#data = utils.prepare_data(data, target)\n", "\n", "# Get data\n", @@ -1107,38 +758,15 @@ }, { "cell_type": "code", - "execution_count": 13, + "execution_count": null, "metadata": { "ExecuteTime": { - "end_time": "2020-01-17T09:14:14.125156Z", - "start_time": "2020-01-17T09:14:14.118966Z" + "end_time": "2020-02-06T20:01:44.380950Z", + "start_time": "2020-02-06T20:01:44.320839Z" }, "hidden": true }, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "{-|x|,-|y|} = 0.551\n", - "{|x|,y} = 0.527\n", - "{-x,|y|} = 0.522\n", - "{|x|,-y} = 0.521\n", - "{x,|y|} = 0.515\n", - "{x,-y} = 0.492\n", - "{x,y} = 0.489\n", - "{-x,y} = 0.472\n", - "{|x|,|y|} = 0.471\n", - "{-x,-y} = 0.465\n", - "{|x|,-|y|} = 0.411\n", - "{x,-|y|} = 0.406\n", - "{-|x|,|y|} = 0.392\n", - "{-x,-|y|} = 0.391\n", - "{-|x|,-y} = 0.373\n", - "{-|x|,y} = 0.370\n" - ] - } - ], + "outputs": [], "source": [ "estimator = DependenceEstimator(method='tcmi', n_jobs=-1)\n", "key = 'gaussian_optimal_subspace'\n", @@ -1180,93 +808,15 @@ }, { "cell_type": "code", - "execution_count": 14, + "execution_count": null, "metadata": { "ExecuteTime": { - "end_time": "2020-01-17T09:14:14.139619Z", - "start_time": "2020-01-17T09:14:14.127330Z" + "end_time": "2020-02-06T20:01:44.492316Z", + "start_time": "2020-02-06T20:01:44.415300Z" }, "hidden": true }, - "outputs": [ - { - "data": { - "text/html": [ - "<div>\n", - "<style scoped>\n", - " .dataframe tbody tr th:only-of-type {\n", - " vertical-align: middle;\n", - " }\n", - "\n", - " .dataframe tbody tr th {\n", - " vertical-align: top;\n", - " }\n", - "\n", - " .dataframe thead th {\n", - " text-align: right;\n", - " }\n", - "</style>\n", - "<table border=\"1\" class=\"dataframe\">\n", - " <thead>\n", - " <tr style=\"text-align: right;\">\n", - " <th></th>\n", - " <th>tcmi</th>\n", - " <th>cmi</th>\n", - " <th>mac</th>\n", - " <th>uds</th>\n", - " <th>mcde</th>\n", - " </tr>\n", - " </thead>\n", - " <tbody>\n", - " <tr>\n", - " <th>50</th>\n", - " <td>0.372399</td>\n", - " <td>0.371234</td>\n", - " <td>0.837915</td>\n", - " <td>0.0</td>\n", - " <td>0.701287</td>\n", - " </tr>\n", - " <tr>\n", - " <th>100</th>\n", - " <td>0.511599</td>\n", - " <td>0.263093</td>\n", - " <td>0.723145</td>\n", - " <td>0.0</td>\n", - " <td>0.797058</td>\n", - " </tr>\n", - " <tr>\n", - " <th>200</th>\n", - " <td>0.514346</td>\n", - " <td>0.261742</td>\n", - " <td>0.627305</td>\n", - " <td>0.0</td>\n", - " <td>0.856580</td>\n", - " </tr>\n", - " <tr>\n", - " <th>500</th>\n", - " <td>0.489740</td>\n", - " <td>0.223698</td>\n", - " <td>0.358752</td>\n", - " <td>0.0</td>\n", - " <td>0.910764</td>\n", - " </tr>\n", - " </tbody>\n", - "</table>\n", - "</div>" - ], - "text/plain": [ - " tcmi cmi mac uds mcde\n", - "50 0.372399 0.371234 0.837915 0.0 0.701287\n", - "100 0.511599 0.263093 0.723145 0.0 0.797058\n", - "200 0.514346 0.261742 0.627305 0.0 0.856580\n", - "500 0.489740 0.223698 0.358752 0.0 0.910764" - ] - }, - "execution_count": 14, - "metadata": {}, - "output_type": "execute_result" - } - ], + "outputs": [], "source": [ "results = {}\n", "for size in sizes:\n", @@ -1310,158 +860,16 @@ }, { "cell_type": "code", - "execution_count": 15, + "execution_count": null, "metadata": { "ExecuteTime": { - "end_time": "2020-01-17T09:14:14.165781Z", - "start_time": "2020-01-17T09:14:14.141425Z" + "end_time": "2020-02-06T20:01:44.598306Z", + "start_time": "2020-02-06T20:01:44.501472Z" }, "hidden": true, "scrolled": true }, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "\n", - "/**\n", - " * Data points = 50\n", - " */\n", - "\n", - "Method: tcmi\n", - " [ 2] {logistic,x} = 0.54\n", - " [ 2] {rayleigh,weibull} = 0.53\n", - " [ 2] {x,y} = 0.52\n", - " [ 1] {x} = 0.38\n", - "\n", - "Method: cmi\n", - " [ 1] {y} = 1.00\n", - " [ 1] {logistic} = 1.00\n", - " [ 1] {triangular} = 1.00\n", - " [ 1] {laplace} = 1.00\n", - "\n", - "Method: mac\n", - " [ 2] {y,laplace} = 0.90\n", - " [ 2] {y,x} = 0.89\n", - " [ 2] {y,triangular} = 0.89\n", - " [ 2] {y,exponential} = 0.89\n", - " [ 2] {y,normal} = 0.89\n", - " [ 2] {y,rayleigh} = 0.89\n", - " [ 2] {y,uniform} = 0.89\n", - " [ 2] {y,weibull} = 0.89\n", - " [ 2] {y,logistic} = 0.89\n", - " [ 1] {y} = 0.88\n", - "\n", - "Method: uds\n", - " [ 1] {laplace} = 0.52\n", - "\n", - "Method: mcde\n", - " [ 1] {y} = 0.84\n", - "\n", - "\n", - "/**\n", - " * Data points = 100\n", - " */\n", - "\n", - "Method: tcmi\n", - " [ 2] {y,rayleigh} = 0.57\n", - " [ 2] {y,laplace} = 0.55\n", - " [ 2] {y,uniform} = 0.55\n", - " [ 1] {y} = 0.46\n", - "\n", - "Method: cmi\n", - " [ 1] {y} = 1.00\n", - " [ 1] {logistic} = 1.00\n", - " [ 1] {normal} = 1.00\n", - " [ 1] {triangular} = 1.00\n", - " [ 1] {laplace} = 1.00\n", - "\n", - "Method: mac\n", - " [ 1] {x} = 0.84\n", - " [ 2] {x,laplace} = 0.82\n", - " [ 2] {x,logistic} = 0.82\n", - " [ 2] {x,weibull} = 0.82\n", - " [ 2] {x,triangular} = 0.82\n", - " [ 2] {x,exponential} = 0.82\n", - " [ 2] {x,normal} = 0.82\n", - " [ 2] {x,rayleigh} = 0.82\n", - " [ 2] {x,uniform} = 0.82\n", - " [ 2] {x,y} = 0.82\n", - " [ 1] {y} = 0.81\n", - "\n", - "Method: uds\n", - " [ 1] {y} = 0.49\n", - " [ 1] {normal} = 0.48\n", - "\n", - "Method: mcde\n", - " [ 1] {x} = 0.88\n", - " [ 1] {y} = 0.86\n", - "\n", - "\n", - "/**\n", - " * Data points = 200\n", - " */\n", - "\n", - "Method: tcmi\n", - " [ 2] {x,y} = 0.58\n", - " [ 2] {y,exponential} = 0.55\n", - " [ 1] {y} = 0.48\n", - " [ 2] {y,poisson} = 0.47\n", - "\n", - "Method: cmi\n", - " [ 1] {y} = 1.00\n", - " [ 1] {x} = 1.00\n", - " [ 1] {triangular} = 1.00\n", - " [ 1] {laplace} = 1.00\n", - " [ 1] {logistic} = 1.00\n", - " [ 1] {normal} = 1.00\n", - "\n", - "Method: mac\n", - " [ 1] {y} = 0.83\n", - " [ 1] {x} = 0.83\n", - "\n", - "Method: uds\n", - " [ 1] {normal} = 0.47\n", - "\n", - "Method: mcde\n", - " [ 1] {x} = 0.92\n", - " [ 1] {y} = 0.88\n", - "\n", - "\n", - "/**\n", - " * Data points = 500\n", - " */\n", - "\n", - "Method: tcmi\n", - " [ 2] {y,x} = 0.60\n", - " [ 2] {normal,triangular} = 0.57\n", - " [ 2] {x,normal} = 0.57\n", - " [ 1] {x} = 0.38\n", - "\n", - "Method: cmi\n", - " [ 1] {y} = 1.00\n", - " [ 1] {x} = 1.00\n", - " [ 1] {logistic} = 1.00\n", - " [ 1] {triangular} = 1.00\n", - " [ 1] {laplace} = 1.00\n", - " [ 1] {normal} = 1.00\n", - "\n", - "Method: mac\n", - " [ 1] {y} = 0.81\n", - " [ 1] {weibull} = 0.78\n", - "\n", - "Method: uds\n", - " [ 1] {normal} = 0.45\n", - " [ 1] {logistic} = 0.44\n", - "\n", - "Method: mcde\n", - " [ 1] {y} = 0.94\n", - " [ 1] {x} = 0.92\n", - "\n" - ] - } - ], + "outputs": [], "source": [ "results = {}\n", "for size in sizes:\n", @@ -1538,11 +946,11 @@ }, { "cell_type": "code", - "execution_count": 16, + "execution_count": null, "metadata": { "ExecuteTime": { - "end_time": "2020-01-17T09:14:14.233510Z", - "start_time": "2020-01-17T09:14:14.167295Z" + "end_time": "2020-02-06T20:01:44.841663Z", + "start_time": "2020-02-06T20:01:44.602723Z" }, "code_folding": [ 0, @@ -1551,71 +959,7 @@ "hidden": true, "scrolled": true }, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "\n", - "/**\n", - " * Data set: gaussian\n", - " */\n", - "\n", - "Data points = 500\n", - "Noise levels = 6\n", - "\n", - "Method: tcmi\n", - " 0.00, 0.20, 0.40, 0.60, 0.80, 1.00, \n", - "\n", - "Noise = 0.00, Power = 1.00, Score = 0.55, Score0 = 0.43\n", - "Noise = 0.20, Power = 0.89, Score = 0.50, Score0 = 0.43\n", - "Noise = 0.40, Power = 0.25, Score = 0.44, Score0 = 0.43\n", - "Noise = 0.60, Power = 0.02, Score = 0.39, Score0 = 0.43\n", - "Noise = 0.80, Power = 0.00, Score = 0.36, Score0 = 0.43\n", - "Noise = 1.00, Power = 0.00, Score = 0.34, Score0 = 0.43\n", - "\n", - "Method: cmi\n", - " 0.00, 0.20, 0.40, 0.60, 0.80, 1.00, \n", - "\n", - "Noise = 0.00, Power = 1.00, Score = 0.21, Score0 = 0.01\n", - "Noise = 0.20, Power = 1.00, Score = 0.21, Score0 = 0.01\n", - "Noise = 0.40, Power = 1.00, Score = 0.21, Score0 = 0.01\n", - "Noise = 0.60, Power = 1.00, Score = 0.20, Score0 = 0.01\n", - "Noise = 0.80, Power = 1.00, Score = 0.20, Score0 = 0.01\n", - "Noise = 1.00, Power = 1.00, Score = 0.19, Score0 = 0.01\n", - "\n", - "Method: mac\n", - " 0.00, 0.20, 0.40, 0.60, 0.80, 1.00, \n", - "\n", - "Noise = 0.00, Power = 1.00, Score = 0.55, Score0 = 0.52\n", - "Noise = 0.20, Power = 1.00, Score = 0.55, Score0 = 0.52\n", - "Noise = 0.40, Power = 1.00, Score = 0.54, Score0 = 0.52\n", - "Noise = 0.60, Power = 1.00, Score = 0.53, Score0 = 0.52\n", - "Noise = 0.80, Power = 0.99, Score = 0.53, Score0 = 0.52\n", - "Noise = 1.00, Power = 0.96, Score = 0.53, Score0 = 0.52\n", - "\n", - "Method: uds\n", - " 0.00, 0.20, 0.40, 0.60, 0.80, 1.00, \n", - "\n", - "Noise = 0.00, Power = 0.00, Score = 0.00, Score0 = 0.00\n", - "Noise = 0.20, Power = 0.00, Score = 0.00, Score0 = 0.00\n", - "Noise = 0.40, Power = 0.00, Score = 0.00, Score0 = 0.00\n", - "Noise = 0.60, Power = 0.00, Score = 0.00, Score0 = 0.00\n", - "Noise = 0.80, Power = 0.00, Score = 0.00, Score0 = 0.00\n", - "Noise = 1.00, Power = 0.00, Score = 0.00, Score0 = 0.00\n", - "\n", - "Method: mcde\n", - " 0.00, 0.20, 0.40, 0.60, 0.80, 1.00, \n", - "\n", - "Noise = 0.00, Power = 1.00, Score = 0.90, Score0 = 0.50\n", - "Noise = 0.20, Power = 1.00, Score = 0.90, Score0 = 0.50\n", - "Noise = 0.40, Power = 1.00, Score = 0.89, Score0 = 0.50\n", - "Noise = 0.60, Power = 1.00, Score = 0.88, Score0 = 0.50\n", - "Noise = 0.80, Power = 1.00, Score = 0.87, Score0 = 0.50\n", - "Noise = 1.00, Power = 1.00, Score = 0.85, Score0 = 0.50\n" - ] - } - ], + "outputs": [], "source": [ "def generate_dataset(data, target, noise, n_splits=10, n_repeats=100, seed=None):\n", " \"\"\"Generate data set.\n", @@ -1678,7 +1022,7 @@ "\n", "\n", "datasets = [\n", - " ('Bivariate normal distribution: $\\{x,y\\}$', 'gaussian', 'data/2d_gaussian.csv', \n", + " ('Bivariate normal distribution: $\\{x,y\\}$', 'gaussian', 'data/tcmi/2d_gaussian.csv', \n", " 'Gaussian', ('-|x|', '-|y|'))\n", "]\n", "\n", @@ -1761,29 +1105,16 @@ }, { "cell_type": "code", - "execution_count": 17, + "execution_count": null, "metadata": { "ExecuteTime": { - "end_time": "2020-01-17T09:14:14.715484Z", - "start_time": "2020-01-17T09:14:14.234956Z" + "end_time": "2020-02-06T20:01:45.959653Z", + "start_time": "2020-02-06T20:01:44.869348Z" }, "code_folding": [], "hidden": true }, - "outputs": [ - { - "data": { - "image/png": "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\n", - "text/plain": [ - "<Figure size 1152x216 with 5 Axes>" - ] - }, - "metadata": { - "needs_background": "light" - }, - "output_type": "display_data" - } - ], + "outputs": [], "source": [ "##\n", "# Plot\n", @@ -1934,11 +1265,11 @@ }, { "cell_type": "code", - "execution_count": 18, + "execution_count": null, "metadata": { "ExecuteTime": { - "end_time": "2020-01-17T09:14:14.718967Z", - "start_time": "2020-01-17T09:14:14.716748Z" + "end_time": "2020-02-06T20:01:46.090055Z", + "start_time": "2020-02-06T20:01:45.980293Z" }, "hidden": true }, @@ -1959,36 +1290,15 @@ }, { "cell_type": "code", - "execution_count": 19, + "execution_count": null, "metadata": { "ExecuteTime": { - "end_time": "2020-01-17T09:14:14.745691Z", - "start_time": "2020-01-17T09:14:14.720322Z" + "end_time": "2020-02-06T20:01:46.302638Z", + "start_time": "2020-02-06T20:01:46.105784Z" }, "hidden": true }, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "x1 : r2=0.16 rho= 0.41 pval=2.8e-21\n", - "x2 : r2=0.18 rho= 0.44 pval=5.4e-25\n", - "x3 : r2=0.00 rho=-0.03 pval=0.44\n", - "x4 : r2=0.30 rho= 0.53 pval=4.3e-37\n", - "x5 : r2=0.06 rho= 0.22 pval=1e-06\n", - "x6 : r2=0.00 rho=-0.03 pval=0.44\n", - "x7 : r2=0.01 rho= 0.09 pval=0.057\n", - "x8 : r2=0.02 rho= 0.13 pval=0.0049\n", - "x9 : r2=0.01 rho= 0.09 pval=0.034\n", - "x10: r2=0.00 rho= 0.02 pval=0.59\n", - "x11: r2=0.16 rho= 0.41 pval=2.7e-21\n", - "x12: r2=0.18 rho= 0.44 pval=1.2e-24\n", - "x13: r2=0.00 rho=-0.04 pval=0.39\n", - "x14: r2=0.30 rho= 0.53 pval=1.5e-37\n" - ] - } - ], + "outputs": [], "source": [ "# https://blog.datadive.net/selecting-good-features-part-iv-stability-selection-rfe-and-everything-side-by-side/\n", "# Friedman #1 regression dataset \n", @@ -2048,326 +1358,22 @@ }, { "cell_type": "code", - "execution_count": 20, + "execution_count": null, "metadata": { "ExecuteTime": { - "end_time": "2020-01-17T09:14:14.834041Z", - "start_time": "2020-01-17T09:14:14.747010Z" + "end_time": "2020-02-06T20:01:46.620148Z", + "start_time": "2020-02-06T20:01:46.315591Z" }, "hidden": true, "scrolled": true }, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "\n", - "/**\n", - " * Model: friedman\n", - " */\n", - "\n", - "Keys: {x1,x2,x3,x4,x5,x6,x7,x8,x9,x10,x11,x12,x13,x14,y,-x1,-x10,|x11|,-x11,-|x11|,|x12|,-x12,-|x12|,|x13|,-x13,-|x13|,|x14|,-x14,-|x14|,-x2,-x3,-x4,-x5,-x6,-x7,-x8,-x9}\n", - "Size: 36\n", - "\n", - "Method: tcmi\n", - " [ 4] {|x14|,|x12|,x1,x5} = 0.77\n", - " [ 4] {|x14|,|x12|,x1,-x7} = 0.75\n", - " [ 4] {|x14|,|x12|,x1,x6} = 0.74\n", - " [ 4] {|x14|,|x12|,x1,-x9} = 0.74\n", - " [ 3] {|x14|,|x12|,x1} = 0.69\n", - " [ 3] {|x14|,|x12|,|x11|} = 0.68\n", - " [ 3] {|x14|,|x12|,x11} = 0.68\n", - " [ 3] {|x14|,|x12|,x5} = 0.66\n", - " [ 2] {|x14|,|x12|} = 0.53\n", - " [ 2] {|x14|,x12} = 0.53\n", - " [ 2] {|x14|,x2} = 0.52\n", - " [ 2] {|x14|,x5} = 0.52\n", - " [ 2] {|x14|,-x8} = 0.51\n", - " [ 1] {|x14|} = 0.35\n", - " [ 1] {x14} = 0.35\n", - " [ 1] {x4} = 0.34\n", - "\n", - "Method: cmi\n", - " [ 1] {-x14} = 1.00\n", - " [ 1] {-x4} = 1.00\n", - " [ 1] {-|x14|} = 1.00\n", - " [ 1] {x14} = 1.00\n", - " [ 1] {x4} = 1.00\n", - " [ 1] {|x14|} = 1.00\n", - "\n", - "Method: mac\n", - " More than 119982 subsets.\n", - "\n", - "Method: uds\n", - " [ 1] {-x1} = 0.00\n", - "\n", - "Method: mcde\n", - " [ 1] {x2} = 0.78\n", - " [ 1] {-x2} = 0.77\n", - " [ 1] {x12} = 0.77\n", - " [ 1] {-x12} = 0.77\n", - " [ 1] {x11} = 0.77\n", - " [ 1] {|x11|} = 0.77\n", - " [ 1] {|x12|} = 0.77\n", - " [ 1] {-|x12|} = 0.77\n", - " [ 1] {x1} = 0.77\n", - " [ 1] {-x11} = 0.76\n", - " [ 1] {-|x11|} = 0.76\n", - " [ 1] {-x1} = 0.76\n", - "\n", - "/**\n", - " * Model: concrete\n", - " */\n", - "\n", - "Keys: {cement,slag,fly_ash,water,sp,coarse_aggr,fine_aggr,age,compressive_strength,-age,-cement,-coarse_aggr,-fine_aggr,-fly_ash,-slag,-sp,-water}\n", - "Size: 16\n", - "\n", - "Method: tcmi\n", - " [ 5] {cement,-sp,-water,-coarse_aggr,-fine_aggr} = 0.68\n", - " [ 5] {-fine_aggr,-water,-sp,-coarse_aggr,-fly_ash} = 0.68\n", - " [ 5] {-fine_aggr,-water,-sp,-coarse_aggr,age} = 0.68\n", - " [ 5] {cement,coarse_aggr,-water,-slag,-fine_aggr} = 0.68\n", - " [ 5] {-fine_aggr,-slag,-water,-coarse_aggr,age} = 0.67\n", - " [ 5] {cement,coarse_aggr,-water,-sp,age} = 0.67\n", - " [ 5] {cement,coarse_aggr,fine_aggr,-sp,age} = 0.67\n", - " [ 5] {coarse_aggr,-cement,-fine_aggr,-water,-sp} = 0.66\n", - " [ 4] {-fine_aggr,-water,-sp,-coarse_aggr} = 0.66\n", - " [ 4] {cement,coarse_aggr,fine_aggr,-sp} = 0.64\n", - " [ 4] {cement,-sp,-water,fine_aggr} = 0.64\n", - " [ 4] {cement,coarse_aggr,-water,-sp} = 0.63\n", - " [ 4] {cement,-sp,fine_aggr,-coarse_aggr} = 0.63\n", - " [ 4] {-fine_aggr,-slag,-water,-coarse_aggr} = 0.63\n", - " [ 3] {cement,-sp,-water} = 0.57\n", - " [ 3] {cement,-sp,fine_aggr} = 0.57\n", - " [ 3] {cement,fly_ash,slag} = 0.57\n", - " [ 3] {cement,coarse_aggr,-water} = 0.56\n", - " [ 3] {-fine_aggr,-slag,-water} = 0.56\n", - " [ 3] {-fine_aggr,-water,-sp} = 0.56\n", - " [ 3] {-water,-sp,-coarse_aggr} = 0.55\n", - " [ 3] {cement,coarse_aggr,fine_aggr} = 0.55\n", - " [ 3] {cement,water,sp} = 0.55\n", - " [ 2] {cement,fly_ash} = 0.43\n", - " [ 2] {-water,-sp} = 0.43\n", - " [ 2] {coarse_aggr,slag} = 0.41\n", - " [ 1] {cement} = 0.30\n", - "\n", - "Method: cmi\n", - " [ 1] {-age} = 1.00\n", - " [ 1] {-cement} = 1.00\n", - " [ 1] {-coarse_aggr} = 1.00\n", - " [ 1] {-fine_aggr} = 1.00\n", - " [ 1] {-slag} = 1.00\n", - " [ 1] {-water} = 1.00\n", - " [ 1] {age} = 1.00\n", - " [ 1] {cement} = 1.00\n", - " [ 1] {water} = 1.00\n", - " [ 1] {-sp} = 0.98\n", - "\n", - "Method: mac\n", - " [ 6] {water,coarse_aggr,fine_aggr,-cement,-sp,-slag} = 0.74\n", - " [ 6] {cement,coarse_aggr,fine_aggr,water,-sp,-slag} = 0.74\n", - " [ 6] {cement,coarse_aggr,fine_aggr,water,-sp,slag} = 0.74\n", - " [ 5] {water,coarse_aggr,fine_aggr,-cement,-sp} = 0.73\n", - " [ 5] {water,coarse_aggr,fine_aggr,-cement,sp} = 0.73\n", - " [ 5] {cement,coarse_aggr,fine_aggr,water,-sp} = 0.72\n", - " [ 5] {cement,coarse_aggr,fine_aggr,water,-slag} = 0.72\n", - " [ 5] {cement,coarse_aggr,fine_aggr,water,sp} = 0.72\n", - " [ 5] {cement,coarse_aggr,fine_aggr,water,-fly_ash} = 0.72\n", - " [ 5] {cement,coarse_aggr,fine_aggr,water,fly_ash} = 0.72\n", - " [ 5] {cement,coarse_aggr,fine_aggr,water,slag} = 0.71\n", - " [ 4] {water,coarse_aggr,fine_aggr,-cement} = 0.70\n", - " [ 4] {cement,coarse_aggr,fine_aggr,water} = 0.70\n", - " [ 4] {cement,coarse_aggr,fine_aggr,-water} = 0.69\n", - " [ 3] {water,coarse_aggr,fine_aggr} = 0.61\n", - " [ 3] {cement,coarse_aggr,fine_aggr} = 0.60\n", - " [ 3] {cement,coarse_aggr,-fine_aggr} = 0.60\n", - " [ 3] {cement,coarse_aggr,water} = 0.59\n", - " [ 3] {cement,coarse_aggr,-water} = 0.59\n", - " [ 2] {coarse_aggr,fine_aggr} = 0.44\n", - " [ 2] {-coarse_aggr,fine_aggr} = 0.44\n", - " [ 2] {-coarse_aggr,-fine_aggr} = 0.44\n", - " [ 2] {water,coarse_aggr} = 0.42\n", - " [ 2] {water,fine_aggr} = 0.42\n", - " [ 2] {water,-coarse_aggr} = 0.42\n", - " [ 2] {cement,coarse_aggr} = 0.42\n", - " [ 2] {cement,fine_aggr} = 0.42\n", - " [ 2] {cement,-coarse_aggr} = 0.42\n", - " [ 1] {-age} = 0.25\n", - " [ 1] {age} = 0.25\n", - "\n", - "Method: uds\n", - " [ 1] {-age} = 0.00\n", - "\n", - "Method: mcde\n", - " [ 1] {-age} = 0.90\n", - " [ 1] {age} = 0.89\n", - "\n", - "/**\n", - " * Model: forest_fires\n", - " */\n", - "\n", - "Keys: {X,Y,month,day,FFMC,DMC,DC,ISI,temp,RH,wind,rain,area,-DC,-DMC,-FFMC,-ISI,-RH,-X,-Y,-day,-month,-rain,-temp,-wind}\n", - "Size: 24\n", - "\n", - "Method: tcmi\n", - " [ 6] {DMC,-RH,ISI,-temp,-wind,-DC} = 0.53\n", - " [ 6] {DMC,-RH,-DC,-temp,FFMC,-wind} = 0.51\n", - " [ 5] {temp,RH,-day,-DC,-FFMC} = 0.44\n", - " [ 5] {temp,RH,-day,-DC,-ISI} = 0.43\n", - " [ 5] {temp,RH,-day,-DC,-DMC} = 0.43\n", - " [ 5] {temp,RH,-day,-FFMC,wind} = 0.43\n", - " [ 5] {temp,RH,-day,-FFMC,X} = 0.43\n", - " [ 5] {temp,RH,-FFMC,-DC,wind} = 0.42\n", - " [ 5] {temp,RH,-FFMC,-DC,X} = 0.42\n", - " [ 5] {temp,RH,-day,-ISI,wind} = 0.42\n", - " [ 5] {temp,RH,-day,-FFMC,-month} = 0.42\n", - " [ 5] {temp,RH,-FFMC,X,-DMC} = 0.42\n", - " [ 5] {temp,RH,-DC,-ISI,wind} = 0.42\n", - " [ 5] {DMC,-RH,-DC,-temp,FFMC} = 0.42\n", - " [ 5] {temp,RH,-day,-FFMC,-DMC} = 0.42\n", - " [ 5] {DMC,-RH,ISI,-temp,-wind} = 0.42\n", - " [ 5] {DMC,-RH,ISI,-temp,-DC} = 0.42\n", - " [ 4] {temp,RH,-day,-DC} = 0.41\n", - " [ 4] {temp,RH,-day,-FFMC} = 0.40\n", - " [ 4] {temp,RH,-day,-DMC} = 0.39\n", - " [ 4] {temp,RH,-FFMC,-DC} = 0.39\n", - " [ 4] {temp,RH,-FFMC,X} = 0.39\n", - " [ 3] {temp,RH,-day} = 0.35\n", - " [ 3] {temp,RH,-FFMC} = 0.33\n", - " [ 2] {DMC,-RH} = 0.26\n", - " [ 1] {temp} = 0.12\n", - " [ 1] {DMC} = 0.11\n", - "\n", - "Method: cmi\n", - " [ 2] {temp,-DC} = 1.00\n", - " [ 2] {temp,-DMC} = 1.00\n", - " [ 2] {temp,-RH} = 1.00\n", - " [ 2] {temp,DC} = 1.00\n", - " [ 2] {temp,DMC} = 1.00\n", - " [ 2] {temp,FFMC} = 1.00\n", - " [ 2] {temp,RH} = 1.00\n", - " [ 2] {FFMC,-DC} = 1.00\n", - " [ 2] {FFMC,-DMC} = 1.00\n", - " [ 2] {FFMC,-RH} = 1.00\n", - " [ 2] {FFMC,-temp} = 1.00\n", - " [ 2] {FFMC,DC} = 1.00\n", - " [ 2] {FFMC,DMC} = 1.00\n", - " [ 2] {FFMC,RH} = 1.00\n", - " [ 2] {-temp,-DC} = 1.00\n", - " [ 2] {-temp,-DMC} = 1.00\n", - " [ 2] {-temp,-RH} = 1.00\n", - " [ 2] {-temp,DC} = 1.00\n", - " [ 2] {-temp,DMC} = 1.00\n", - " [ 2] {-temp,RH} = 1.00\n", - " [ 2] {DMC,-DC} = 1.00\n", - " [ 2] {DMC,-ISI} = 1.00\n", - " [ 2] {DMC,-RH} = 1.00\n", - " [ 2] {DMC,DC} = 1.00\n", - " [ 2] {DMC,ISI} = 1.00\n", - " [ 2] {DMC,month} = 1.00\n", - " [ 2] {-DC,-DMC} = 1.00\n", - " [ 2] {-DC,-ISI} = 1.00\n", - " [ 2] {-DC,-RH} = 1.00\n", - " [ 2] {-DC,-month} = 1.00\n", - " [ 2] {-DC,month} = 1.00\n", - " [ 2] {-RH,-DMC} = 1.00\n", - " [ 2] {-RH,DC} = 1.00\n", - " [ 2] {-ISI,-DMC} = 1.00\n", - " [ 2] {-ISI,DC} = 1.00\n", - " [ 2] {RH,-DMC} = 1.00\n", - " [ 2] {DC,-DMC} = 1.00\n", - " [ 2] {DC,-month} = 1.00\n", - " [ 2] {DC,month} = 1.00\n", - " [ 2] {-DMC,ISI} = 1.00\n", - " [ 2] {-DMC,month} = 1.00\n", - " [ 2] {temp,month} = 1.00\n", - " [ 1] {temp} = 0.40\n", - "\n", - "Method: mac\n", - " [ 9] {-temp,-RH,-DMC,FFMC,DC,-ISI,-wind,-day,-X} = 0.85\n", - " [ 9] {-temp,-RH,-DMC,FFMC,DC,-ISI,-wind,-day,X} = 0.85\n", - " [ 9] {-temp,-RH,-DMC,FFMC,DC,-ISI,-wind,day,-X} = 0.85\n", - " [ 9] {-temp,-RH,-DMC,FFMC,DC,-ISI,-wind,day,X} = 0.85\n", - " [ 9] {-temp,-RH,-DMC,-DC,FFMC,-ISI,-wind,-day,-X} = 0.85\n", - " [ 9] {-temp,-RH,-DMC,-DC,FFMC,-ISI,-wind,-day,X} = 0.85\n", - " [ 9] {-temp,-RH,-DMC,-DC,FFMC,-ISI,-wind,day,-X} = 0.85\n", - " [ 9] {-temp,-RH,-DMC,-DC,FFMC,-ISI,-wind,day,X} = 0.85\n", - " [ 8] {-temp,-RH,-DMC,FFMC,DC,-ISI,-wind,-day} = 0.83\n", - " [ 8] {-temp,-RH,-DMC,FFMC,DC,-ISI,-wind,day} = 0.83\n", - " [ 8] {-temp,-RH,-DMC,-DC,FFMC,-ISI,-wind,-day} = 0.83\n", - " [ 8] {-temp,-RH,-DMC,-DC,FFMC,-ISI,-wind,day} = 0.83\n", - " [ 8] {-temp,-RH,-DMC,-DC,FFMC,-ISI,-wind,-X} = 0.83\n", - " [ 8] {-temp,-RH,-DMC,-DC,FFMC,-ISI,-wind,X} = 0.83\n", - " [ 7] {-temp,-RH,-DMC,FFMC,DC,-ISI,-wind} = 0.82\n", - " [ 7] {-temp,-RH,-DMC,-DC,FFMC,-ISI,-wind} = 0.82\n", - " [ 7] {-temp,-RH,-DMC,-DC,FFMC,-ISI,wind} = 0.81\n", - " [ 7] {-temp,-RH,-DMC,-DC,FFMC,-ISI,-X} = 0.81\n", - " [ 7] {-temp,-RH,-DMC,-DC,FFMC,-ISI,X} = 0.81\n", - " [ 6] {-temp,-RH,-DMC,-DC,FFMC,-ISI} = 0.79\n", - " [ 6] {-temp,-RH,-DMC,FFMC,DC,-ISI} = 0.79\n", - " [ 6] {-temp,-RH,-DMC,-DC,FFMC,-wind} = 0.78\n", - " [ 6] {-temp,-RH,-DMC,-DC,FFMC,-day} = 0.78\n", - " [ 6] {-temp,-RH,-DMC,-DC,FFMC,day} = 0.78\n", - " [ 6] {-temp,-RH,-DMC,-DC,FFMC,wind} = 0.78\n", - " [ 6] {-temp,-RH,-DMC,-DC,FFMC,-X} = 0.78\n", - " [ 6] {-temp,-RH,-DMC,-DC,FFMC,X} = 0.78\n", - " [ 5] {FFMC,ISI,DC,-day,-DMC} = 0.77\n", - " [ 5] {FFMC,ISI,DC,day,-DMC} = 0.77\n", - " [ 5] {FFMC,ISI,DC,-day,DMC} = 0.77\n", - " [ 5] {-temp,-DMC,-DC,-ISI,FFMC} = 0.75\n", - " [ 5] {-temp,-DMC,-DC,-ISI,-FFMC} = 0.75\n", - " [ 5] {-temp,-RH,-DMC,ISI,DC} = 0.75\n", - " [ 5] {-temp,-RH,-DMC,-ISI,DC} = 0.75\n", - " [ 5] {-temp,-RH,-DMC,FFMC,DC} = 0.75\n", - " [ 5] {-temp,-RH,-DMC,-DC,FFMC} = 0.75\n", - " [ 5] {-temp,-RH,-DMC,-FFMC,DC} = 0.75\n", - " [ 5] {-temp,-RH,-DMC,-DC,-FFMC} = 0.75\n", - " [ 5] {-temp,-RH,-DMC,-DC,-ISI} = 0.75\n", - " [ 5] {-temp,-RH,-DMC,-DC,ISI} = 0.75\n", - " [ 4] {FFMC,ISI,-DMC,-DC} = 0.74\n", - " [ 3] {FFMC,-DC,-DMC} = 0.63\n", - " [ 3] {FFMC,DC,-ISI} = 0.63\n", - " [ 3] {FFMC,ISI,DC} = 0.63\n", - " [ 3] {FFMC,ISI,-DMC} = 0.63\n", - " [ 3] {FFMC,ISI,-DC} = 0.63\n", - " [ 3] {FFMC,ISI,DMC} = 0.62\n", - " [ 2] {-DMC,-DC} = 0.46\n", - " [ 2] {FFMC,ISI} = 0.45\n", - " [ 1] {-temp} = 0.15\n", - " [ 1] {FFMC} = 0.15\n", - " [ 1] {-DMC} = 0.15\n", - " [ 1] {-ISI} = 0.15\n", - " [ 1] {temp} = 0.14\n", - " [ 1] {ISI} = 0.14\n", - "\n", - "Method: uds\n", - " [ 1] {-rain} = 0.35\n", - "\n", - "Method: mcde\n", - " [ 3] {-DMC,temp,RH} = 0.84\n", - " [ 3] {-DMC,temp,-RH} = 0.83\n", - " [ 3] {-DMC,temp,-DC} = 0.82\n", - " [ 3] {-DMC,temp,DC} = 0.81\n", - " [ 3] {-DMC,temp,-FFMC} = 0.81\n", - " [ 3] {-DMC,temp,FFMC} = 0.81\n", - " [ 2] {temp,RH} = 0.81\n", - " [ 2] {temp,-RH} = 0.80\n", - " [ 2] {-DMC,temp} = 0.79\n", - " [ 2] {-DMC,-temp} = 0.79\n", - " [ 2] {-DMC,-DC} = 0.79\n", - " [ 2] {-DMC,DC} = 0.79\n", - " [ 1] {-DMC} = 0.69\n", - " [ 1] {DMC} = 0.69\n" - ] - } - ], + "outputs": [], "source": [ "# Load datasets\n", "datasets = [\n", - " ('friedman', 'data/friedman.csv', 'y'),\n", - " ('concrete', 'data/concrete.csv', 'compressive_strength'),\n", - " ('forest_fires', 'data/forestfires.csv', 'area')\n", + " ('friedman', 'data/tcmi/friedman.csv', 'y'),\n", + " ('concrete', 'data/tcmi/concrete.csv', 'compressive_strength'),\n", + " ('forest_fires', 'data/tcmi/forestfires.csv', 'area')\n", "]\n", "\n", "for name, filename, target in datasets:\n", @@ -2462,98 +1468,18 @@ }, { "cell_type": "code", - "execution_count": 3, + "execution_count": null, "metadata": { "ExecuteTime": { - "end_time": "2020-01-17T09:32:40.231624Z", - "start_time": "2020-01-17T09:32:40.167835Z" + "end_time": "2020-02-06T20:01:46.843772Z", + "start_time": "2020-02-06T20:01:46.664345Z" }, - "scrolled": true + "scrolled": false }, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "\n", - "/**\n", - " * Model: Octet-binary compound semiconductor\n", - " */\n", - "\n", - "Keys: {Delta E,EA(A),EN(A),H(A),IP(A),L(A),rd(A),rp(A),rs(A),EA(B),EN(B),H(B),IP(B),L(B),rd(B),rp(B),rs(B),|EA(A)|,-EA(A),-|EA(A)|,|EA(B)|,-EN(A),-EN(B),|H(A)|,|H(B)|,|IP(A)|,|IP(B)|,|L(A)|,-L(A),-|L(A)|,-L(B),-rd(A),-rd(B),-rp(A),-rp(B),-rs(A),-rs(B)}\n", - "Size: 36\n", - "\n", - "Method: tcmi\n", - " [ 6] {rs(A),L(B),rp(A),EA(A),IP(B),rs(B)} = 0.82\n", - " [ 6] {rs(A),L(B),rp(A),EA(A),IP(B),rp(B)} = 0.82\n", - " [ 6] {rs(A),L(B),rp(A),EA(A),EN(B),rs(B)} = 0.82\n", - " [ 6] {rs(A),L(B),rp(A),EA(A),H(B),rs(B)} = 0.82\n", - " [ 6] {rp(A),L(B),EA(A),H(A),rd(A),rd(B)} = 0.82\n", - " [ 6] {rs(A),L(B),rp(A),EA(A),EN(B),rp(B)} = 0.82\n", - " [ 6] {rs(A),L(B),rp(A),EA(A),H(B),rp(B)} = 0.82\n", - " [ 6] {rs(A),L(B),EA(A),H(A),rd(A),rd(B)} = 0.82\n", - " [ 6] {rp(A),L(B),EA(A),H(A),rp(B),rd(B)} = 0.81\n", - " [ 6] {rp(A),L(B),EA(A),H(A),rs(B),rd(B)} = 0.81\n", - " [ 6] {rp(A),L(B),EA(A),H(A),rd(A),EA(B)} = 0.80\n", - " [ 6] {rs(A),L(B),EA(A),H(A),rd(A),IP(B)} = 0.80\n", - " [ 6] {rp(A),L(B),EA(A),H(A),rd(A),IP(B)} = 0.80\n", - " [ 6] {rs(A),L(B),EA(A),H(A),rd(A),EA(B)} = 0.80\n", - " [ 6] {rs(A),L(B),EA(A),H(A),L(A),EA(B)} = 0.80\n", - " [ 6] {rs(A),L(B),rp(A),EA(A),EN(B),EN(A)} = 0.80\n", - " [ 6] {rs(A),L(B),rp(A),EA(A),IP(B),EN(A)} = 0.80\n", - " [ 6] {rs(A),L(B),rp(A),EN(A),IP(B),rd(A)} = 0.80\n", - " [ 6] {rs(A),L(B),rp(A),EA(A),H(B),EN(A)} = 0.80\n", - " [ 6] {rs(A),L(B),EA(A),H(A),L(A),IP(A)} = 0.79\n", - " [ 6] {rp(A),L(B),EA(A),H(A),rs(B),EA(B)} = 0.79\n", - " [ 6] {rs(A),L(B),EA(A),H(A),L(A),IP(B)} = 0.79\n", - " [ 6] {rp(A),L(B),EA(A),H(A),rp(B),EA(B)} = 0.79\n", - " [ 5] {rs(A),L(B),rp(A),EA(A),IP(B)} = 0.79\n", - " [ 5] {rs(A),L(B),rp(A),EA(A),H(B)} = 0.79\n", - " [ 5] {rs(A),L(B),rp(A),EA(A),EN(B)} = 0.79\n", - " [ 5] {rs(A),L(B),EA(A),H(A),L(A)} = 0.78\n", - " [ 5] {rs(A),L(B),rp(A),EN(A),IP(B)} = 0.78\n", - " [ 5] {rs(A),L(B),EA(A),H(A),rd(A)} = 0.78\n", - " [ 5] {rp(A),L(B),EA(A),H(A),rd(A)} = 0.78\n", - " [ 5] {rp(A),L(B),EA(A),H(A),rs(B)} = 0.78\n", - " [ 5] {rp(A),L(B),EA(A),H(A),rp(B)} = 0.78\n", - " [ 4] {rs(A),L(B),rp(A),EA(A)} = 0.78\n", - " [ 4] {rs(A),L(B),rp(A),EN(A)} = 0.76\n", - " [ 4] {rs(A),rp(B),rp(A),L(A)} = 0.76\n", - " [ 4] {rs(A),rs(B),rp(A),L(A)} = 0.76\n", - " [ 4] {rp(A),EN(A),L(B),IP(B)} = 0.75\n", - " [ 4] {rs(A),L(B),rp(A),IP(B)} = 0.75\n", - " [ 4] {rs(A),L(B),EA(A),H(A)} = 0.75\n", - " [ 4] {rs(A),L(B),rp(A),H(B)} = 0.75\n", - " [ 4] {rs(A),L(B),rp(A),EN(B)} = 0.75\n", - " [ 4] {rp(A),L(B),EA(A),H(A)} = 0.75\n", - " [ 4] {rs(A),L(B),rp(A),IP(A)} = 0.74\n", - " [ 4] {rp(A),L(B),EA(A),rd(A)} = 0.74\n", - " [ 4] {rp(A),EN(A),L(B),EN(B)} = 0.74\n", - " [ 3] {rp(A),EN(A),L(B)} = 0.73\n", - " [ 3] {rp(A),L(B),IP(A)} = 0.73\n", - " [ 3] {rs(A),rs(B),rp(A)} = 0.73\n", - " [ 3] {rs(A),L(B),rp(A)} = 0.73\n", - " [ 3] {rs(A),rp(B),rp(A)} = 0.72\n", - " [ 3] {rs(A),L(B),IP(A)} = 0.72\n", - " [ 3] {rs(A),L(B),EN(A)} = 0.72\n", - " [ 3] {rs(A),L(B),EA(A)} = 0.71\n", - " [ 3] {rp(A),L(B),EA(A)} = 0.71\n", - " [ 3] {rs(A),L(B),rs(B)} = 0.71\n", - " [ 3] {rs(A),L(B),rp(B)} = 0.71\n", - " [ 3] {rs(A),rs(B),IP(A)} = 0.70\n", - " [ 3] {rp(A),rs(B),EN(A)} = 0.70\n", - " [ 3] {rp(A),rs(B),IP(A)} = 0.70\n", - " [ 2] {rs(A),L(B)} = 0.69\n", - " [ 2] {rs(A),rs(B)} = 0.67\n", - " [ 2] {rs(A),rp(B)} = 0.66\n", - " [ 1] {rs(A)} = 0.56\n", - " [ 1] {rp(A)} = 0.55\n" - ] - } - ], + "outputs": [], "source": [ "methods = ['tcmi', 'cmi', 'mac', 'uds', 'mcde']\n", - "data = pd.read_csv('data/octet-binary-compound-semiconductors.csv', low_memory=False)\n", + "data = pd.read_csv('data/tcmi/octet-binary-compound-semiconductors.csv', low_memory=False)\n", "materials = data.drop(columns='Combination', inplace=True)\n", "target = 'Delta E'\n", "\n", @@ -2618,13 +1544,6 @@ " key=lambda x: (x[-2], x))\n", "candidates.insert(0, (keys, 1))" ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] } ], "metadata": { @@ -2644,7 +1563,7 @@ "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", - "version": "3.6.10" + "version": "3.7.3" }, "toc": { "base_numbering": "0",