Meta-Features.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/howl/miniconda3/envs/Tesis/lib/python3.7/site-packages/sklearn/utils/__init__.py:4: DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated, and in 3.8 it will stop working\n",
      "  from collections import Sequence\n"
     ]
    }
   ],
   "source": [
    "from project.meta_feature import MetaFeature\n",
    "from project.meta_feature import MetaFeatureExtractor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "meta_features = [\n",
    "    MetaFeature(\"X\", \"num_obs\", \"none\"),\n",
    "    MetaFeature(\"X\", \"num_attrs\", \"none\"),\n",
    "    MetaFeature(\"y\", \"num_classes\", \"none\"),\n",
    "    MetaFeature(\"X\", \"ratio_obs_attrs\", \"none\"),\n",
    "    \n",
    "    MetaFeature(\"y\", \"norm_class_entropy\", \"none\"),    \n",
    "    \n",
    "    MetaFeature(\"X\", \"stand_dev\", \"min\"),\n",
    "    MetaFeature(\"X\", \"stand_dev\", \"mean\"),\n",
    "    MetaFeature(\"X\", \"stand_dev\", \"max\"),\n",
    "    \n",
    "    MetaFeature(\"X_minmax\", \"stand_dev\", \"min\"),\n",
    "    MetaFeature(\"X_minmax\", \"stand_dev\", \"mean\"),\n",
    "    MetaFeature(\"X_minmax\", \"stand_dev\", \"max\"),\n",
    "    \n",
    "    MetaFeature(\"X\", \"var_coef\", \"min\"),\n",
    "    MetaFeature(\"X\", \"var_coef\", \"mean\"),    \n",
    "    MetaFeature(\"X\", \"var_coef\", \"max\"),\n",
    "    \n",
    "    MetaFeature(\"X_minmax\", \"var_coef\", \"min\"),\n",
    "    MetaFeature(\"X_minmax\", \"var_coef\", \"mean\"),\n",
    "    MetaFeature(\"X_minmax\", \"var_coef\", \"max\"),\n",
    "    \n",
    "    MetaFeature(\"X\", \"covariance\", \"min\"),\n",
    "    MetaFeature(\"X\", \"covariance\", \"mean\"),\n",
    "    MetaFeature(\"X\", \"covariance\", \"max\"),\n",
    "    \n",
    "    MetaFeature(\"X_std\", \"covariance\", \"min\"),\n",
    "    MetaFeature(\"X_std\", \"covariance\", \"mean\"),\n",
    "    MetaFeature(\"X_std\", \"covariance\", \"max\"),\n",
    "    \n",
    "    MetaFeature(\"X_minmax\", \"covariance\", \"min\"),\n",
    "    MetaFeature(\"X_minmax\", \"covariance\", \"mean\"),\n",
    "    MetaFeature(\"X_minmax\", \"covariance\", \"max\"),    \n",
    "    \n",
    "    MetaFeature(\"X\", \"correlation\", \"min\"),\n",
    "    MetaFeature(\"X\", \"correlation\", \"mean\"),\n",
    "    MetaFeature(\"X\", \"correlation\", \"max\"),\n",
    "    \n",
    "    # Valores similares a los obtenidos a utilizar el objeto X\n",
    "    # MetaFeature(\"X_std\", \"correlation\", \"mean\"),  \n",
    "    # MetaFeature(\"X_minmax\", \"correlation\", \"mean\"),\n",
    "    \n",
    "    MetaFeature(\"X\", \"skewness\", \"min\"),\n",
    "    MetaFeature(\"X\", \"skewness\", \"mean\"),\n",
    "    MetaFeature(\"X\", \"skewness\", \"max\"),\n",
    "    \n",
    "    # Valores similares a los obtenidos a utilizar el objeto X\n",
    "    # MetaFeature(\"X_std\", \"skewness\", \"mean\"),\n",
    "    # MetaFeature(\"X_minmax\", \"skewness\", \"mean\"),    \n",
    "    \n",
    "    MetaFeature(\"X\", \"kurtosis\", \"min\"),\n",
    "    MetaFeature(\"X\", \"kurtosis\", \"mean\"),\n",
    "    MetaFeature(\"X\", \"kurtosis\", \"max\"),\n",
    "    \n",
    "    # Valores similares a los obtenidos a utilizar el objeto X    \n",
    "    #MetaFeature(\"X_std\", \"kurtosis\", \"mean\"),\n",
    "    #MetaFeature(\"X_minmax\", \"kurtosis\", \"mean\"),\n",
    "        \n",
    "    MetaFeature(\"X\", \"exp_var\", \"max\"),\n",
    "    MetaFeature(\"X_std\", \"exp_var\", \"max\"),\n",
    "    MetaFeature(\"X_minmax\", \"exp_var\", \"max\"),\n",
    "    \n",
    "    MetaFeature(\"X\", \"exp_var\", \"n_t80_cumsum\"),\n",
    "    #MetaFeature(\"X\", \"exp_var\", \"n_t90_cumsum\"),\n",
    "    ]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<project.meta_feature.extractor.MFRCollection at 0x7f2d495f7e10>"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "extractor = MetaFeatureExtractor(meta_features)\n",
    "\n",
    "extractor.run(\"ma\", reset=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "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>X_correlation_max</th>\n",
       "      <th>X_correlation_mean</th>\n",
       "      <th>X_correlation_min</th>\n",
       "      <th>X_covariance_max</th>\n",
       "      <th>X_covariance_mean</th>\n",
       "      <th>X_covariance_min</th>\n",
       "      <th>X_exp_var_max</th>\n",
       "      <th>X_exp_var_n_t80_cumsum</th>\n",
       "      <th>X_kurtosis_max</th>\n",
       "      <th>X_kurtosis_mean</th>\n",
       "      <th>...</th>\n",
       "      <th>X_stand_dev_min</th>\n",
       "      <th>X_std_covariance_max</th>\n",
       "      <th>X_std_covariance_mean</th>\n",
       "      <th>X_std_covariance_min</th>\n",
       "      <th>X_std_exp_var_max</th>\n",
       "      <th>X_var_coef_max</th>\n",
       "      <th>X_var_coef_mean</th>\n",
       "      <th>X_var_coef_min</th>\n",
       "      <th>y_norm_class_entropy_none</th>\n",
       "      <th>y_num_classes_none</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>dataset</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>dataset_prostate_singh</th>\n",
       "      <td>0.994905</td>\n",
       "      <td>0.001368</td>\n",
       "      <td>-0.989840</td>\n",
       "      <td>6.501921e+06</td>\n",
       "      <td>29.654376</td>\n",
       "      <td>-1.662059e+06</td>\n",
       "      <td>0.538233</td>\n",
       "      <td>3</td>\n",
       "      <td>96.809515</td>\n",
       "      <td>3.634957</td>\n",
       "      <td>...</td>\n",
       "      <td>45.069954</td>\n",
       "      <td>0.994905</td>\n",
       "      <td>0.001368</td>\n",
       "      <td>-0.989840</td>\n",
       "      <td>0.635789</td>\n",
       "      <td>24.644940</td>\n",
       "      <td>0.968532</td>\n",
       "      <td>0.393882</td>\n",
       "      <td>0.999723</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>dataset_glioma_phillips</th>\n",
       "      <td>0.993589</td>\n",
       "      <td>0.026853</td>\n",
       "      <td>-0.690733</td>\n",
       "      <td>4.323435e+08</td>\n",
       "      <td>245013.502247</td>\n",
       "      <td>-6.592495e+07</td>\n",
       "      <td>0.311808</td>\n",
       "      <td>16</td>\n",
       "      <td>94.998000</td>\n",
       "      <td>6.624304</td>\n",
       "      <td>...</td>\n",
       "      <td>849.046524</td>\n",
       "      <td>0.993589</td>\n",
       "      <td>0.026853</td>\n",
       "      <td>-0.690733</td>\n",
       "      <td>0.144017</td>\n",
       "      <td>6.143025</td>\n",
       "      <td>0.497883</td>\n",
       "      <td>0.196919</td>\n",
       "      <td>0.795040</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>dataset_leukemia_armstrong</th>\n",
       "      <td>0.994280</td>\n",
       "      <td>0.084879</td>\n",
       "      <td>-0.916524</td>\n",
       "      <td>1.154806e+08</td>\n",
       "      <td>556524.637475</td>\n",
       "      <td>-7.348629e+07</td>\n",
       "      <td>0.249095</td>\n",
       "      <td>19</td>\n",
       "      <td>64.783987</td>\n",
       "      <td>3.472936</td>\n",
       "      <td>...</td>\n",
       "      <td>1480.763026</td>\n",
       "      <td>0.994280</td>\n",
       "      <td>0.084879</td>\n",
       "      <td>-0.916524</td>\n",
       "      <td>0.241996</td>\n",
       "      <td>2.250262</td>\n",
       "      <td>0.511498</td>\n",
       "      <td>0.239257</td>\n",
       "      <td>0.991532</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>dataset_breast_sotiriou</th>\n",
       "      <td>0.997310</td>\n",
       "      <td>0.703431</td>\n",
       "      <td>-0.803858</td>\n",
       "      <td>6.244864e+00</td>\n",
       "      <td>1.451080</td>\n",
       "      <td>-5.063628e+00</td>\n",
       "      <td>0.721969</td>\n",
       "      <td>3</td>\n",
       "      <td>10.362369</td>\n",
       "      <td>-0.156562</td>\n",
       "      <td>...</td>\n",
       "      <td>1.308462</td>\n",
       "      <td>0.997310</td>\n",
       "      <td>0.703431</td>\n",
       "      <td>-0.803858</td>\n",
       "      <td>0.745195</td>\n",
       "      <td>0.389483</td>\n",
       "      <td>0.202296</td>\n",
       "      <td>0.182322</td>\n",
       "      <td>0.993717</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>dataset_lymphoma_dave_2</th>\n",
       "      <td>0.985847</td>\n",
       "      <td>0.019387</td>\n",
       "      <td>-0.739594</td>\n",
       "      <td>5.893252e+00</td>\n",
       "      <td>0.027503</td>\n",
       "      <td>-3.396945e+00</td>\n",
       "      <td>0.124228</td>\n",
       "      <td>93</td>\n",
       "      <td>35.139636</td>\n",
       "      <td>1.273938</td>\n",
       "      <td>...</td>\n",
       "      <td>0.610939</td>\n",
       "      <td>0.985847</td>\n",
       "      <td>0.019387</td>\n",
       "      <td>-0.739594</td>\n",
       "      <td>0.140919</td>\n",
       "      <td>0.525447</td>\n",
       "      <td>0.158073</td>\n",
       "      <td>0.086560</td>\n",
       "      <td>0.882052</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 39 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                            X_correlation_max  X_correlation_mean  \\\n",
       "dataset                                                             \n",
       "dataset_prostate_singh               0.994905            0.001368   \n",
       "dataset_glioma_phillips              0.993589            0.026853   \n",
       "dataset_leukemia_armstrong           0.994280            0.084879   \n",
       "dataset_breast_sotiriou              0.997310            0.703431   \n",
       "dataset_lymphoma_dave_2              0.985847            0.019387   \n",
       "\n",
       "                            X_correlation_min  X_covariance_max  \\\n",
       "dataset                                                           \n",
       "dataset_prostate_singh              -0.989840      6.501921e+06   \n",
       "dataset_glioma_phillips             -0.690733      4.323435e+08   \n",
       "dataset_leukemia_armstrong          -0.916524      1.154806e+08   \n",
       "dataset_breast_sotiriou             -0.803858      6.244864e+00   \n",
       "dataset_lymphoma_dave_2             -0.739594      5.893252e+00   \n",
       "\n",
       "                            X_covariance_mean  X_covariance_min  \\\n",
       "dataset                                                           \n",
       "dataset_prostate_singh              29.654376     -1.662059e+06   \n",
       "dataset_glioma_phillips         245013.502247     -6.592495e+07   \n",
       "dataset_leukemia_armstrong      556524.637475     -7.348629e+07   \n",
       "dataset_breast_sotiriou              1.451080     -5.063628e+00   \n",
       "dataset_lymphoma_dave_2              0.027503     -3.396945e+00   \n",
       "\n",
       "                            X_exp_var_max  X_exp_var_n_t80_cumsum  \\\n",
       "dataset                                                             \n",
       "dataset_prostate_singh           0.538233                       3   \n",
       "dataset_glioma_phillips          0.311808                      16   \n",
       "dataset_leukemia_armstrong       0.249095                      19   \n",
       "dataset_breast_sotiriou          0.721969                       3   \n",
       "dataset_lymphoma_dave_2          0.124228                      93   \n",
       "\n",
       "                            X_kurtosis_max  X_kurtosis_mean  \\\n",
       "dataset                                                       \n",
       "dataset_prostate_singh           96.809515         3.634957   \n",
       "dataset_glioma_phillips          94.998000         6.624304   \n",
       "dataset_leukemia_armstrong       64.783987         3.472936   \n",
       "dataset_breast_sotiriou          10.362369        -0.156562   \n",
       "dataset_lymphoma_dave_2          35.139636         1.273938   \n",
       "\n",
       "                                   ...          X_stand_dev_min  \\\n",
       "dataset                            ...                            \n",
       "dataset_prostate_singh             ...                45.069954   \n",
       "dataset_glioma_phillips            ...               849.046524   \n",
       "dataset_leukemia_armstrong         ...              1480.763026   \n",
       "dataset_breast_sotiriou            ...                 1.308462   \n",
       "dataset_lymphoma_dave_2            ...                 0.610939   \n",
       "\n",
       "                            X_std_covariance_max  X_std_covariance_mean  \\\n",
       "dataset                                                                   \n",
       "dataset_prostate_singh                  0.994905               0.001368   \n",
       "dataset_glioma_phillips                 0.993589               0.026853   \n",
       "dataset_leukemia_armstrong              0.994280               0.084879   \n",
       "dataset_breast_sotiriou                 0.997310               0.703431   \n",
       "dataset_lymphoma_dave_2                 0.985847               0.019387   \n",
       "\n",
       "                            X_std_covariance_min  X_std_exp_var_max  \\\n",
       "dataset                                                               \n",
       "dataset_prostate_singh                 -0.989840           0.635789   \n",
       "dataset_glioma_phillips                -0.690733           0.144017   \n",
       "dataset_leukemia_armstrong             -0.916524           0.241996   \n",
       "dataset_breast_sotiriou                -0.803858           0.745195   \n",
       "dataset_lymphoma_dave_2                -0.739594           0.140919   \n",
       "\n",
       "                            X_var_coef_max  X_var_coef_mean  X_var_coef_min  \\\n",
       "dataset                                                                       \n",
       "dataset_prostate_singh           24.644940         0.968532        0.393882   \n",
       "dataset_glioma_phillips           6.143025         0.497883        0.196919   \n",
       "dataset_leukemia_armstrong        2.250262         0.511498        0.239257   \n",
       "dataset_breast_sotiriou           0.389483         0.202296        0.182322   \n",
       "dataset_lymphoma_dave_2           0.525447         0.158073        0.086560   \n",
       "\n",
       "                            y_norm_class_entropy_none  y_num_classes_none  \n",
       "dataset                                                                    \n",
       "dataset_prostate_singh                       0.999723                   2  \n",
       "dataset_glioma_phillips                      0.795040                   2  \n",
       "dataset_leukemia_armstrong                   0.991532                   3  \n",
       "dataset_breast_sotiriou                      0.993717                   3  \n",
       "dataset_lymphoma_dave_2                      0.882052                   4  \n",
       "\n",
       "[5 rows x 39 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "extractor.results.toDataFrame().head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = extractor.result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}