subgroup_classfier_archiv.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pysubgroup as ps\n",
    "import pandas as pd\n",
    "from graph_pipeline.main import *\n",
    "from tqdm.auto import tqdm\n",
    "import numpy as np\n",
    "import multiprocessing as mp\n",
    "from itertools import product\n",
    "import numpy as np\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.metrics import confusion_matrix, plot_confusion_matrix, classification_report\n",
    "from sklearn.model_selection import GridSearchCV, StratifiedKFold\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from collections import Counter\n",
    "import seaborn as sns\n",
    "import statsmodels.api as sm\n",
    "import pickle\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "from pprint import pprint\n",
    "\n",
    "def filter_on_stats(stat_df, min_replies=10, max_replies=350, language='en', acyclic=True):\n",
    "    filter_array = np.logical_and(stat_df.replies >= min_replies ,stat_df.language == language)\n",
    "    filter_array = np.logical_and(filter_array, stat_df.is_acyclic == acyclic)\n",
    "    filter_array = np.logical_and(filter_array,stat_df.replies <= max_replies)\n",
    "    filter_array = filter_array.to_numpy()\n",
    "    return filter_array\n",
    "\n",
    "def drop_useless(graph_feature_df):\n",
    "    return graph_feature_df.drop(columns=[\"network_diameter\", \"network_radius\", \"closeness_centralization\", \"102\", \"111D\", \"111U\", \n",
    "                                          \"030C\", \"201\", \"120D\", \"120U\", \"120C\", \"210\", \"300\" ])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "stat_df = pd.read_pickle(\"archive_collection_extracted/stat_df\")\n",
    "graph_features_all = pd.read_pickle(\"archive_collection_extracted/graph_features_all\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "graph_features_clean = drop_useless(graph_features_all).drop(columns=[\"003\",\"max_degree\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>average_degree</th>\n",
       "      <th>density</th>\n",
       "      <th>longest_path</th>\n",
       "      <th>transitivity</th>\n",
       "      <th>average_clustering</th>\n",
       "      <th>degree_centralization</th>\n",
       "      <th>betweenness_centralization</th>\n",
       "      <th>012</th>\n",
       "      <th>021D</th>\n",
       "      <th>021U</th>\n",
       "      <th>021C</th>\n",
       "      <th>030T</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>788034.000000</td>\n",
       "      <td>788034.000000</td>\n",
       "      <td>788034.000000</td>\n",
       "      <td>788034.000000</td>\n",
       "      <td>788034.000000</td>\n",
       "      <td>695437.000000</td>\n",
       "      <td>788034.000000</td>\n",
       "      <td>695437.000000</td>\n",
       "      <td>695437.000000</td>\n",
       "      <td>695437.000000</td>\n",
       "      <td>695437.000000</td>\n",
       "      <td>695437.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>1.733873</td>\n",
       "      <td>0.170053</td>\n",
       "      <td>0.287100</td>\n",
       "      <td>0.054778</td>\n",
       "      <td>0.009852</td>\n",
       "      <td>0.600335</td>\n",
       "      <td>0.039149</td>\n",
       "      <td>0.172408</td>\n",
       "      <td>0.002700</td>\n",
       "      <td>0.219646</td>\n",
       "      <td>0.065254</td>\n",
       "      <td>0.003039</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.395908</td>\n",
       "      <td>0.155705</td>\n",
       "      <td>0.168515</td>\n",
       "      <td>0.130479</td>\n",
       "      <td>0.035750</td>\n",
       "      <td>0.277440</td>\n",
       "      <td>0.090696</td>\n",
       "      <td>0.165858</td>\n",
       "      <td>0.023062</td>\n",
       "      <td>0.294452</td>\n",
       "      <td>0.185972</td>\n",
       "      <td>0.039478</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.666667</td>\n",
       "      <td>0.000255</td>\n",
       "      <td>0.003282</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>1.500000</td>\n",
       "      <td>0.045238</td>\n",
       "      <td>0.153846</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.362401</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.015641</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.018725</td>\n",
       "      <td>0.000064</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>1.800000</td>\n",
       "      <td>0.119048</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.569170</td>\n",
       "      <td>0.007700</td>\n",
       "      <td>0.135835</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.091667</td>\n",
       "      <td>0.005647</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>2.000000</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.861111</td>\n",
       "      <td>0.033333</td>\n",
       "      <td>0.290909</td>\n",
       "      <td>0.000007</td>\n",
       "      <td>0.285714</td>\n",
       "      <td>0.033333</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>24.181818</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.958333</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       average_degree        density   longest_path   transitivity  \\\n",
       "count   788034.000000  788034.000000  788034.000000  788034.000000   \n",
       "mean         1.733873       0.170053       0.287100       0.054778   \n",
       "std          0.395908       0.155705       0.168515       0.130479   \n",
       "min          0.666667       0.000255       0.003282       0.000000   \n",
       "25%          1.500000       0.045238       0.153846       0.000000   \n",
       "50%          1.800000       0.119048       0.250000       0.000000   \n",
       "75%          2.000000       0.250000       0.400000       0.000000   \n",
       "max         24.181818       0.500000       0.958333       0.500000   \n",
       "\n",
       "       average_clustering  degree_centralization  betweenness_centralization  \\\n",
       "count       788034.000000          695437.000000               788034.000000   \n",
       "mean             0.009852               0.600335                    0.039149   \n",
       "std              0.035750               0.277440                    0.090696   \n",
       "min              0.000000               0.000000                    0.000000   \n",
       "25%              0.000000               0.362401                    0.000000   \n",
       "50%              0.000000               0.569170                    0.007700   \n",
       "75%              0.000000               0.861111                    0.033333   \n",
       "max              0.500000               1.000000                    0.500000   \n",
       "\n",
       "                 012           021D           021U           021C  \\\n",
       "count  695437.000000  695437.000000  695437.000000  695437.000000   \n",
       "mean        0.172408       0.002700       0.219646       0.065254   \n",
       "std         0.165858       0.023062       0.294452       0.185972   \n",
       "min         0.000000       0.000000       0.000000       0.000000   \n",
       "25%         0.015641       0.000000       0.018725       0.000064   \n",
       "50%         0.135835       0.000000       0.091667       0.005647   \n",
       "75%         0.290909       0.000007       0.285714       0.033333   \n",
       "max         1.000000       1.000000       1.000000       1.000000   \n",
       "\n",
       "                030T  \n",
       "count  695437.000000  \n",
       "mean        0.003039  \n",
       "std         0.039478  \n",
       "min         0.000000  \n",
       "25%         0.000000  \n",
       "50%         0.000000  \n",
       "75%         0.000000  \n",
       "max         1.000000  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "graph_features_clean.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "graph_features_clean.describe().to_excel(\"archive_collection_extracted/graph_features_description.xlsx\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Random Forests "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d170db510cdb445b8692bdd75d44ebd7",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/14 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10 50\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/hengelhardt/masterarbeit/.venv/lib/python3.6/site-packages/joblib/externals/loky/process_executor.py:691: UserWarning: A worker stopped while some jobs were given to the executor. This can be caused by a too short worker timeout or by a memory leak.\n",
      "  \"timeout or by a memory leak.\", UserWarning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'class_weight': {'v': 2.4514009756575884e-05, 'mu': 6.858710562414266e-05, 'int': 4.112180277983387e-05, 'pol': 1.8767007600638077e-05, 'tv': 3.922030042750128e-05, 'biz': 4.9509852460639665e-05, 'r9k': 6.692992436918547e-05}, 'criterion': 'gini', 'max_depth': 15, 'max_features': 'auto', 'n_estimators': 200, 'random_state': 42}\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         biz       0.14      0.16      0.15      5050\n",
      "         int       0.15      0.12      0.13      6079\n",
      "          mu       0.13      0.13      0.13      3645\n",
      "         pol       0.39      0.38      0.38     13321\n",
      "         r9k       0.16      0.32      0.21      3735\n",
      "          tv       0.17      0.25      0.20      6375\n",
      "           v       0.24      0.11      0.16     10199\n",
      "\n",
      "    accuracy                           0.23     48404\n",
      "   macro avg       0.20      0.21      0.20     48404\n",
      "weighted avg       0.24      0.23      0.22     48404\n",
      "\n",
      "10 100\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/hengelhardt/masterarbeit/.venv/lib/python3.6/site-packages/joblib/externals/loky/process_executor.py:691: UserWarning: A worker stopped while some jobs were given to the executor. This can be caused by a too short worker timeout or by a memory leak.\n",
      "  \"timeout or by a memory leak.\", UserWarning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'class_weight': {'tv': 3.37997701615629e-05, 'v': 2.0376558806748717e-05, 'pol': 1.524343770007012e-05, 'int': 3.622925874936599e-05, 'mu': 6.185439475474733e-05, 'r9k': 6.0727515637335276e-05, 'biz': 4.516915849857717e-05}, 'criterion': 'gini', 'max_depth': 15, 'max_features': 'auto', 'n_estimators': 175, 'random_state': 42}\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         biz       0.13      0.18      0.15      5534\n",
      "         int       0.16      0.14      0.15      6901\n",
      "          mu       0.13      0.13      0.13      4041\n",
      "         pol       0.44      0.40      0.42     16401\n",
      "         r9k       0.16      0.31      0.21      4117\n",
      "          tv       0.18      0.25      0.21      7397\n",
      "           v       0.29      0.13      0.18     12269\n",
      "\n",
      "    accuracy                           0.24     56660\n",
      "   macro avg       0.21      0.22      0.21     56660\n",
      "weighted avg       0.27      0.24      0.25     56660\n",
      "\n",
      "10 350\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/hengelhardt/masterarbeit/.venv/lib/python3.6/site-packages/joblib/externals/loky/process_executor.py:691: UserWarning: A worker stopped while some jobs were given to the executor. This can be caused by a too short worker timeout or by a memory leak.\n",
      "  \"timeout or by a memory leak.\", UserWarning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'class_weight': {'int': 3.195194427580918e-05, 'tv': 2.9887324785558443e-05, 'v': 1.745413924912293e-05, 'pol': 1.1361570623522995e-05, 'biz': 4.242321398269133e-05, 'mu': 5.302789267154523e-05, 'r9k': 5.81767409389726e-05}, 'criterion': 'gini', 'max_depth': 15, 'max_features': 'auto', 'n_estimators': 200, 'random_state': 42}\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         biz       0.13      0.22      0.16      5893\n",
      "         int       0.17      0.14      0.15      7824\n",
      "          mu       0.22      0.18      0.20      4714\n",
      "         pol       0.56      0.41      0.48     22005\n",
      "         r9k       0.14      0.36      0.20      4297\n",
      "          tv       0.18      0.28      0.22      8365\n",
      "           v       0.33      0.17      0.22     14324\n",
      "\n",
      "    accuracy                           0.28     67422\n",
      "   macro avg       0.25      0.25      0.23     67422\n",
      "weighted avg       0.33      0.28      0.29     67422\n",
      "\n",
      "50 100\n",
      "{'class_weight': 'balanced', 'criterion': 'gini', 'max_depth': 15, 'max_features': 'auto', 'n_estimators': 200, 'random_state': 42}\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         biz       0.11      0.06      0.07       509\n",
      "         int       0.18      0.24      0.21       857\n",
      "          mu       0.34      0.15      0.20       415\n",
      "         pol       0.53      0.59      0.56      3181\n",
      "         r9k       0.17      0.16      0.17       399\n",
      "          tv       0.20      0.24      0.22      1069\n",
      "           v       0.34      0.28      0.31      2141\n",
      "\n",
      "    accuracy                           0.36      8571\n",
      "   macro avg       0.27      0.25      0.25      8571\n",
      "weighted avg       0.36      0.36      0.35      8571\n",
      "\n",
      "50 150\n",
      "{'class_weight': {'pol': 5.55000555000555e-05, 'r9k': 0.0005012531328320802, 'v': 8.32570144034635e-05, 'int': 0.00023030861354214648, 'tv': 0.00017838030681412772, 'biz': 0.0004043671653861706, 'mu': 0.0005002501250625312}, 'criterion': 'gini', 'max_depth': 15, 'max_features': 'auto', 'n_estimators': 200, 'random_state': 42}\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         biz       0.09      0.06      0.07       618\n",
      "         int       0.18      0.29      0.23      1086\n",
      "          mu       0.30      0.12      0.17       500\n",
      "         pol       0.55      0.60      0.57      4505\n",
      "         r9k       0.18      0.17      0.18       498\n",
      "          tv       0.22      0.26      0.24      1402\n",
      "           v       0.36      0.26      0.30      3003\n",
      "\n",
      "    accuracy                           0.37     11612\n",
      "   macro avg       0.27      0.25      0.25     11612\n",
      "weighted avg       0.38      0.37      0.37     11612\n",
      "\n",
      "50 250\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/hengelhardt/masterarbeit/.venv/lib/python3.6/site-packages/joblib/externals/loky/process_executor.py:691: UserWarning: A worker stopped while some jobs were given to the executor. This can be caused by a too short worker timeout or by a memory leak.\n",
      "  \"timeout or by a memory leak.\", UserWarning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'class_weight': {'tv': 0.0001490979573579842, 'v': 6.585879873551106e-05, 'biz': 0.0003663003663003663, 'pol': 4.358627903935841e-05, 'int': 0.00020181634712411706, 'r9k': 0.0004470272686633885, 'mu': 0.00045045045045045046}, 'criterion': 'gini', 'max_depth': 15, 'max_features': 'auto', 'n_estimators': 200, 'random_state': 42}\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         biz       0.08      0.07      0.07       682\n",
      "         int       0.19      0.30      0.23      1239\n",
      "          mu       0.27      0.14      0.18       555\n",
      "         pol       0.58      0.61      0.59      5736\n",
      "         r9k       0.15      0.17      0.16       559\n",
      "          tv       0.22      0.25      0.23      1677\n",
      "           v       0.40      0.30      0.34      3796\n",
      "\n",
      "    accuracy                           0.39     14244\n",
      "   macro avg       0.27      0.26      0.26     14244\n",
      "weighted avg       0.40      0.39      0.39     14244\n",
      "\n",
      "50 350\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/hengelhardt/masterarbeit/.venv/lib/python3.6/site-packages/joblib/externals/loky/process_executor.py:691: UserWarning: A worker stopped while some jobs were given to the executor. This can be caused by a too short worker timeout or by a memory leak.\n",
      "  \"timeout or by a memory leak.\", UserWarning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'class_weight': 'balanced', 'criterion': 'gini', 'max_depth': 15, 'max_features': 'auto', 'n_estimators': 200, 'random_state': 42}\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         biz       0.14      0.16      0.15       867\n",
      "         int       0.25      0.18      0.21      1780\n",
      "          mu       0.47      0.37      0.41      1087\n",
      "         pol       0.67      0.60      0.63      8785\n",
      "         r9k       0.12      0.32      0.18       580\n",
      "          tv       0.23      0.26      0.24      2037\n",
      "           v       0.36      0.39      0.37      4196\n",
      "\n",
      "    accuracy                           0.44     19332\n",
      "   macro avg       0.32      0.33      0.31     19332\n",
      "weighted avg       0.47      0.44      0.45     19332\n",
      "\n",
      "100 150\n",
      "{'class_weight': {'int': 0.0010570824524312897, 'v': 0.00028312570781426955, 'pol': 0.0001836884643644379, 'tv': 0.00072992700729927, 'mu': 0.002824858757062147, 'r9k': 0.0024154589371980675, 'biz': 0.002207505518763797}, 'criterion': 'gini', 'max_depth': 10, 'max_features': 'auto', 'n_estimators': 175, 'random_state': 42}\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         biz       0.07      0.04      0.05       114\n",
      "         int       0.15      0.35      0.21       237\n",
      "          mu       0.21      0.16      0.18        88\n",
      "         pol       0.63      0.59      0.61      1361\n",
      "         r9k       0.12      0.20      0.15       103\n",
      "          tv       0.24      0.33      0.28       343\n",
      "           v       0.40      0.24      0.30       883\n",
      "\n",
      "    accuracy                           0.40      3129\n",
      "   macro avg       0.26      0.27      0.25      3129\n",
      "weighted avg       0.44      0.40      0.41      3129\n",
      "\n",
      "100 250\n",
      "{'class_weight': 'balanced', 'criterion': 'gini', 'max_depth': 13, 'max_features': 'auto', 'n_estimators': 175, 'random_state': 42}\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         biz       0.09      0.03      0.04       178\n",
      "         int       0.17      0.35      0.23       390\n",
      "          mu       0.30      0.10      0.15       144\n",
      "         pol       0.62      0.64      0.63      2592\n",
      "         r9k       0.12      0.12      0.12       164\n",
      "          tv       0.26      0.24      0.25       618\n",
      "           v       0.41      0.36      0.39      1676\n",
      "\n",
      "    accuracy                           0.45      5762\n",
      "   macro avg       0.28      0.26      0.26      5762\n",
      "weighted avg       0.45      0.45      0.45      5762\n",
      "\n",
      "100 350\n",
      "{'class_weight': 'balanced', 'criterion': 'gini', 'max_depth': 13, 'max_features': 'auto', 'n_estimators': 150, 'random_state': 42}\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         biz       0.26      0.22      0.24       362\n",
      "         int       0.32      0.20      0.25       932\n",
      "          mu       0.45      0.53      0.49       676\n",
      "         pol       0.74      0.59      0.65      5641\n",
      "         r9k       0.10      0.35      0.16       184\n",
      "          tv       0.29      0.21      0.24       979\n",
      "           v       0.37      0.60      0.46      2076\n",
      "\n",
      "    accuracy                           0.50     10850\n",
      "   macro avg       0.36      0.39      0.35     10850\n",
      "weighted avg       0.55      0.50      0.51     10850\n",
      "\n",
      "150 250\n",
      "{'class_weight': 'balanced', 'criterion': 'gini', 'max_depth': 13, 'max_features': 'auto', 'n_estimators': 125, 'random_state': 42}\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         biz       0.00      0.00      0.00        66\n",
      "         int       0.19      0.31      0.23       157\n",
      "          mu       0.50      0.07      0.12        56\n",
      "         pol       0.62      0.67      0.64      1252\n",
      "         r9k       0.19      0.13      0.15        62\n",
      "          tv       0.24      0.17      0.20       278\n",
      "           v       0.43      0.43      0.43       806\n",
      "\n",
      "    accuracy                           0.48      2677\n",
      "   macro avg       0.31      0.25      0.26      2677\n",
      "weighted avg       0.47      0.48      0.47      2677\n",
      "\n",
      "150 350\n",
      "{'class_weight': 'balanced', 'criterion': 'gini', 'max_depth': 13, 'max_features': 'auto', 'n_estimators': 125, 'random_state': 42}\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         biz       0.27      0.32      0.29       250\n",
      "         int       0.37      0.22      0.28       698\n",
      "          mu       0.48      0.58      0.52       589\n",
      "         pol       0.75      0.64      0.69      4301\n",
      "         r9k       0.10      0.23      0.14        82\n",
      "          tv       0.30      0.23      0.26       639\n",
      "           v       0.38      0.63      0.47      1206\n",
      "\n",
      "    accuracy                           0.54      7765\n",
      "   macro avg       0.38      0.41      0.38      7765\n",
      "weighted avg       0.58      0.54      0.55      7765\n",
      "\n",
      "200 350\n",
      "{'class_weight': 'balanced', 'criterion': 'gini', 'max_depth': 13, 'max_features': 'auto', 'n_estimators': 150, 'random_state': 42}\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         biz       0.28      0.40      0.33       208\n",
      "         int       0.44      0.28      0.34       598\n",
      "          mu       0.52      0.62      0.57       552\n",
      "         pol       0.76      0.68      0.72      3544\n",
      "         r9k       0.07      0.12      0.09        42\n",
      "          tv       0.27      0.23      0.25       469\n",
      "           v       0.37      0.57      0.45       724\n",
      "\n",
      "    accuracy                           0.58      6137\n",
      "   macro avg       0.39      0.42      0.39      6137\n",
      "weighted avg       0.60      0.58      0.58      6137\n",
      "\n",
      "250 350\n",
      "{'class_weight': {'v': 0.0006165228113440197, 'pol': 8.177283506419168e-05, 'mu': 0.00046948356807511736, 'int': 0.00046125461254612545, 'tv': 0.0006891798759476223, 'biz': 0.0013531799729364006, 'r9k': 0.011904761904761904}, 'criterion': 'gini', 'max_depth': 13, 'max_features': 'auto', 'n_estimators': 200, 'random_state': 42}\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         biz       0.27      0.43      0.33       185\n",
      "         int       0.44      0.27      0.34       542\n",
      "          mu       0.49      0.58      0.53       533\n",
      "         pol       0.75      0.71      0.73      3057\n",
      "         r9k       0.18      0.10      0.12        21\n",
      "          tv       0.21      0.25      0.23       363\n",
      "           v       0.38      0.47      0.42       405\n",
      "\n",
      "    accuracy                           0.59      5106\n",
      "   macro avg       0.39      0.40      0.39      5106\n",
      "weighted avg       0.60      0.59      0.59      5106\n",
      "\n"
     ]
    }
   ],
   "source": [
    "graph_features_all = graph_features_clean\n",
    "graph_features_all['replies'] = stat_df.set_index('thread_id').replies\n",
    "scaler = StandardScaler()\n",
    "\n",
    "\n",
    "clf = RandomForestClassifier(random_state=42)\n",
    "\n",
    "cv = StratifiedKFold(\n",
    "    n_splits=5\n",
    ")\n",
    "cv_results_rf = {}\n",
    "for min_replies, max_replies in tqdm([(10, 50), (10, 100), (10, 350), (50,100), \n",
    "                                      (50, 150), (50, 250), (50, 350), (100, 150), \n",
    "                                      (100, 250), (100, 350), (150,250), (150, 350), \n",
    "                                      (200, 350), (250, 350)]):\n",
    "    print(min_replies, max_replies)\n",
    "    filter_array = filter_on_stats(stat_df, min_replies=min_replies, max_replies=max_replies)\n",
    "    graph_features = graph_features_all.loc[filter_array]\n",
    "    graph_features = graph_features.dropna()\n",
    "    X = graph_features.drop(columns=[\"board\", \"replies\"]).to_numpy()\n",
    "    y = graph_features.board.to_numpy()\n",
    "    scaler.fit(X)\n",
    "    X_transformed = scaler.transform(X)\n",
    "    X_train, X_test, y_train, y_test = train_test_split(X_transformed, y, test_size=0.2, \n",
    "                                                        random_state=42, stratify=y)\n",
    "    count = Counter(y_train)\n",
    "    \n",
    "    param_grid = { \n",
    "    'n_estimators': [50, 100, 125, 150, 175, 200],\n",
    "    'max_features': ['auto'],\n",
    "    'max_depth' : [10, 13, 15, None],\n",
    "    'criterion' :['gini'],\n",
    "    'class_weight': ['balanced', {label:1/count[label] for label in count.keys()}],\n",
    "    'random_state': [42]\n",
    "    }\n",
    "\n",
    "    search = GridSearchCV(\n",
    "        estimator=clf, param_grid=param_grid,\n",
    "        scoring='f1_macro', cv=cv,\n",
    "        n_jobs=20, refit = True\n",
    "    )\n",
    "    search.fit(X_train, y_train)\n",
    "    print(search.best_params_)\n",
    "    key = f\"{min_replies}-{max_replies}\"\n",
    "    cv_results_rf[key] = {\"model\":search, \n",
    "                      'X_test': X_test, \n",
    "                      'X_train': X_train, \n",
    "                      'y_test':y_test, \n",
    "                      \"y_train\":y_train\n",
    "                     }\n",
    "    cv_results_rf[key]['report']  = classification_report(cv_results_rf[key]['y_test'], cv_results_rf[key]['model'].predict(cv_results_rf[key]['X_test']))\n",
    "    cv_results_rf[key]['report_dict']  = classification_report(cv_results_rf[key]['y_test'], cv_results_rf[key]['model'].predict(cv_results_rf[key]['X_test']), output_dict=True)\n",
    "    cv_results_rf[key]['best_params'] = cv_results_rf[key]['model'].best_params_\n",
    "    print(classification_report(cv_results_rf[key]['y_test'], cv_results_rf[key]['model'].predict(cv_results_rf[key]['X_test'])))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(\"archive_collection_extracted/cv_restults_rf\", \"wb\") as f:\n",
    "    pickle.dump(cv_results_rf, f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(\"archive_collection_extracted/cv_restults_rf\", \"rb\") as f:\n",
    "    cv_results_rf = pickle.load(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "multi_dict = {}\n",
    "param_dict = {}\n",
    "for key in cv_results_rf.keys():\n",
    "    multi_dict[key] = pd.DataFrame(cv_results_rf[key]['report_dict'])\n",
    "    param_dict[key] = cv_results_rf[key]['best_params']\n",
    "df = pd.concat(multi_dict.values(), keys=multi_dict.keys())\n",
    "param_df = pd.DataFrame(param_dict)\n",
    "df.round(2).to_excel(\"archive_collection_extracted/classification_report_rf.xlsx\")\n",
    "param_df.to_excel(\"archive_collection_extracted/classification_params_rf.xlsx\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "viz_dict_rf = {\"precision\":{},\n",
    "                \"precision_weighted\":{},\n",
    "                \"recall\":{},\n",
    "                \"recall_weighted\":{},\n",
    "                \"f1\":{},\n",
    "                \"f1_weigthed\":{},\n",
    "                #\"accuracy\":{}\n",
    "               }\n",
    "for key in cv_results_rf.keys():\n",
    "    df =  pd.DataFrame(cv_results_rf[key]['report_dict']).transpose()\n",
    "    viz_dict_rf[\"precision\"][key] = df.precision.loc[\"macro avg\"]\n",
    "    viz_dict_rf[\"precision_weighted\"][key] = df.precision.loc[\"weighted avg\"]\n",
    "    viz_dict_rf[\"recall\"][key] = df.recall.loc[\"macro avg\"]\n",
    "    viz_dict_rf[\"recall_weighted\"][key] = df.recall.loc[\"weighted avg\"]\n",
    "    viz_dict_rf[\"f1\"][key] = df[\"f1-score\"].loc[\"macro avg\"]\n",
    "    viz_dict_rf[\"f1_weigthed\"][key] = df[\"f1-score\"].loc[\"weighted avg\"]\n",
    "    #viz_dict_rf[\"accuracy\"][key] = df[\"f1-score\"].loc[\"accuracy\"]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f957e659048>"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x504 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set_style(\"whitegrid\")\n",
    "plt.figure(figsize=(12,7))\n",
    "sns.lineplot(data=viz_dict_rf).set_xlabel(\"Spanne an Antworten\")\n",
    "plt.legend([\"Genauigkeit\", \"Genauigkeit gewichtet\", \"Sensitivität\", \"Sensitivität gewichtet\", \"F1\", \"F1 gewichtet\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [],
   "source": [
    "pd.DataFrame(cv_results_rf[\"10-350\"]['report_dict']).transpose().round(2).join(pd.DataFrame(cv_results_rf[\"250-350\"]['report_dict']).transpose().round(2), lsuffix=\" (10-350)\", rsuffix=\" (250-350)\", how=\"left\").to_excel(\"archive_collection_extracted/classification_report_rf_10-350_250-350.xlsx\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Threads mit 250-350 Antworten')"
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x360 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set_style(\"whitegrid\", {'axes.grid' : False})\n",
    "fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14,5))\n",
    "g1=plot_confusion_matrix(estimator=cv_results_rf[\"10-350\"][\"model\"], X=cv_results_rf[\"10-350\"][\"X_test\"], y_true=cv_results_rf[\"10-350\"][\"y_test\"],cmap=\"magma\", ax=ax1)\n",
    "g1.ax_.set_xlabel('Vorhergesagtes Board')\n",
    "g1.ax_.set_ylabel('Tatsächliches Board')\n",
    "g1.ax_.set_title(\"Threads mit 10-350 Antworten\")\n",
    "g2=plot_confusion_matrix(estimator=cv_results_rf[\"250-350\"][\"model\"], X=cv_results_rf[\"250-350\"][\"X_test\"], y_true=cv_results_rf[\"250-350\"][\"y_test\"],cmap=\"magma\", ax=ax2)\n",
    "g2.ax_.set_xlabel('Vorhergesagtes Board')\n",
    "g2.ax_.set_ylabel('Tatsächliches Board')\n",
    "g2.ax_.set_title(\"Threads mit 250-350 Antworten\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# binary log mit statsmodels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "cc8cfa2b9aa445f493f58b92ce74b969",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/2 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10 350\n",
      "10 350\n",
      "110021\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.596724\n",
      "         Iterations 6\n",
      "10 350\n",
      "41824\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.369151\n",
      "         Iterations 7\n",
      "10 350\n",
      "71617\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.514381\n",
      "         Iterations 30\n",
      "250 350\n",
      "250 350\n",
      "2663\n",
      "Warning: Maximum number of iterations has been exceeded.\n",
      "         Current function value: 0.292359\n",
      "         Iterations: 35\n",
      "250 350\n",
      "15286\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/hengelhardt/masterarbeit/.venv/lib/python3.6/site-packages/statsmodels/base/model.py:568: ConvergenceWarning: Maximum Likelihood optimization failed to converge. Check mle_retvals\n",
      "  ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Warning: Maximum number of iterations has been exceeded.\n",
      "         Current function value: 0.667332\n",
      "         Iterations: 35\n",
      "250 350\n",
      "2027\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/hengelhardt/masterarbeit/.venv/lib/python3.6/site-packages/statsmodels/base/model.py:568: ConvergenceWarning: Maximum Likelihood optimization failed to converge. Check mle_retvals\n",
      "  ConvergenceWarning)\n",
      "/home/hengelhardt/masterarbeit/.venv/lib/python3.6/site-packages/statsmodels/discrete/discrete_model.py:1810: RuntimeWarning: overflow encountered in exp\n",
      "  return 1/(1+np.exp(-X))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Warning: Maximum number of iterations has been exceeded.\n",
      "         Current function value: inf\n",
      "         Iterations: 35\n",
      "EXCEPTION!!! Singular matrix !!!\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/hengelhardt/masterarbeit/.venv/lib/python3.6/site-packages/statsmodels/discrete/discrete_model.py:1863: RuntimeWarning: divide by zero encountered in log\n",
      "  return np.sum(np.log(self.cdf(q*np.dot(X,params))))\n"
     ]
    }
   ],
   "source": [
    "scaler = StandardScaler()\n",
    "\n",
    "result_dict_stats_binary = {}\n",
    "for min_replies, max_replies in tqdm([(10, 350), (250, 350)]):\n",
    "    print(min_replies, max_replies)\n",
    "    filter_array = filter_on_stats(stat_df, min_replies=min_replies)\n",
    "    graph_features = graph_features_clean.loc[filter_array]\n",
    "    graph_features = graph_features.dropna()\n",
    "    X = graph_features.drop(columns=[\"board\"]).astype(np.float64)\n",
    "    y = graph_features.board \n",
    "    scaler.fit(X)\n",
    "    X_transformed = scaler.transform(X)\n",
    "    key = f\"{min_replies}-{max_replies}\"\n",
    "    result_dict_stats_binary[key] = {}\n",
    "    try:\n",
    "        temp_dict = {}\n",
    "        for board in [\"pol\", \"tv\", \"v\"] if min_replies == 10 else [\"mu\", \"pol\", \"v\"]:\n",
    "            y_train_binary = y == board\n",
    "            print(min_replies, max_replies)\n",
    "            print(y_train_binary.sum())\n",
    "            model = sm.Logit(endog=y_train_binary, exog=sm.add_constant(X))\n",
    "            result = model.fit()\n",
    "            temp_dict[board] = {'model': model,\n",
    "                                'result': result,\n",
    "                                }\n",
    "        result_dict_stats_binary[key]['predictors'] = temp_dict\n",
    "    except Exception as e:\n",
    "        print(f\"EXCEPTION!!! {e} !!!\")\n",
    "        continue"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "pol\n",
      "tv\n",
      "v\n"
     ]
    }
   ],
   "source": [
    "for min_replies in [10]:\n",
    "    for key in result_dict_stats_binary[f\"{min_replies}-350\"][\"predictors\"].keys():\n",
    "        print(key)\n",
    "        pd.DataFrame(result_dict_stats_binary[f\"{min_replies}-350\"][\"predictors\"][key][\"result\"].summary().tables[1]).transpose().to_excel(f\"archive_collection_extracted/log_reg_result{min_replies}_board_{key}.xlsx\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Subgroup Discovery"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "def search_subgroups(data, min_replies, target_attribute, target, a):\n",
    "    result_dict = {}\n",
    "    ps_target = ps.BinaryTarget(target_attribute=target_attribute, target_value=target)\n",
    "    ps_search_space = ps.create_selectors(data, ignore=[target_attribute], nbins =30)\n",
    "    ps_task = ps.SubgroupDiscoveryTask(data, ps_target, ps_search_space, qf=ps.GeneralizationAware_StandardQF(a))\n",
    "    result_dict[f\"{min_replies}_{target}_GA_{a}_SDFS\"] = ps.SimpleDFS().execute(ps_task).to_dataframe()\n",
    "    ps_task = ps.SubgroupDiscoveryTask(data, ps_target, ps_search_space, qf=ps.StandardQF(a))\n",
    "    result_dict[f\"{min_replies}_{target}_SQF_{a}_BFS\"] = ps.BestFirstSearch().execute(ps_task).to_dataframe()\n",
    "    return result_dict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "da428aae3ef044b3b8bb54ae3bf87d45",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "adding search to mp pool:   0%|          | 0/18 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "70af0fad5c484ab78d555304fe9a4b34",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "retrieving subgroups:   0%|          | 0/18 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "subgroup_combs = []\n",
    "[subgroup_combs.append(prod) for prod in product([10], [\"board\"], [\"pol\", \"tv\", \"v\"], [0.5, 1])]\n",
    "[subgroup_combs.append(prod) for prod in product([250], [\"board\"], [\"mu\", \"pol\", \"v\"], [0.5, 1])]\n",
    "pool = mp.Pool(processes=18)\n",
    "subgroup_archive_results = [pool.apply_async(search_subgroups, (graph_features_clean.loc[filter_on_stats(stat_df, min_replies=item[0])], item[0], item[1], item[2], item[3],))\n",
    "           for item in tqdm(subgroup_combs,desc=\"adding search to mp pool\")]\n",
    "subgroup_archive_results = [result.get() for result in tqdm(subgroup_archive_results, desc=\"retrieving subgroups\")]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[{'10_pol_GA_0_SDFS':     quality                                           subgroup  size_sg  \\\n",
       "  0  0.673631  021U: [0.02:0.03[ AND average_degree: [1.82:1....      1.0   \n",
       "  1  0.673631  021U: [0.02:0.03[ AND average_degree: [1.82:1....      1.0   \n",
       "  2  0.673631  021U: [0.03:0.03[ AND average_degree: [1.82:1....      1.0   \n",
       "  3  0.673631  average_degree: [1.82:1.83[ AND degree_central...      3.0   \n",
       "  4  0.673631  average_degree: [1.82:1.83[ AND betweenness_ce...      1.0   \n",
       "  5  0.673631  average_degree: [1.82:1.83[ AND density: [0.03...      1.0   \n",
       "  6  0.673631  average_degree: [1.82:1.83[ AND degree_central...      1.0   \n",
       "  7  0.673631  average_degree: [1.82:1.83[ AND density: [0.03...      2.0   \n",
       "  8  0.673631  average_degree: [1.82:1.83[ AND longest_path: ...      1.0   \n",
       "  9  0.673631  average_degree: [1.82:1.83[ AND betweenness_ce...      1.0   \n",
       "  \n",
       "     size_dataset  positives_sg  positives_dataset  size_complement  \\\n",
       "  0      337106.0           1.0           110021.0         337105.0   \n",
       "  1      337106.0           1.0           110021.0         337105.0   \n",
       "  2      337106.0           1.0           110021.0         337105.0   \n",
       "  3      337106.0           3.0           110021.0         337103.0   \n",
       "  4      337106.0           1.0           110021.0         337105.0   \n",
       "  5      337106.0           1.0           110021.0         337105.0   \n",
       "  6      337106.0           1.0           110021.0         337105.0   \n",
       "  7      337106.0           2.0           110021.0         337104.0   \n",
       "  8      337106.0           1.0           110021.0         337105.0   \n",
       "  9      337106.0           1.0           110021.0         337105.0   \n",
       "  \n",
       "     relative_size_sg  relative_size_complement  coverage_sg  \\\n",
       "  0          0.000003                  0.999997     0.000009   \n",
       "  1          0.000003                  0.999997     0.000009   \n",
       "  2          0.000003                  0.999997     0.000009   \n",
       "  3          0.000009                  0.999991     0.000027   \n",
       "  4          0.000003                  0.999997     0.000009   \n",
       "  5          0.000003                  0.999997     0.000009   \n",
       "  6          0.000003                  0.999997     0.000009   \n",
       "  7          0.000006                  0.999994     0.000018   \n",
       "  8          0.000003                  0.999997     0.000009   \n",
       "  9          0.000003                  0.999997     0.000009   \n",
       "  \n",
       "     coverage_complement  target_share_sg  target_share_complement  \\\n",
       "  0             0.999991              1.0                 0.326367   \n",
       "  1             0.999991              1.0                 0.326367   \n",
       "  2             0.999991              1.0                 0.326367   \n",
       "  3             0.999973              1.0                 0.326363   \n",
       "  4             0.999991              1.0                 0.326367   \n",
       "  5             0.999991              1.0                 0.326367   \n",
       "  6             0.999991              1.0                 0.326367   \n",
       "  7             0.999982              1.0                 0.326365   \n",
       "  8             0.999991              1.0                 0.326367   \n",
       "  9             0.999991              1.0                 0.326367   \n",
       "  \n",
       "     target_share_dataset      lift  \n",
       "  0              0.326369  3.064015  \n",
       "  1              0.326369  3.064015  \n",
       "  2              0.326369  3.064015  \n",
       "  3              0.326369  3.064015  \n",
       "  4              0.326369  3.064015  \n",
       "  5              0.326369  3.064015  \n",
       "  6              0.326369  3.064015  \n",
       "  7              0.326369  3.064015  \n",
       "  8              0.326369  3.064015  \n",
       "  9              0.326369  3.064015  ,\n",
       "  '10_pol_SQF_0_BFS':     quality                     subgroup  size_sg  size_dataset  positives_sg  \\\n",
       "  0  0.673631  average_degree: [2.01:2.01[      1.0      337106.0           1.0   \n",
       "  1  0.673631    transitivity: [0.00:0.00[      1.0      337106.0           1.0   \n",
       "  2  0.673631  average_degree: [2.01:2.01[      1.0      337106.0           1.0   \n",
       "  3  0.673631    transitivity: [0.00:0.00[      1.0      337106.0           1.0   \n",
       "  4  0.673631    transitivity: [0.00:0.00[      1.0      337106.0           1.0   \n",
       "  5  0.673631    transitivity: [0.00:0.00[      1.0      337106.0           1.0   \n",
       "  6  0.673631    transitivity: [0.00:0.00[      1.0      337106.0           1.0   \n",
       "  7  0.673631    transitivity: [0.00:0.00[      1.0      337106.0           1.0   \n",
       "  8  0.673631    transitivity: [0.00:0.00[      1.0      337106.0           1.0   \n",
       "  9  0.673631    transitivity: [0.00:0.00[      1.0      337106.0           1.0   \n",
       "  \n",
       "     positives_dataset  size_complement  relative_size_sg  \\\n",
       "  0           110021.0         337105.0          0.000003   \n",
       "  1           110021.0         337105.0          0.000003   \n",
       "  2           110021.0         337105.0          0.000003   \n",
       "  3           110021.0         337105.0          0.000003   \n",
       "  4           110021.0         337105.0          0.000003   \n",
       "  5           110021.0         337105.0          0.000003   \n",
       "  6           110021.0         337105.0          0.000003   \n",
       "  7           110021.0         337105.0          0.000003   \n",
       "  8           110021.0         337105.0          0.000003   \n",
       "  9           110021.0         337105.0          0.000003   \n",
       "  \n",
       "     relative_size_complement  coverage_sg  coverage_complement  \\\n",
       "  0                  0.999997     0.000009             0.999991   \n",
       "  1                  0.999997     0.000009             0.999991   \n",
       "  2                  0.999997     0.000009             0.999991   \n",
       "  3                  0.999997     0.000009             0.999991   \n",
       "  4                  0.999997     0.000009             0.999991   \n",
       "  5                  0.999997     0.000009             0.999991   \n",
       "  6                  0.999997     0.000009             0.999991   \n",
       "  7                  0.999997     0.000009             0.999991   \n",
       "  8                  0.999997     0.000009             0.999991   \n",
       "  9                  0.999997     0.000009             0.999991   \n",
       "  \n",
       "     target_share_sg  target_share_complement  target_share_dataset      lift  \n",
       "  0              1.0                 0.326367              0.326369  3.064015  \n",
       "  1              1.0                 0.326367              0.326369  3.064015  \n",
       "  2              1.0                 0.326367              0.326369  3.064015  \n",
       "  3              1.0                 0.326367              0.326369  3.064015  \n",
       "  4              1.0                 0.326367              0.326369  3.064015  \n",
       "  5              1.0                 0.326367              0.326369  3.064015  \n",
       "  6              1.0                 0.326367              0.326369  3.064015  \n",
       "  7              1.0                 0.326367              0.326369  3.064015  \n",
       "  8              1.0                 0.326367              0.326369  3.064015  \n",
       "  9              1.0                 0.326367              0.326369  3.064015  },\n",
       " {'10_pol_GA_0.5_SDFS':     quality                                 subgroup  size_sg  size_dataset  \\\n",
       "  0  0.066989                     density: [0.00:0.00[  11250.0      337106.0   \n",
       "  1  0.060182                        021C: [0.00:0.00[  11234.0      337106.0   \n",
       "  2  0.059265                         012: [0.02:0.02[  11237.0      337106.0   \n",
       "  3  0.052736                        030T: [0.00:0.00[  11226.0      337106.0   \n",
       "  4  0.047925                longest_path: [0.03:0.05[  11119.0      337106.0   \n",
       "  5  0.046911                        021C: [0.00:0.00[  11240.0      337106.0   \n",
       "  6  0.046754                         012: [0.02:0.03[  11237.0      337106.0   \n",
       "  7  0.046509                        longest_path<0.03  11224.0      337106.0   \n",
       "  8  0.044528  betweenness_centralization: [0.00:0.00[  11237.0      337106.0   \n",
       "  9  0.044354                        030T: [0.00:0.00[  11214.0      337106.0   \n",
       "  \n",
       "     positives_sg  positives_dataset  size_complement  relative_size_sg  \\\n",
       "  0        7797.0           110021.0         325856.0          0.033372   \n",
       "  1        7370.0           110021.0         325872.0          0.033325   \n",
       "  2        7315.0           110021.0         325869.0          0.033334   \n",
       "  3        6908.0           110021.0         325880.0          0.033301   \n",
       "  4        6563.0           110021.0         325987.0          0.032984   \n",
       "  5        6556.0           110021.0         325866.0          0.033343   \n",
       "  6        6545.0           110021.0         325869.0          0.033334   \n",
       "  7        6524.0           110021.0         325882.0          0.033295   \n",
       "  8        6408.0           110021.0         325869.0          0.033334   \n",
       "  9        6387.0           110021.0         325892.0          0.033266   \n",
       "  \n",
       "     relative_size_complement  coverage_sg  coverage_complement  \\\n",
       "  0                  0.966628     0.070868             0.929132   \n",
       "  1                  0.966675     0.066987             0.933013   \n",
       "  2                  0.966666     0.066487             0.933513   \n",
       "  3                  0.966699     0.062788             0.937212   \n",
       "  4                  0.967016     0.059652             0.940348   \n",
       "  5                  0.966657     0.059589             0.940411   \n",
       "  6                  0.966666     0.059489             0.940511   \n",
       "  7                  0.966705     0.059298             0.940702   \n",
       "  8                  0.966666     0.058243             0.941757   \n",
       "  9                  0.966734     0.058053             0.941947   \n",
       "  \n",
       "     target_share_sg  target_share_complement  target_share_dataset      lift  \n",
       "  0         0.693067                 0.313709              0.326369  2.123567  \n",
       "  1         0.656044                 0.315004              0.326369  2.010129  \n",
       "  2         0.650974                 0.315176              0.326369  1.994596  \n",
       "  3         0.615357                 0.316414              0.326369  1.885464  \n",
       "  4         0.590251                 0.317368              0.326369  1.808538  \n",
       "  5         0.583274                 0.317508              0.326369  1.787160  \n",
       "  6         0.582451                 0.317539              0.326369  1.784638  \n",
       "  7         0.581254                 0.317590              0.326369  1.780972  \n",
       "  8         0.570259                 0.317959              0.326369  1.747282  \n",
       "  9         0.569556                 0.318001              0.326369  1.745128  ,\n",
       "  '10_pol_SQF_0.5_BFS':     quality                                    subgroup  size_sg  \\\n",
       "  0  0.066989                        density: [0.00:0.00[  11250.0   \n",
       "  1  0.060182                           021C: [0.00:0.00[  11234.0   \n",
       "  2  0.059265                            012: [0.02:0.02[  11237.0   \n",
       "  3  0.052736                           030T: [0.00:0.00[  11226.0   \n",
       "  4  0.052587      012: [0.02:0.02[ AND 021C: [0.00:0.00[   7950.0   \n",
       "  5  0.047925                   longest_path: [0.03:0.05[  11119.0   \n",
       "  6  0.047523  021C: [0.00:0.00[ AND density: [0.00:0.00[   5825.0   \n",
       "  7  0.047397   012: [0.02:0.02[ AND density: [0.00:0.00[   5354.0   \n",
       "  8  0.046911                           021C: [0.00:0.00[  11240.0   \n",
       "  9  0.046754                            012: [0.02:0.03[  11237.0   \n",
       "  \n",
       "     size_dataset  positives_sg  positives_dataset  size_complement  \\\n",
       "  0      337106.0        7797.0           110021.0         325856.0   \n",
       "  1      337106.0        7370.0           110021.0         325872.0   \n",
       "  2      337106.0        7315.0           110021.0         325869.0   \n",
       "  3      337106.0        6908.0           110021.0         325880.0   \n",
       "  4      337106.0        5317.0           110021.0         329156.0   \n",
       "  5      337106.0        6563.0           110021.0         325987.0   \n",
       "  6      337106.0        4007.0           110021.0         331281.0   \n",
       "  7      337106.0        3761.0           110021.0         331752.0   \n",
       "  8      337106.0        6556.0           110021.0         325866.0   \n",
       "  9      337106.0        6545.0           110021.0         325869.0   \n",
       "  \n",
       "     relative_size_sg  relative_size_complement  coverage_sg  \\\n",
       "  0          0.033372                  0.966628     0.070868   \n",
       "  1          0.033325                  0.966675     0.066987   \n",
       "  2          0.033334                  0.966666     0.066487   \n",
       "  3          0.033301                  0.966699     0.062788   \n",
       "  4          0.023583                  0.976417     0.048327   \n",
       "  5          0.032984                  0.967016     0.059652   \n",
       "  6          0.017279                  0.982721     0.036420   \n",
       "  7          0.015882                  0.984118     0.034184   \n",
       "  8          0.033343                  0.966657     0.059589   \n",
       "  9          0.033334                  0.966666     0.059489   \n",
       "  \n",
       "     coverage_complement  target_share_sg  target_share_complement  \\\n",
       "  0             0.929132         0.693067                 0.313709   \n",
       "  1             0.933013         0.656044                 0.315004   \n",
       "  2             0.933513         0.650974                 0.315176   \n",
       "  3             0.937212         0.615357                 0.316414   \n",
       "  4             0.951673         0.668805                 0.318098   \n",
       "  5             0.940348         0.590251                 0.317368   \n",
       "  6             0.963580         0.687897                 0.320012   \n",
       "  7             0.965816         0.702465                 0.320300   \n",
       "  8             0.940411         0.583274                 0.317508   \n",
       "  9             0.940511         0.582451                 0.317539   \n",
       "  \n",
       "     target_share_dataset      lift  \n",
       "  0              0.326369  2.123567  \n",
       "  1              0.326369  2.010129  \n",
       "  2              0.326369  1.994596  \n",
       "  3              0.326369  1.885464  \n",
       "  4              0.326369  2.049229  \n",
       "  5              0.326369  1.808538  \n",
       "  6              0.326369  2.107727  \n",
       "  7              0.326369  2.152365  \n",
       "  8              0.326369  1.787160  \n",
       "  9              0.326369  1.784638  },\n",
       " {'10_pol_GA_1_SDFS':     quality                                 subgroup  size_sg  size_dataset  \\\n",
       "  0  0.012238                     density: [0.00:0.00[  11250.0      337106.0   \n",
       "  1  0.010986                        021C: [0.00:0.00[  11234.0      337106.0   \n",
       "  2  0.010820                         012: [0.02:0.02[  11237.0      337106.0   \n",
       "  3  0.009624                        030T: [0.00:0.00[  11226.0      337106.0   \n",
       "  4  0.008704                longest_path: [0.03:0.05[  11119.0      337106.0   \n",
       "  5  0.008566                        021C: [0.00:0.00[  11240.0      337106.0   \n",
       "  6  0.008536                         012: [0.02:0.03[  11237.0      337106.0   \n",
       "  7  0.008486                        longest_path<0.03  11224.0      337106.0   \n",
       "  8  0.008130  betweenness_centralization: [0.00:0.00[  11237.0      337106.0   \n",
       "  9  0.008090                        030T: [0.00:0.00[  11214.0      337106.0   \n",
       "  \n",
       "     positives_sg  positives_dataset  size_complement  relative_size_sg  \\\n",
       "  0        7797.0           110021.0         325856.0          0.033372   \n",
       "  1        7370.0           110021.0         325872.0          0.033325   \n",
       "  2        7315.0           110021.0         325869.0          0.033334   \n",
       "  3        6908.0           110021.0         325880.0          0.033301   \n",
       "  4        6563.0           110021.0         325987.0          0.032984   \n",
       "  5        6556.0           110021.0         325866.0          0.033343   \n",
       "  6        6545.0           110021.0         325869.0          0.033334   \n",
       "  7        6524.0           110021.0         325882.0          0.033295   \n",
       "  8        6408.0           110021.0         325869.0          0.033334   \n",
       "  9        6387.0           110021.0         325892.0          0.033266   \n",
       "  \n",
       "     relative_size_complement  coverage_sg  coverage_complement  \\\n",
       "  0                  0.966628     0.070868             0.929132   \n",
       "  1                  0.966675     0.066987             0.933013   \n",
       "  2                  0.966666     0.066487             0.933513   \n",
       "  3                  0.966699     0.062788             0.937212   \n",
       "  4                  0.967016     0.059652             0.940348   \n",
       "  5                  0.966657     0.059589             0.940411   \n",
       "  6                  0.966666     0.059489             0.940511   \n",
       "  7                  0.966705     0.059298             0.940702   \n",
       "  8                  0.966666     0.058243             0.941757   \n",
       "  9                  0.966734     0.058053             0.941947   \n",
       "  \n",
       "     target_share_sg  target_share_complement  target_share_dataset      lift  \n",
       "  0         0.693067                 0.313709              0.326369  2.123567  \n",
       "  1         0.656044                 0.315004              0.326369  2.010129  \n",
       "  2         0.650974                 0.315176              0.326369  1.994596  \n",
       "  3         0.615357                 0.316414              0.326369  1.885464  \n",
       "  4         0.590251                 0.317368              0.326369  1.808538  \n",
       "  5         0.583274                 0.317508              0.326369  1.787160  \n",
       "  6         0.582451                 0.317539              0.326369  1.784638  \n",
       "  7         0.581254                 0.317590              0.326369  1.780972  \n",
       "  8         0.570259                 0.317959              0.326369  1.747282  \n",
       "  9         0.569556                 0.318001              0.326369  1.745128  ,\n",
       "  '10_pol_SQF_1_BFS':     quality                                 subgroup  size_sg  size_dataset  \\\n",
       "  0  0.012238                     density: [0.00:0.00[  11250.0      337106.0   \n",
       "  1  0.010986                        021C: [0.00:0.00[  11234.0      337106.0   \n",
       "  2  0.010820                         012: [0.02:0.02[  11237.0      337106.0   \n",
       "  3  0.009624                        030T: [0.00:0.00[  11226.0      337106.0   \n",
       "  4  0.008704                longest_path: [0.03:0.05[  11119.0      337106.0   \n",
       "  5  0.008566                        021C: [0.00:0.00[  11240.0      337106.0   \n",
       "  6  0.008536                         012: [0.02:0.03[  11237.0      337106.0   \n",
       "  7  0.008486                        longest_path<0.03  11224.0      337106.0   \n",
       "  8  0.008130  betweenness_centralization: [0.00:0.00[  11237.0      337106.0   \n",
       "  9  0.008090                        030T: [0.00:0.00[  11214.0      337106.0   \n",
       "  \n",
       "     positives_sg  positives_dataset  size_complement  relative_size_sg  \\\n",
       "  0        7797.0           110021.0         325856.0          0.033372   \n",
       "  1        7370.0           110021.0         325872.0          0.033325   \n",
       "  2        7315.0           110021.0         325869.0          0.033334   \n",
       "  3        6908.0           110021.0         325880.0          0.033301   \n",
       "  4        6563.0           110021.0         325987.0          0.032984   \n",
       "  5        6556.0           110021.0         325866.0          0.033343   \n",
       "  6        6545.0           110021.0         325869.0          0.033334   \n",
       "  7        6524.0           110021.0         325882.0          0.033295   \n",
       "  8        6408.0           110021.0         325869.0          0.033334   \n",
       "  9        6387.0           110021.0         325892.0          0.033266   \n",
       "  \n",
       "     relative_size_complement  coverage_sg  coverage_complement  \\\n",
       "  0                  0.966628     0.070868             0.929132   \n",
       "  1                  0.966675     0.066987             0.933013   \n",
       "  2                  0.966666     0.066487             0.933513   \n",
       "  3                  0.966699     0.062788             0.937212   \n",
       "  4                  0.967016     0.059652             0.940348   \n",
       "  5                  0.966657     0.059589             0.940411   \n",
       "  6                  0.966666     0.059489             0.940511   \n",
       "  7                  0.966705     0.059298             0.940702   \n",
       "  8                  0.966666     0.058243             0.941757   \n",
       "  9                  0.966734     0.058053             0.941947   \n",
       "  \n",
       "     target_share_sg  target_share_complement  target_share_dataset      lift  \n",
       "  0         0.693067                 0.313709              0.326369  2.123567  \n",
       "  1         0.656044                 0.315004              0.326369  2.010129  \n",
       "  2         0.650974                 0.315176              0.326369  1.994596  \n",
       "  3         0.615357                 0.316414              0.326369  1.885464  \n",
       "  4         0.590251                 0.317368              0.326369  1.808538  \n",
       "  5         0.583274                 0.317508              0.326369  1.787160  \n",
       "  6         0.582451                 0.317539              0.326369  1.784638  \n",
       "  7         0.581254                 0.317590              0.326369  1.780972  \n",
       "  8         0.570259                 0.317959              0.326369  1.747282  \n",
       "  9         0.569556                 0.318001              0.326369  1.745128  },\n",
       " {'10_tv_GA_0_SDFS':     quality                                           subgroup  size_sg  \\\n",
       "  0  0.875932  012: [0.03:0.04[ AND average_clustering: [0.01...      1.0   \n",
       "  1  0.875932  012: [0.06:0.07[ AND average_degree: [1.96:2.0...      1.0   \n",
       "  2  0.875932  average_clustering: [0.01:0.01[ AND average_de...      1.0   \n",
       "  3  0.875932  021D: [0.00:0.00[ AND average_degree<1.82 AND ...      1.0   \n",
       "  4  0.875932  021C: [0.00:0.00[ AND average_degree: [1.96:2....      1.0   \n",
       "  5  0.875932  average_clustering: [0.01:0.01[ AND average_de...      1.0   \n",
       "  6  0.875932  average_degree<1.82 AND betweenness_centraliza...      1.0   \n",
       "  7  0.875932  021U: [0.00:0.00[ AND average_degree: [1.96:2....      1.0   \n",
       "  8  0.875932  average_degree: [1.96:2.0[ AND betweenness_cen...      1.0   \n",
       "  9  0.875932  030T: [0.00:0.00[ AND average_clustering: [0.0...      1.0   \n",
       "  \n",
       "     size_dataset  positives_sg  positives_dataset  size_complement  \\\n",
       "  0      337106.0           1.0            41824.0         337105.0   \n",
       "  1      337106.0           1.0            41824.0         337105.0   \n",
       "  2      337106.0           1.0            41824.0         337105.0   \n",
       "  3      337106.0           1.0            41824.0         337105.0   \n",
       "  4      337106.0           1.0            41824.0         337105.0   \n",
       "  5      337106.0           1.0            41824.0         337105.0   \n",
       "  6      337106.0           1.0            41824.0         337105.0   \n",
       "  7      337106.0           1.0            41824.0         337105.0   \n",
       "  8      337106.0           1.0            41824.0         337105.0   \n",
       "  9      337106.0           1.0            41824.0         337105.0   \n",
       "  \n",
       "     relative_size_sg  relative_size_complement  coverage_sg  \\\n",
       "  0          0.000003                  0.999997     0.000024   \n",
       "  1          0.000003                  0.999997     0.000024   \n",
       "  2          0.000003                  0.999997     0.000024   \n",
       "  3          0.000003                  0.999997     0.000024   \n",
       "  4          0.000003                  0.999997     0.000024   \n",
       "  5          0.000003                  0.999997     0.000024   \n",
       "  6          0.000003                  0.999997     0.000024   \n",
       "  7          0.000003                  0.999997     0.000024   \n",
       "  8          0.000003                  0.999997     0.000024   \n",
       "  9          0.000003                  0.999997     0.000024   \n",
       "  \n",
       "     coverage_complement  target_share_sg  target_share_complement  \\\n",
       "  0             0.999976              1.0                 0.124065   \n",
       "  1             0.999976              1.0                 0.124065   \n",
       "  2             0.999976              1.0                 0.124065   \n",
       "  3             0.999976              1.0                 0.124065   \n",
       "  4             0.999976              1.0                 0.124065   \n",
       "  5             0.999976              1.0                 0.124065   \n",
       "  6             0.999976              1.0                 0.124065   \n",
       "  7             0.999976              1.0                 0.124065   \n",
       "  8             0.999976              1.0                 0.124065   \n",
       "  9             0.999976              1.0                 0.124065   \n",
       "  \n",
       "     target_share_dataset      lift  \n",
       "  0              0.124068  8.060109  \n",
       "  1              0.124068  8.060109  \n",
       "  2              0.124068  8.060109  \n",
       "  3              0.124068  8.060109  \n",
       "  4              0.124068  8.060109  \n",
       "  5              0.124068  8.060109  \n",
       "  6              0.124068  8.060109  \n",
       "  7              0.124068  8.060109  \n",
       "  8              0.124068  8.060109  \n",
       "  9              0.124068  8.060109  ,\n",
       "  '10_tv_SQF_0_BFS':     quality                                           subgroup  size_sg  \\\n",
       "  0  0.875932  012: [0.02:0.02[ AND 021C: [0.00:0.00[ AND 021...      1.0   \n",
       "  1  0.875932  012: [0.02:0.02[ AND 021C: [0.00:0.00[ AND 021...      1.0   \n",
       "  2  0.875932  012: [0.02:0.02[ AND 021C: [0.00:0.00[ AND 030...      1.0   \n",
       "  3  0.875932  012: [0.02:0.02[ AND 021C: [0.00:0.00[ AND 030...      1.0   \n",
       "  4  0.875932  012: [0.02:0.03[ AND 021D: [0.00:0.00[ AND 021...      1.0   \n",
       "  5  0.875932             012: [0.02:0.02[ AND 030T: [0.00:0.00[      1.0   \n",
       "  6  0.875932  012: [0.02:0.02[ AND 021U<0.00 AND 030T: [0.00...      1.0   \n",
       "  7  0.875932  012: [0.02:0.02[ AND 021D: [0.00:0.00[ AND 030...      1.0   \n",
       "  8  0.875932  012: [0.02:0.03[ AND 021D: [0.00:0.00[ AND 021...      1.0   \n",
       "  9  0.875932  012: [0.02:0.03[ AND 021D: [0.00:0.00[ AND 021...      1.0   \n",
       "  \n",
       "     size_dataset  positives_sg  positives_dataset  size_complement  \\\n",
       "  0      337106.0           1.0            41824.0         337105.0   \n",
       "  1      337106.0           1.0            41824.0         337105.0   \n",
       "  2      337106.0           1.0            41824.0         337105.0   \n",
       "  3      337106.0           1.0            41824.0         337105.0   \n",
       "  4      337106.0           1.0            41824.0         337105.0   \n",
       "  5      337106.0           1.0            41824.0         337105.0   \n",
       "  6      337106.0           1.0            41824.0         337105.0   \n",
       "  7      337106.0           1.0            41824.0         337105.0   \n",
       "  8      337106.0           1.0            41824.0         337105.0   \n",
       "  9      337106.0           1.0            41824.0         337105.0   \n",
       "  \n",
       "     relative_size_sg  relative_size_complement  coverage_sg  \\\n",
       "  0          0.000003                  0.999997     0.000024   \n",
       "  1          0.000003                  0.999997     0.000024   \n",
       "  2          0.000003                  0.999997     0.000024   \n",
       "  3          0.000003                  0.999997     0.000024   \n",
       "  4          0.000003                  0.999997     0.000024   \n",
       "  5          0.000003                  0.999997     0.000024   \n",
       "  6          0.000003                  0.999997     0.000024   \n",
       "  7          0.000003                  0.999997     0.000024   \n",
       "  8          0.000003                  0.999997     0.000024   \n",
       "  9          0.000003                  0.999997     0.000024   \n",
       "  \n",
       "     coverage_complement  target_share_sg  target_share_complement  \\\n",
       "  0             0.999976              1.0                 0.124065   \n",
       "  1             0.999976              1.0                 0.124065   \n",
       "  2             0.999976              1.0                 0.124065   \n",
       "  3             0.999976              1.0                 0.124065   \n",
       "  4             0.999976              1.0                 0.124065   \n",
       "  5             0.999976              1.0                 0.124065   \n",
       "  6             0.999976              1.0                 0.124065   \n",
       "  7             0.999976              1.0                 0.124065   \n",
       "  8             0.999976              1.0                 0.124065   \n",
       "  9             0.999976              1.0                 0.124065   \n",
       "  \n",
       "     target_share_dataset      lift  \n",
       "  0              0.124068  8.060109  \n",
       "  1              0.124068  8.060109  \n",
       "  2              0.124068  8.060109  \n",
       "  3              0.124068  8.060109  \n",
       "  4              0.124068  8.060109  \n",
       "  5              0.124068  8.060109  \n",
       "  6              0.124068  8.060109  \n",
       "  7              0.124068  8.060109  \n",
       "  8              0.124068  8.060109  \n",
       "  9              0.124068  8.060109  },\n",
       " {'10_tv_GA_0.5_SDFS':     quality                                           subgroup   size_sg  \\\n",
       "  0  0.011220                                   021D: [0.0:0.00[  160721.0   \n",
       "  1  0.008668                                 transitivity>=0.50   31903.0   \n",
       "  2  0.007487   021C: [0.00:0.00[ AND degree_centralization<0.18     920.0   \n",
       "  3  0.006548   021D: [0.00:0.00[ AND degree_centralization<0.18     708.0   \n",
       "  4  0.006358   021D: [0.00:0.00[ AND degree_centralization<0.18     196.0   \n",
       "  5  0.005911        degree_centralization<0.18 AND density<0.00     782.0   \n",
       "  6  0.005833    012: [0.02:0.02[ AND degree_centralization<0.18     838.0   \n",
       "  7  0.005547                    021D: [0.00:0.00[ AND 021U<0.00    1256.0   \n",
       "  8  0.005118  betweenness_centralization: [0.00:0.00[ AND de...     192.0   \n",
       "  9  0.005115  betweenness_centralization: [0.00:0.00[ AND de...     146.0   \n",
       "  \n",
       "     size_dataset  positives_sg  positives_dataset  size_complement  \\\n",
       "  0      337106.0       22552.0            41824.0         176385.0   \n",
       "  1      337106.0        4857.0            41824.0         305203.0   \n",
       "  2      337106.0         246.0            41824.0         336186.0   \n",
       "  3      337106.0         189.0            41824.0         336398.0   \n",
       "  4      337106.0          76.0            41824.0         336910.0   \n",
       "  5      337106.0         193.0            41824.0         336324.0   \n",
       "  6      337106.0         202.0            41824.0         336268.0   \n",
       "  7      337106.0         276.0            41824.0         335850.0   \n",
       "  8      337106.0          65.0            41824.0         336914.0   \n",
       "  9      337106.0          54.0            41824.0         336960.0   \n",
       "  \n",
       "     relative_size_sg  relative_size_complement  coverage_sg  \\\n",
       "  0          0.476767                  0.523233     0.539212   \n",
       "  1          0.094638                  0.905362     0.116129   \n",
       "  2          0.002729                  0.997271     0.005882   \n",
       "  3          0.002100                  0.997900     0.004519   \n",
       "  4          0.000581                  0.999419     0.001817   \n",
       "  5          0.002320                  0.997680     0.004615   \n",
       "  6          0.002486                  0.997514     0.004830   \n",
       "  7          0.003726                  0.996274     0.006599   \n",
       "  8          0.000570                  0.999430     0.001554   \n",
       "  9          0.000433                  0.999567     0.001291   \n",
       "  \n",
       "     coverage_complement  target_share_sg  target_share_complement  \\\n",
       "  0             0.460788         0.140318                 0.109261   \n",
       "  1             0.883871         0.152243                 0.121123   \n",
       "  2             0.994118         0.267391                 0.123676   \n",
       "  3             0.995481         0.266949                 0.123767   \n",
       "  4             0.998183         0.387755                 0.123914   \n",
       "  5             0.995385         0.246803                 0.123782   \n",
       "  6             0.995170         0.241050                 0.123776   \n",
       "  7             0.993401         0.219745                 0.123710   \n",
       "  8             0.998446         0.338542                 0.123946   \n",
       "  9             0.998709         0.369863                 0.123961   \n",
       "  \n",
       "     target_share_dataset      lift  \n",
       "  0              0.124068  1.130976  \n",
       "  1              0.124068  1.227093  \n",
       "  2              0.124068  2.155203  \n",
       "  3              0.124068  2.151639  \n",
       "  4              0.124068  3.125348  \n",
       "  5              0.124068  1.989260  \n",
       "  6              0.124068  1.942890  \n",
       "  7              0.124068  1.771170  \n",
       "  8              0.124068  2.728683  \n",
       "  9              0.124068  2.981136  ,\n",
       "  '10_tv_SQF_0.5_BFS':     quality                                           subgroup   size_sg  \\\n",
       "  0  0.011220                                   021D: [0.0:0.00[  160721.0   \n",
       "  1  0.008668                                 transitivity>=0.50   31903.0   \n",
       "  2  0.008668            021D: [0.0:0.00[ AND transitivity>=0.50   31903.0   \n",
       "  3  0.008219  021D: [0.0:0.00[ AND 030T: [0.0:0.00[ AND aver...  128818.0   \n",
       "  4  0.008219              021D: [0.0:0.00[ AND 030T: [0.0:0.00[  128818.0   \n",
       "  5  0.008219  021D: [0.0:0.00[ AND average_clustering: [0.0:...  128818.0   \n",
       "  6  0.008219  021D: [0.0:0.00[ AND 030T: [0.0:0.00[ AND tran...  128818.0   \n",
       "  7  0.008219  021D: [0.0:0.00[ AND average_clustering: [0.0:...  128818.0   \n",
       "  8  0.008219      021D: [0.0:0.00[ AND transitivity: [0.0:0.00[  128818.0   \n",
       "  9  0.007605  average_degree: [2.0:2.00[ AND transitivity>=0.50   25429.0   \n",
       "  \n",
       "     size_dataset  positives_sg  positives_dataset  size_complement  \\\n",
       "  0      337106.0       22552.0            41824.0         176385.0   \n",
       "  1      337106.0        4857.0            41824.0         305203.0   \n",
       "  2      337106.0        4857.0            41824.0         305203.0   \n",
       "  3      337106.0       17695.0            41824.0         208288.0   \n",
       "  4      337106.0       17695.0            41824.0         208288.0   \n",
       "  5      337106.0       17695.0            41824.0         208288.0   \n",
       "  6      337106.0       17695.0            41824.0         208288.0   \n",
       "  7      337106.0       17695.0            41824.0         208288.0   \n",
       "  8      337106.0       17695.0            41824.0         208288.0   \n",
       "  9      337106.0        3859.0            41824.0         311677.0   \n",
       "  \n",
       "     relative_size_sg  relative_size_complement  coverage_sg  \\\n",
       "  0          0.476767                  0.523233     0.539212   \n",
       "  1          0.094638                  0.905362     0.116129   \n",
       "  2          0.094638                  0.905362     0.116129   \n",
       "  3          0.382129                  0.617871     0.423082   \n",
       "  4          0.382129                  0.617871     0.423082   \n",
       "  5          0.382129                  0.617871     0.423082   \n",
       "  6          0.382129                  0.617871     0.423082   \n",
       "  7          0.382129                  0.617871     0.423082   \n",
       "  8          0.382129                  0.617871     0.423082   \n",
       "  9          0.075433                  0.924567     0.092268   \n",
       "  \n",
       "     coverage_complement  target_share_sg  target_share_complement  \\\n",
       "  0             0.460788         0.140318                 0.109261   \n",
       "  1             0.883871         0.152243                 0.121123   \n",
       "  2             0.883871         0.152243                 0.121123   \n",
       "  3             0.576918         0.137364                 0.115844   \n",
       "  4             0.576918         0.137364                 0.115844   \n",
       "  5             0.576918         0.137364                 0.115844   \n",
       "  6             0.576918         0.137364                 0.115844   \n",
       "  7             0.576918         0.137364                 0.115844   \n",
       "  8             0.576918         0.137364                 0.115844   \n",
       "  9             0.907732         0.151756                 0.121809   \n",
       "  \n",
       "     target_share_dataset      lift  \n",
       "  0              0.124068  1.130976  \n",
       "  1              0.124068  1.227093  \n",
       "  2              0.124068  1.227093  \n",
       "  3              0.124068  1.107172  \n",
       "  4              0.124068  1.107172  \n",
       "  5              0.124068  1.107172  \n",
       "  6              0.124068  1.107172  \n",
       "  7              0.124068  1.107172  \n",
       "  8              0.124068  1.107172  \n",
       "  9              0.124068  1.223169  },\n",
       " {'10_tv_GA_1_SDFS':     quality                                         subgroup   size_sg  \\\n",
       "  0  0.007747                                 021D: [0.0:0.00[  160721.0   \n",
       "  1  0.002666                               transitivity>=0.50   31903.0   \n",
       "  2  0.001431                       average_degree: [2.0:2.00[   59173.0   \n",
       "  3  0.001181                                 030T: [0.0:0.00[  195625.0   \n",
       "  4  0.001181                         transitivity: [0.0:0.00[  195625.0   \n",
       "  5  0.001181                   average_clustering: [0.0:0.00[  195625.0   \n",
       "  6  0.000993                        longest_path: [0.18:0.19[   19603.0   \n",
       "  7  0.000870                        longest_path: [0.15:0.16[   15102.0   \n",
       "  8  0.000863  021D: [0.0:0.00[ AND average_degree: [2.0:2.00[   25429.0   \n",
       "  9  0.000802                        longest_path: [0.13:0.14[   11523.0   \n",
       "  \n",
       "     size_dataset  positives_sg  positives_dataset  size_complement  \\\n",
       "  0      337106.0       22552.0            41824.0         176385.0   \n",
       "  1      337106.0        4857.0            41824.0         305203.0   \n",
       "  2      337106.0        7824.0            41824.0         277933.0   \n",
       "  3      337106.0       24669.0            41824.0         141481.0   \n",
       "  4      337106.0       24669.0            41824.0         141481.0   \n",
       "  5      337106.0       24669.0            41824.0         141481.0   \n",
       "  6      337106.0        2767.0            41824.0         317503.0   \n",
       "  7      337106.0        2167.0            41824.0         322004.0   \n",
       "  8      337106.0        3859.0            41824.0         311677.0   \n",
       "  9      337106.0        1700.0            41824.0         325583.0   \n",
       "  \n",
       "     relative_size_sg  relative_size_complement  coverage_sg  \\\n",
       "  0          0.476767                  0.523233     0.539212   \n",
       "  1          0.094638                  0.905362     0.116129   \n",
       "  2          0.175532                  0.824468     0.187070   \n",
       "  3          0.580307                  0.419693     0.589829   \n",
       "  4          0.580307                  0.419693     0.589829   \n",
       "  5          0.580307                  0.419693     0.589829   \n",
       "  6          0.058151                  0.941849     0.066158   \n",
       "  7          0.044799                  0.955201     0.051812   \n",
       "  8          0.075433                  0.924567     0.092268   \n",
       "  9          0.034182                  0.965818     0.040647   \n",
       "  \n",
       "     coverage_complement  target_share_sg  target_share_complement  \\\n",
       "  0             0.460788         0.140318                 0.109261   \n",
       "  1             0.883871         0.152243                 0.121123   \n",
       "  2             0.812930         0.132222                 0.122332   \n",
       "  3             0.410171         0.126104                 0.121253   \n",
       "  4             0.410171         0.126104                 0.121253   \n",
       "  5             0.410171         0.126104                 0.121253   \n",
       "  6             0.933842         0.141152                 0.123013   \n",
       "  7             0.948188         0.143491                 0.123157   \n",
       "  8             0.907732         0.151756                 0.121809   \n",
       "  9             0.959353         0.147531                 0.123237   \n",
       "  \n",
       "     target_share_dataset      lift  \n",
       "  0              0.124068  1.130976  \n",
       "  1              0.124068  1.227093  \n",
       "  2              0.124068  1.065727  \n",
       "  3              0.124068  1.016408  \n",
       "  4              0.124068  1.016408  \n",
       "  5              0.124068  1.016408  \n",
       "  6              0.124068  1.137699  \n",
       "  7              0.124068  1.156553  \n",
       "  8              0.124068  1.223169  \n",
       "  9              0.124068  1.189116  ,\n",
       "  '10_tv_SQF_1_BFS':     quality                                           subgroup   size_sg  \\\n",
       "  0  0.007747                                   021D: [0.0:0.00[  160721.0   \n",
       "  1  0.005081  021D: [0.0:0.00[ AND average_clustering: [0.0:...  128818.0   \n",
       "  2  0.005081  021D: [0.0:0.00[ AND 030T: [0.0:0.00[ AND aver...  128818.0   \n",
       "  3  0.005081  021D: [0.0:0.00[ AND average_clustering: [0.0:...  128818.0   \n",
       "  4  0.005081              021D: [0.0:0.00[ AND 030T: [0.0:0.00[  128818.0   \n",
       "  5  0.005081      021D: [0.0:0.00[ AND transitivity: [0.0:0.00[  128818.0   \n",
       "  6  0.005081  021D: [0.0:0.00[ AND 030T: [0.0:0.00[ AND tran...  128818.0   \n",
       "  7  0.002666            021D: [0.0:0.00[ AND transitivity>=0.50   31903.0   \n",
       "  8  0.002666                                 transitivity>=0.50   31903.0   \n",
       "  9  0.002089  average_degree: [2.0:2.00[ AND transitivity>=0.50   25429.0   \n",
       "  \n",
       "     size_dataset  positives_sg  positives_dataset  size_complement  \\\n",
       "  0      337106.0       22552.0            41824.0         176385.0   \n",
       "  1      337106.0       17695.0            41824.0         208288.0   \n",
       "  2      337106.0       17695.0            41824.0         208288.0   \n",
       "  3      337106.0       17695.0            41824.0         208288.0   \n",
       "  4      337106.0       17695.0            41824.0         208288.0   \n",
       "  5      337106.0       17695.0            41824.0         208288.0   \n",
       "  6      337106.0       17695.0            41824.0         208288.0   \n",
       "  7      337106.0        4857.0            41824.0         305203.0   \n",
       "  8      337106.0        4857.0            41824.0         305203.0   \n",
       "  9      337106.0        3859.0            41824.0         311677.0   \n",
       "  \n",
       "     relative_size_sg  relative_size_complement  coverage_sg  \\\n",
       "  0          0.476767                  0.523233     0.539212   \n",
       "  1          0.382129                  0.617871     0.423082   \n",
       "  2          0.382129                  0.617871     0.423082   \n",
       "  3          0.382129                  0.617871     0.423082   \n",
       "  4          0.382129                  0.617871     0.423082   \n",
       "  5          0.382129                  0.617871     0.423082   \n",
       "  6          0.382129                  0.617871     0.423082   \n",
       "  7          0.094638                  0.905362     0.116129   \n",
       "  8          0.094638                  0.905362     0.116129   \n",
       "  9          0.075433                  0.924567     0.092268   \n",
       "  \n",
       "     coverage_complement  target_share_sg  target_share_complement  \\\n",
       "  0             0.460788         0.140318                 0.109261   \n",
       "  1             0.576918         0.137364                 0.115844   \n",
       "  2             0.576918         0.137364                 0.115844   \n",
       "  3             0.576918         0.137364                 0.115844   \n",
       "  4             0.576918         0.137364                 0.115844   \n",
       "  5             0.576918         0.137364                 0.115844   \n",
       "  6             0.576918         0.137364                 0.115844   \n",
       "  7             0.883871         0.152243                 0.121123   \n",
       "  8             0.883871         0.152243                 0.121123   \n",
       "  9             0.907732         0.151756                 0.121809   \n",
       "  \n",
       "     target_share_dataset      lift  \n",
       "  0              0.124068  1.130976  \n",
       "  1              0.124068  1.107172  \n",
       "  2              0.124068  1.107172  \n",
       "  3              0.124068  1.107172  \n",
       "  4              0.124068  1.107172  \n",
       "  5              0.124068  1.107172  \n",
       "  6              0.124068  1.107172  \n",
       "  7              0.124068  1.227093  \n",
       "  8              0.124068  1.227093  \n",
       "  9              0.124068  1.223169  },\n",
       " {'10_v_GA_0_SDFS':     quality                                           subgroup  size_sg  \\\n",
       "  0  0.787553  012: [0.31:0.35[ AND average_degree<1.82 AND d...      1.0   \n",
       "  1  0.787553  021C: [0.01:0.01[ AND average_degree<1.82 AND ...      1.0   \n",
       "  2  0.787553  021C: [0.01:0.01[ AND average_degree<1.82 AND ...      1.0   \n",
       "  3  0.787553  average_degree<1.82 AND betweenness_centraliza...      1.0   \n",
       "  4  0.787553  021C: [0.01:0.01[ AND average_degree<1.82 AND ...      1.0   \n",
       "  5  0.787553  average_degree<1.82 AND density: [0.07:0.07[ A...      1.0   \n",
       "  6  0.787553  021U: [0.03:0.04[ AND average_degree<1.82 AND ...      2.0   \n",
       "  7  0.787553  average_degree<1.82 AND density: [0.06:0.07[ A...      1.0   \n",
       "  8  0.787553  average_degree<1.82 AND degree_centralization<...      1.0   \n",
       "  9  0.787553  021U: [0.07:0.08[ AND average_degree<1.82 AND ...      1.0   \n",
       "  \n",
       "     size_dataset  positives_sg  positives_dataset  size_complement  \\\n",
       "  0      337106.0           1.0            71617.0         337105.0   \n",
       "  1      337106.0           1.0            71617.0         337105.0   \n",
       "  2      337106.0           1.0            71617.0         337105.0   \n",
       "  3      337106.0           1.0            71617.0         337105.0   \n",
       "  4      337106.0           1.0            71617.0         337105.0   \n",
       "  5      337106.0           1.0            71617.0         337105.0   \n",
       "  6      337106.0           2.0            71617.0         337104.0   \n",
       "  7      337106.0           1.0            71617.0         337105.0   \n",
       "  8      337106.0           1.0            71617.0         337105.0   \n",
       "  9      337106.0           1.0            71617.0         337105.0   \n",
       "  \n",
       "     relative_size_sg  relative_size_complement  coverage_sg  \\\n",
       "  0          0.000003                  0.999997     0.000014   \n",
       "  1          0.000003                  0.999997     0.000014   \n",
       "  2          0.000003                  0.999997     0.000014   \n",
       "  3          0.000003                  0.999997     0.000014   \n",
       "  4          0.000003                  0.999997     0.000014   \n",
       "  5          0.000003                  0.999997     0.000014   \n",
       "  6          0.000006                  0.999994     0.000028   \n",
       "  7          0.000003                  0.999997     0.000014   \n",
       "  8          0.000003                  0.999997     0.000014   \n",
       "  9          0.000003                  0.999997     0.000014   \n",
       "  \n",
       "     coverage_complement  target_share_sg  target_share_complement  \\\n",
       "  0             0.999986              1.0                 0.212444   \n",
       "  1             0.999986              1.0                 0.212444   \n",
       "  2             0.999986              1.0                 0.212444   \n",
       "  3             0.999986              1.0                 0.212444   \n",
       "  4             0.999986              1.0                 0.212444   \n",
       "  5             0.999986              1.0                 0.212444   \n",
       "  6             0.999972              1.0                 0.212442   \n",
       "  7             0.999986              1.0                 0.212444   \n",
       "  8             0.999986              1.0                 0.212444   \n",
       "  9             0.999986              1.0                 0.212444   \n",
       "  \n",
       "     target_share_dataset      lift  \n",
       "  0              0.212447  4.707067  \n",
       "  1              0.212447  4.707067  \n",
       "  2              0.212447  4.707067  \n",
       "  3              0.212447  4.707067  \n",
       "  4              0.212447  4.707067  \n",
       "  5              0.212447  4.707067  \n",
       "  6              0.212447  4.707067  \n",
       "  7              0.212447  4.707067  \n",
       "  8              0.212447  4.707067  \n",
       "  9              0.212447  4.707067  ,\n",
       "  '10_v_SQF_0_BFS':     quality                                           subgroup  size_sg  \\\n",
       "  0  0.787553  012: [0.02:0.02[ AND 021C: [0.00:0.00[ AND 021...      1.0   \n",
       "  1  0.787553  012: [0.02:0.02[ AND 021C: [0.00:0.00[ AND 021...      1.0   \n",
       "  2  0.787553  012: [0.02:0.02[ AND 021D: [0.0:0.00[ AND 030T...      1.0   \n",
       "  3  0.787553  012: [0.02:0.02[ AND 021D: [0.0:0.00[ AND 030T...      1.0   \n",
       "  4  0.787553  012: [0.02:0.02[ AND 021C: [0.00:0.00[ AND 021...      1.0   \n",
       "  5  0.787553                    average_clustering: [0.00:0.00[      1.0   \n",
       "  6  0.787553  012: [0.02:0.02[ AND 021D: [0.00:0.00[ AND 021...      1.0   \n",
       "  7  0.787553                    average_clustering: [0.00:0.00[      1.0   \n",
       "  8  0.787553  012: [0.02:0.02[ AND 021D: [0.00:0.00[ AND 021...      1.0   \n",
       "  9  0.787553  012: [0.02:0.02[ AND 021U: [0.02:0.02[ AND 030...      1.0   \n",
       "  \n",
       "     size_dataset  positives_sg  positives_dataset  size_complement  \\\n",
       "  0      337106.0           1.0            71617.0         337105.0   \n",
       "  1      337106.0           1.0            71617.0         337105.0   \n",
       "  2      337106.0           1.0            71617.0         337105.0   \n",
       "  3      337106.0           1.0            71617.0         337105.0   \n",
       "  4      337106.0           1.0            71617.0         337105.0   \n",
       "  5      337106.0           1.0            71617.0         337105.0   \n",
       "  6      337106.0           1.0            71617.0         337105.0   \n",
       "  7      337106.0           1.0            71617.0         337105.0   \n",
       "  8      337106.0           1.0            71617.0         337105.0   \n",
       "  9      337106.0           1.0            71617.0         337105.0   \n",
       "  \n",
       "     relative_size_sg  relative_size_complement  coverage_sg  \\\n",
       "  0          0.000003                  0.999997     0.000014   \n",
       "  1          0.000003                  0.999997     0.000014   \n",
       "  2          0.000003                  0.999997     0.000014   \n",
       "  3          0.000003                  0.999997     0.000014   \n",
       "  4          0.000003                  0.999997     0.000014   \n",
       "  5          0.000003                  0.999997     0.000014   \n",
       "  6          0.000003                  0.999997     0.000014   \n",
       "  7          0.000003                  0.999997     0.000014   \n",
       "  8          0.000003                  0.999997     0.000014   \n",
       "  9          0.000003                  0.999997     0.000014   \n",
       "  \n",
       "     coverage_complement  target_share_sg  target_share_complement  \\\n",
       "  0             0.999986              1.0                 0.212444   \n",
       "  1             0.999986              1.0                 0.212444   \n",
       "  2             0.999986              1.0                 0.212444   \n",
       "  3             0.999986              1.0                 0.212444   \n",
       "  4             0.999986              1.0                 0.212444   \n",
       "  5             0.999986              1.0                 0.212444   \n",
       "  6             0.999986              1.0                 0.212444   \n",
       "  7             0.999986              1.0                 0.212444   \n",
       "  8             0.999986              1.0                 0.212444   \n",
       "  9             0.999986              1.0                 0.212444   \n",
       "  \n",
       "     target_share_dataset      lift  \n",
       "  0              0.212447  4.707067  \n",
       "  1              0.212447  4.707067  \n",
       "  2              0.212447  4.707067  \n",
       "  3              0.212447  4.707067  \n",
       "  4              0.212447  4.707067  \n",
       "  5              0.212447  4.707067  \n",
       "  6              0.212447  4.707067  \n",
       "  7              0.212447  4.707067  \n",
       "  8              0.212447  4.707067  \n",
       "  9              0.212447  4.707067  },\n",
       " {'10_v_GA_0.5_SDFS':     quality              subgroup  size_sg  size_dataset  positives_sg  \\\n",
       "  0  0.015244  density: [0.01:0.01[  11243.0      337106.0        3327.0   \n",
       "  1  0.015082  density: [0.00:0.01[  11227.0      337106.0        3313.0   \n",
       "  2  0.012860     021C: [0.00:0.00[  11238.0      337106.0        3179.0   \n",
       "  3  0.012458     021C: [0.00:0.00[  11237.0      337106.0        3154.0   \n",
       "  4  0.012350     030T: [0.00:0.00[  11061.0      337106.0        3104.0   \n",
       "  5  0.011853  density: [0.01:0.01[  11238.0      337106.0        3117.0   \n",
       "  6  0.011385      012: [0.03:0.04[  11237.0      337106.0        3088.0   \n",
       "  7  0.011190      012: [0.04:0.05[  11237.0      337106.0        3076.0   \n",
       "  8  0.010509     030T: [0.00:0.00[  11299.0      337106.0        3049.0   \n",
       "  9  0.009911     021D: [0.00:0.00[  11236.0      337106.0        2997.0   \n",
       "  \n",
       "     positives_dataset  size_complement  relative_size_sg  \\\n",
       "  0            71617.0         325863.0          0.033352   \n",
       "  1            71617.0         325879.0          0.033304   \n",
       "  2            71617.0         325868.0          0.033337   \n",
       "  3            71617.0         325869.0          0.033334   \n",
       "  4            71617.0         326045.0          0.032812   \n",
       "  5            71617.0         325868.0          0.033337   \n",
       "  6            71617.0         325869.0          0.033334   \n",
       "  7            71617.0         325869.0          0.033334   \n",
       "  8            71617.0         325807.0          0.033518   \n",
       "  9            71617.0         325870.0          0.033331   \n",
       "  \n",
       "     relative_size_complement  coverage_sg  coverage_complement  \\\n",
       "  0                  0.966648     0.046455             0.953545   \n",
       "  1                  0.966696     0.046260             0.953740   \n",
       "  2                  0.966663     0.044389             0.955611   \n",
       "  3                  0.966666     0.044040             0.955960   \n",
       "  4                  0.967188     0.043342             0.956658   \n",
       "  5                  0.966663     0.043523             0.956477   \n",
       "  6                  0.966666     0.043118             0.956882   \n",
       "  7                  0.966666     0.042951             0.957049   \n",
       "  8                  0.966482     0.042574             0.957426   \n",
       "  9                  0.966669     0.041848             0.958152   \n",
       "  \n",
       "     target_share_sg  target_share_complement  target_share_dataset      lift  \n",
       "  0         0.295917                 0.209567              0.212447  1.392903  \n",
       "  1         0.295092                 0.209599              0.212447  1.389019  \n",
       "  2         0.282880                 0.210018              0.212447  1.331533  \n",
       "  3         0.280680                 0.210094              0.212447  1.321179  \n",
       "  4         0.280626                 0.210134              0.212447  1.320924  \n",
       "  5         0.277363                 0.210208              0.212447  1.305564  \n",
       "  6         0.274806                 0.210296              0.212447  1.293532  \n",
       "  7         0.273739                 0.210333              0.212447  1.288506  \n",
       "  8         0.269847                 0.210456              0.212447  1.270187  \n",
       "  9         0.266732                 0.210575              0.212447  1.255525  ,\n",
       "  '10_v_SQF_0.5_BFS':     quality                                           subgroup  size_sg  \\\n",
       "  0  0.016171          012: [0.04:0.05[ AND density: [0.01:0.01[   5588.0   \n",
       "  1  0.015244                               density: [0.01:0.01[  11243.0   \n",
       "  2  0.015082                               density: [0.00:0.01[  11227.0   \n",
       "  3  0.014395  012: [0.04:0.05[ AND 021C: [0.00:0.00[ AND den...   4031.0   \n",
       "  4  0.014300          012: [0.03:0.04[ AND density: [0.00:0.01[   4420.0   \n",
       "  5  0.014197         021D: [0.00:0.00[ AND density: [0.00:0.01[   2862.0   \n",
       "  6  0.014177         021U: [0.00:0.00[ AND density: [0.00:0.01[   2651.0   \n",
       "  7  0.014021         021C: [0.00:0.00[ AND density: [0.01:0.01[   4485.0   \n",
       "  8  0.013545  012: [0.03:0.04[ AND 021C: [0.00:0.00[ AND den...   3612.0   \n",
       "  9  0.012996         021U: [0.00:0.00[ AND density: [0.01:0.01[   2849.0   \n",
       "  \n",
       "     size_dataset  positives_sg  positives_dataset  size_complement  \\\n",
       "  0      337106.0        1889.0            71617.0         331518.0   \n",
       "  1      337106.0        3327.0            71617.0         325863.0   \n",
       "  2      337106.0        3313.0            71617.0         325879.0   \n",
       "  3      337106.0        1387.0            71617.0         333075.0   \n",
       "  4      337106.0        1491.0            71617.0         332686.0   \n",
       "  5      337106.0        1049.0            71617.0         334244.0   \n",
       "  6      337106.0         987.0            71617.0         334455.0   \n",
       "  7      337106.0        1498.0            71617.0         332621.0   \n",
       "  8      337106.0        1240.0            71617.0         333494.0   \n",
       "  9      337106.0        1008.0            71617.0         334257.0   \n",
       "  \n",
       "     relative_size_sg  relative_size_complement  coverage_sg  \\\n",
       "  0          0.016576                  0.983424     0.026376   \n",
       "  1          0.033352                  0.966648     0.046455   \n",
       "  2          0.033304                  0.966696     0.046260   \n",
       "  3          0.011958                  0.988042     0.019367   \n",
       "  4          0.013112                  0.986888     0.020819   \n",
       "  5          0.008490                  0.991510     0.014647   \n",
       "  6          0.007864                  0.992136     0.013782   \n",
       "  7          0.013304                  0.986696     0.020917   \n",
       "  8          0.010715                  0.989285     0.017314   \n",
       "  9          0.008451                  0.991549     0.014075   \n",
       "  \n",
       "     coverage_complement  target_share_sg  target_share_complement  \\\n",
       "  0             0.973624         0.338046                 0.210329   \n",
       "  1             0.953545         0.295917                 0.209567   \n",
       "  2             0.953740         0.295092                 0.209599   \n",
       "  3             0.980633         0.344083                 0.210853   \n",
       "  4             0.979181         0.337330                 0.210787   \n",
       "  5             0.985353         0.366527                 0.211127   \n",
       "  6             0.986218         0.372312                 0.211179   \n",
       "  7             0.979083         0.334002                 0.210807   \n",
       "  8             0.982686         0.343300                 0.211029   \n",
       "  9             0.985925         0.353808                 0.211242   \n",
       "  \n",
       "     target_share_dataset      lift  \n",
       "  0              0.212447  1.591204  \n",
       "  1              0.212447  1.392903  \n",
       "  2              0.212447  1.389019  \n",
       "  3              0.212447  1.619623  \n",
       "  4              0.212447  1.587836  \n",
       "  5              0.212447  1.725267  \n",
       "  6              0.212447  1.752499  \n",
       "  7              0.212447  1.572171  \n",
       "  8              0.212447  1.615937  \n",
       "  9              0.212447  1.665400  },\n",
       " {'10_v_GA_1_SDFS':     quality                        subgroup   size_sg  size_dataset  \\\n",
       "  0  0.005106                021D: [0.0:0.00[  160721.0      337106.0   \n",
       "  1  0.002784            density: [0.01:0.01[   11243.0      337106.0   \n",
       "  2  0.002752            density: [0.00:0.01[   11227.0      337106.0   \n",
       "  3  0.002670        transitivity: [0.0:0.00[  195625.0      337106.0   \n",
       "  4  0.002670  average_clustering: [0.0:0.00[  195625.0      337106.0   \n",
       "  5  0.002670                030T: [0.0:0.00[  195625.0      337106.0   \n",
       "  6  0.002348               021C: [0.00:0.00[   11238.0      337106.0   \n",
       "  7  0.002274               021C: [0.00:0.00[   11237.0      337106.0   \n",
       "  8  0.002237               030T: [0.00:0.00[   11061.0      337106.0   \n",
       "  9  0.002164            density: [0.01:0.01[   11238.0      337106.0   \n",
       "  \n",
       "     positives_sg  positives_dataset  size_complement  relative_size_sg  \\\n",
       "  0       35866.0            71617.0         176385.0          0.476767   \n",
       "  1        3327.0            71617.0         325863.0          0.033352   \n",
       "  2        3313.0            71617.0         325879.0          0.033304   \n",
       "  3       42460.0            71617.0         141481.0          0.580307   \n",
       "  4       42460.0            71617.0         141481.0          0.580307   \n",
       "  5       42460.0            71617.0         141481.0          0.580307   \n",
       "  6        3179.0            71617.0         325868.0          0.033337   \n",
       "  7        3154.0            71617.0         325869.0          0.033334   \n",
       "  8        3104.0            71617.0         326045.0          0.032812   \n",
       "  9        3117.0            71617.0         325868.0          0.033337   \n",
       "  \n",
       "     relative_size_complement  coverage_sg  coverage_complement  \\\n",
       "  0                  0.523233     0.500803             0.499197   \n",
       "  1                  0.966648     0.046455             0.953545   \n",
       "  2                  0.966696     0.046260             0.953740   \n",
       "  3                  0.419693     0.592876             0.407124   \n",
       "  4                  0.419693     0.592876             0.407124   \n",
       "  5                  0.419693     0.592876             0.407124   \n",
       "  6                  0.966663     0.044389             0.955611   \n",
       "  7                  0.966666     0.044040             0.955960   \n",
       "  8                  0.967188     0.043342             0.956658   \n",
       "  9                  0.966663     0.043523             0.956477   \n",
       "  \n",
       "     target_share_sg  target_share_complement  target_share_dataset      lift  \n",
       "  0         0.223157                 0.202687              0.212447  1.050414  \n",
       "  1         0.295917                 0.209567              0.212447  1.392903  \n",
       "  2         0.295092                 0.209599              0.212447  1.389019  \n",
       "  3         0.217048                 0.206084              0.212447  1.021659  \n",
       "  4         0.217048                 0.206084              0.212447  1.021659  \n",
       "  5         0.217048                 0.206084              0.212447  1.021659  \n",
       "  6         0.282880                 0.210018              0.212447  1.331533  \n",
       "  7         0.280680                 0.210094              0.212447  1.321179  \n",
       "  8         0.280626                 0.210134              0.212447  1.320924  \n",
       "  9         0.277363                 0.210208              0.212447  1.305564  ,\n",
       "  '10_v_SQF_1_BFS':     quality                                           subgroup   size_sg  \\\n",
       "  0  0.005106                                   021D: [0.0:0.00[  160721.0   \n",
       "  1  0.003406  021D: [0.0:0.00[ AND average_clustering: [0.0:...  128818.0   \n",
       "  2  0.003406  021D: [0.0:0.00[ AND 030T: [0.0:0.00[ AND aver...  128818.0   \n",
       "  3  0.003406              021D: [0.0:0.00[ AND 030T: [0.0:0.00[  128818.0   \n",
       "  4  0.003406  021D: [0.0:0.00[ AND average_clustering: [0.0:...  128818.0   \n",
       "  5  0.003406      021D: [0.0:0.00[ AND transitivity: [0.0:0.00[  128818.0   \n",
       "  6  0.003406  021D: [0.0:0.00[ AND 030T: [0.0:0.00[ AND tran...  128818.0   \n",
       "  7  0.002784                               density: [0.01:0.01[   11243.0   \n",
       "  8  0.002752                               density: [0.00:0.01[   11227.0   \n",
       "  9  0.002670                           transitivity: [0.0:0.00[  195625.0   \n",
       "  \n",
       "     size_dataset  positives_sg  positives_dataset  size_complement  \\\n",
       "  0      337106.0       35866.0            71617.0         176385.0   \n",
       "  1      337106.0       28515.0            71617.0         208288.0   \n",
       "  2      337106.0       28515.0            71617.0         208288.0   \n",
       "  3      337106.0       28515.0            71617.0         208288.0   \n",
       "  4      337106.0       28515.0            71617.0         208288.0   \n",
       "  5      337106.0       28515.0            71617.0         208288.0   \n",
       "  6      337106.0       28515.0            71617.0         208288.0   \n",
       "  7      337106.0        3327.0            71617.0         325863.0   \n",
       "  8      337106.0        3313.0            71617.0         325879.0   \n",
       "  9      337106.0       42460.0            71617.0         141481.0   \n",
       "  \n",
       "     relative_size_sg  relative_size_complement  coverage_sg  \\\n",
       "  0          0.476767                  0.523233     0.500803   \n",
       "  1          0.382129                  0.617871     0.398160   \n",
       "  2          0.382129                  0.617871     0.398160   \n",
       "  3          0.382129                  0.617871     0.398160   \n",
       "  4          0.382129                  0.617871     0.398160   \n",
       "  5          0.382129                  0.617871     0.398160   \n",
       "  6          0.382129                  0.617871     0.398160   \n",
       "  7          0.033352                  0.966648     0.046455   \n",
       "  8          0.033304                  0.966696     0.046260   \n",
       "  9          0.580307                  0.419693     0.592876   \n",
       "  \n",
       "     coverage_complement  target_share_sg  target_share_complement  \\\n",
       "  0             0.499197         0.223157                 0.202687   \n",
       "  1             0.601840         0.221359                 0.206935   \n",
       "  2             0.601840         0.221359                 0.206935   \n",
       "  3             0.601840         0.221359                 0.206935   \n",
       "  4             0.601840         0.221359                 0.206935   \n",
       "  5             0.601840         0.221359                 0.206935   \n",
       "  6             0.601840         0.221359                 0.206935   \n",
       "  7             0.953545         0.295917                 0.209567   \n",
       "  8             0.953740         0.295092                 0.209599   \n",
       "  9             0.407124         0.217048                 0.206084   \n",
       "  \n",
       "     target_share_dataset      lift  \n",
       "  0              0.212447  1.050414  \n",
       "  1              0.212447  1.041951  \n",
       "  2              0.212447  1.041951  \n",
       "  3              0.212447  1.041951  \n",
       "  4              0.212447  1.041951  \n",
       "  5              0.212447  1.041951  \n",
       "  6              0.212447  1.041951  \n",
       "  7              0.212447  1.392903  \n",
       "  8              0.212447  1.389019  \n",
       "  9              0.212447  1.021659  },\n",
       " {'250_mu_GA_0_SDFS':     quality                                           subgroup  size_sg  \\\n",
       "  0  0.895687  012: [0.02:0.02[ AND average_degree: [2.06:2.0...      1.0   \n",
       "  1  0.895687  021C: [0.00:0.00[ AND average_degree: [2.06:2....      1.0   \n",
       "  2  0.895687  030T: [0.00:0.00[ AND average_degree: [2.06:2....      1.0   \n",
       "  3  0.895687  021U: [0.00:0.00[ AND average_degree: [2.06:2....      1.0   \n",
       "  4  0.895687  021U: [0.00:0.00[ AND average_degree: [2.06:2....      1.0   \n",
       "  5  0.895687  average_degree: [2.06:2.07[ AND betweenness_ce...      1.0   \n",
       "  6  0.895687  030T: [0.00:0.00[ AND average_degree: [2.06:2....      1.0   \n",
       "  7  0.895687  average_degree: [2.06:2.07[ AND degree_central...      1.0   \n",
       "  8  0.895687  average_clustering: [0.01:0.01[ AND average_de...      1.0   \n",
       "  9  0.895687  average_degree: [2.06:2.07[ AND degree_central...      1.0   \n",
       "  \n",
       "     size_dataset  positives_sg  positives_dataset  size_complement  \\\n",
       "  0       25529.0           1.0             2663.0          25528.0   \n",
       "  1       25529.0           1.0             2663.0          25528.0   \n",
       "  2       25529.0           1.0             2663.0          25528.0   \n",
       "  3       25529.0           1.0             2663.0          25528.0   \n",
       "  4       25529.0           1.0             2663.0          25528.0   \n",
       "  5       25529.0           1.0             2663.0          25528.0   \n",
       "  6       25529.0           1.0             2663.0          25528.0   \n",
       "  7       25529.0           1.0             2663.0          25528.0   \n",
       "  8       25529.0           1.0             2663.0          25528.0   \n",
       "  9       25529.0           1.0             2663.0          25528.0   \n",
       "  \n",
       "     relative_size_sg  relative_size_complement  coverage_sg  \\\n",
       "  0          0.000039                  0.999961     0.000376   \n",
       "  1          0.000039                  0.999961     0.000376   \n",
       "  2          0.000039                  0.999961     0.000376   \n",
       "  3          0.000039                  0.999961     0.000376   \n",
       "  4          0.000039                  0.999961     0.000376   \n",
       "  5          0.000039                  0.999961     0.000376   \n",
       "  6          0.000039                  0.999961     0.000376   \n",
       "  7          0.000039                  0.999961     0.000376   \n",
       "  8          0.000039                  0.999961     0.000376   \n",
       "  9          0.000039                  0.999961     0.000376   \n",
       "  \n",
       "     coverage_complement  target_share_sg  target_share_complement  \\\n",
       "  0             0.999624              1.0                 0.104278   \n",
       "  1             0.999624              1.0                 0.104278   \n",
       "  2             0.999624              1.0                 0.104278   \n",
       "  3             0.999624              1.0                 0.104278   \n",
       "  4             0.999624              1.0                 0.104278   \n",
       "  5             0.999624              1.0                 0.104278   \n",
       "  6             0.999624              1.0                 0.104278   \n",
       "  7             0.999624              1.0                 0.104278   \n",
       "  8             0.999624              1.0                 0.104278   \n",
       "  9             0.999624              1.0                 0.104278   \n",
       "  \n",
       "     target_share_dataset      lift  \n",
       "  0              0.104313  9.586557  \n",
       "  1              0.104313  9.586557  \n",
       "  2              0.104313  9.586557  \n",
       "  3              0.104313  9.586557  \n",
       "  4              0.104313  9.586557  \n",
       "  5              0.104313  9.586557  \n",
       "  6              0.104313  9.586557  \n",
       "  7              0.104313  9.586557  \n",
       "  8              0.104313  9.586557  \n",
       "  9              0.104313  9.586557  ,\n",
       "  '250_mu_SQF_0_BFS':     quality                                           subgroup  size_sg  \\\n",
       "  0  0.895687  012: [0.01:0.02[ AND 021C: [0.00:0.00[ AND 021...      1.0   \n",
       "  1  0.895687  012: [0.01:0.02[ AND 021C: [0.00:0.00[ AND 021...      1.0   \n",
       "  2  0.895687  012: [0.01:0.02[ AND 021D: [0.00:0.00[ AND 030...      1.0   \n",
       "  3  0.895687  012: [0.01:0.02[ AND 021C: [0.00:0.00[ AND 030...      1.0   \n",
       "  4  0.895687  012: [0.01:0.02[ AND 021D: [0.00:0.00[ AND 030...      2.0   \n",
       "  5  0.895687  012: [0.01:0.02[ AND 021D: [0.00:0.00[ AND 030...      1.0   \n",
       "  6  0.895687  012: [0.01:0.02[ AND 021D: [0.00:0.00[ AND 030...      1.0   \n",
       "  7  0.895687  012: [0.01:0.02[ AND 021D: [0.00:0.00[ AND 030...      1.0   \n",
       "  8  0.895687  012: [0.01:0.02[ AND 021D: [0.00:0.00[ AND 030...      1.0   \n",
       "  9  0.895687  012: [0.01:0.02[ AND 021D: [0.00:0.00[ AND 030...      1.0   \n",
       "  \n",
       "     size_dataset  positives_sg  positives_dataset  size_complement  \\\n",
       "  0       25529.0           1.0             2663.0          25528.0   \n",
       "  1       25529.0           1.0             2663.0          25528.0   \n",
       "  2       25529.0           1.0             2663.0          25528.0   \n",
       "  3       25529.0           1.0             2663.0          25528.0   \n",
       "  4       25529.0           2.0             2663.0          25527.0   \n",
       "  5       25529.0           1.0             2663.0          25528.0   \n",
       "  6       25529.0           1.0             2663.0          25528.0   \n",
       "  7       25529.0           1.0             2663.0          25528.0   \n",
       "  8       25529.0           1.0             2663.0          25528.0   \n",
       "  9       25529.0           1.0             2663.0          25528.0   \n",
       "  \n",
       "     relative_size_sg  relative_size_complement  coverage_sg  \\\n",
       "  0          0.000039                  0.999961     0.000376   \n",
       "  1          0.000039                  0.999961     0.000376   \n",
       "  2          0.000039                  0.999961     0.000376   \n",
       "  3          0.000039                  0.999961     0.000376   \n",
       "  4          0.000078                  0.999922     0.000751   \n",
       "  5          0.000039                  0.999961     0.000376   \n",
       "  6          0.000039                  0.999961     0.000376   \n",
       "  7          0.000039                  0.999961     0.000376   \n",
       "  8          0.000039                  0.999961     0.000376   \n",
       "  9          0.000039                  0.999961     0.000376   \n",
       "  \n",
       "     coverage_complement  target_share_sg  target_share_complement  \\\n",
       "  0             0.999624              1.0                 0.104278   \n",
       "  1             0.999624              1.0                 0.104278   \n",
       "  2             0.999624              1.0                 0.104278   \n",
       "  3             0.999624              1.0                 0.104278   \n",
       "  4             0.999249              1.0                 0.104243   \n",
       "  5             0.999624              1.0                 0.104278   \n",
       "  6             0.999624              1.0                 0.104278   \n",
       "  7             0.999624              1.0                 0.104278   \n",
       "  8             0.999624              1.0                 0.104278   \n",
       "  9             0.999624              1.0                 0.104278   \n",
       "  \n",
       "     target_share_dataset      lift  \n",
       "  0              0.104313  9.586557  \n",
       "  1              0.104313  9.586557  \n",
       "  2              0.104313  9.586557  \n",
       "  3              0.104313  9.586557  \n",
       "  4              0.104313  9.586557  \n",
       "  5              0.104313  9.586557  \n",
       "  6              0.104313  9.586557  \n",
       "  7              0.104313  9.586557  \n",
       "  8              0.104313  9.586557  \n",
       "  9              0.104313  9.586557  },\n",
       " {'250_mu_GA_0.5_SDFS':     quality                                 subgroup  size_sg  size_dataset  \\\n",
       "  0  0.062326                      average_degree<2.01    850.0       25529.0   \n",
       "  1  0.055457          betweenness_centralization<0.00    850.0       25529.0   \n",
       "  2  0.049609              average_degree: [2.01:2.02[    851.0       25529.0   \n",
       "  3  0.045515                longest_path: [0.02:0.03[    847.0       25529.0   \n",
       "  4  0.044149                        021C: [0.00:0.00[    853.0       25529.0   \n",
       "  5  0.041456  betweenness_centralization: [0.00:0.00[    851.0       25529.0   \n",
       "  6  0.039090                        030T: [0.00:0.00[    894.0       25529.0   \n",
       "  7  0.038113                        021C: [0.00:0.00[    849.0       25529.0   \n",
       "  8  0.037979                        021C: [0.00:0.00[    852.0       25529.0   \n",
       "  9  0.037809                        021C: [0.00:0.00[    851.0       25529.0   \n",
       "  \n",
       "     positives_sg  positives_dataset  size_complement  relative_size_sg  \\\n",
       "  0         379.0             2663.0          24679.0          0.033295   \n",
       "  1         347.0             2663.0          24679.0          0.033295   \n",
       "  2         320.0             2663.0          24678.0          0.033335   \n",
       "  3         300.0             2663.0          24682.0          0.033178   \n",
       "  4         295.0             2663.0          24676.0          0.033413   \n",
       "  5         282.0             2663.0          24678.0          0.033335   \n",
       "  6         280.0             2663.0          24635.0          0.035019   \n",
       "  7         266.0             2663.0          24680.0          0.033256   \n",
       "  8         266.0             2663.0          24677.0          0.033374   \n",
       "  9         265.0             2663.0          24678.0          0.033335   \n",
       "  \n",
       "     relative_size_complement  coverage_sg  coverage_complement  \\\n",
       "  0                  0.966705     0.142321             0.857679   \n",
       "  1                  0.966705     0.130304             0.869696   \n",
       "  2                  0.966665     0.120165             0.879835   \n",
       "  3                  0.966822     0.112655             0.887345   \n",
       "  4                  0.966587     0.110777             0.889223   \n",
       "  5                  0.966665     0.105896             0.894104   \n",
       "  6                  0.964981     0.105145             0.894855   \n",
       "  7                  0.966744     0.099887             0.900113   \n",
       "  8                  0.966626     0.099887             0.900113   \n",
       "  9                  0.966665     0.099512             0.900488   \n",
       "  \n",
       "     target_share_sg  target_share_complement  target_share_dataset      lift  \n",
       "  0         0.445882                 0.092548              0.104313  4.274476  \n",
       "  1         0.408235                 0.093845              0.104313  3.913571  \n",
       "  2         0.376028                 0.094943              0.104313  3.604816  \n",
       "  3         0.354191                 0.095738              0.104313  3.395475  \n",
       "  4         0.345838                 0.095964              0.104313  3.315398  \n",
       "  5         0.331375                 0.096483              0.104313  3.176744  \n",
       "  6         0.313199                 0.096732              0.104313  3.002501  \n",
       "  7         0.313310                 0.097123              0.104313  3.003562  \n",
       "  8         0.312207                 0.097135              0.104313  2.992986  \n",
       "  9         0.311398                 0.097172              0.104313  2.985238  ,\n",
       "  '250_mu_SQF_0.5_BFS':     quality                                      subgroup  size_sg  \\\n",
       "  0  0.062326                           average_degree<2.01    850.0   \n",
       "  1  0.055457               betweenness_centralization<0.00    850.0   \n",
       "  2  0.049609                   average_degree: [2.01:2.02[    851.0   \n",
       "  3  0.045515                     longest_path: [0.02:0.03[    847.0   \n",
       "  4  0.044149                             021C: [0.00:0.00[    853.0   \n",
       "  5  0.042943  average_degree<2.01 AND density: [0.00:0.00[    116.0   \n",
       "  6  0.041456       betweenness_centralization: [0.00:0.00[    851.0   \n",
       "  7  0.040036     021D: [0.00:0.00[ AND average_degree<2.01    193.0   \n",
       "  8  0.039090                             030T: [0.00:0.00[    894.0   \n",
       "  9  0.038201        012: [0.02:0.02[ AND 021C: [0.00:0.00[    178.0   \n",
       "  \n",
       "     size_dataset  positives_sg  positives_dataset  size_complement  \\\n",
       "  0       25529.0         379.0             2663.0          24679.0   \n",
       "  1       25529.0         347.0             2663.0          24679.0   \n",
       "  2       25529.0         320.0             2663.0          24678.0   \n",
       "  3       25529.0         300.0             2663.0          24682.0   \n",
       "  4       25529.0         295.0             2663.0          24676.0   \n",
       "  5       25529.0          86.0             2663.0          25413.0   \n",
       "  6       25529.0         282.0             2663.0          24678.0   \n",
       "  7       25529.0         109.0             2663.0          25336.0   \n",
       "  8       25529.0         280.0             2663.0          24635.0   \n",
       "  9       25529.0         100.0             2663.0          25351.0   \n",
       "  \n",
       "     relative_size_sg  relative_size_complement  coverage_sg  \\\n",
       "  0          0.033295                  0.966705     0.142321   \n",
       "  1          0.033295                  0.966705     0.130304   \n",
       "  2          0.033335                  0.966665     0.120165   \n",
       "  3          0.033178                  0.966822     0.112655   \n",
       "  4          0.033413                  0.966587     0.110777   \n",
       "  5          0.004544                  0.995456     0.032294   \n",
       "  6          0.033335                  0.966665     0.105896   \n",
       "  7          0.007560                  0.992440     0.040931   \n",
       "  8          0.035019                  0.964981     0.105145   \n",
       "  9          0.006972                  0.993028     0.037552   \n",
       "  \n",
       "     coverage_complement  target_share_sg  target_share_complement  \\\n",
       "  0             0.857679         0.445882                 0.092548   \n",
       "  1             0.869696         0.408235                 0.093845   \n",
       "  2             0.879835         0.376028                 0.094943   \n",
       "  3             0.887345         0.354191                 0.095738   \n",
       "  4             0.889223         0.345838                 0.095964   \n",
       "  5             0.967706         0.741379                 0.101405   \n",
       "  6             0.894104         0.331375                 0.096483   \n",
       "  7             0.959069         0.564767                 0.100805   \n",
       "  8             0.894855         0.313199                 0.096732   \n",
       "  9             0.962448         0.561798                 0.101101   \n",
       "  \n",
       "     target_share_dataset      lift  \n",
       "  0              0.104313  4.274476  \n",
       "  1              0.104313  3.913571  \n",
       "  2              0.104313  3.604816  \n",
       "  3              0.104313  3.395475  \n",
       "  4              0.104313  3.315398  \n",
       "  5              0.104313  7.107275  \n",
       "  6              0.104313  3.176744  \n",
       "  7              0.104313  5.414169  \n",
       "  8              0.104313  3.002501  \n",
       "  9              0.104313  5.385706  },\n",
       " {'250_mu_GA_1_SDFS':     quality                                 subgroup  size_sg  size_dataset  \\\n",
       "  0  0.011373                      average_degree<2.01    850.0       25529.0   \n",
       "  1  0.010119          betweenness_centralization<0.00    850.0       25529.0   \n",
       "  2  0.009058              average_degree: [2.01:2.02[    851.0       25529.0   \n",
       "  3  0.008290                longest_path: [0.02:0.03[    847.0       25529.0   \n",
       "  4  0.008070                        021C: [0.00:0.00[    853.0       25529.0   \n",
       "  5  0.007623                         030T: [0.0:0.00[   1643.0       25529.0   \n",
       "  6  0.007569  betweenness_centralization: [0.00:0.00[    851.0       25529.0   \n",
       "  7  0.007315                        030T: [0.00:0.00[    894.0       25529.0   \n",
       "  8  0.006950                        021C: [0.00:0.00[    849.0       25529.0   \n",
       "  9  0.006938                        021C: [0.00:0.00[    852.0       25529.0   \n",
       "  \n",
       "     positives_sg  positives_dataset  size_complement  relative_size_sg  \\\n",
       "  0         379.0             2663.0          24679.0          0.033295   \n",
       "  1         347.0             2663.0          24679.0          0.033295   \n",
       "  2         320.0             2663.0          24678.0          0.033335   \n",
       "  3         300.0             2663.0          24682.0          0.033178   \n",
       "  4         295.0             2663.0          24676.0          0.033413   \n",
       "  5         366.0             2663.0          23886.0          0.064358   \n",
       "  6         282.0             2663.0          24678.0          0.033335   \n",
       "  7         280.0             2663.0          24635.0          0.035019   \n",
       "  8         266.0             2663.0          24680.0          0.033256   \n",
       "  9         266.0             2663.0          24677.0          0.033374   \n",
       "  \n",
       "     relative_size_complement  coverage_sg  coverage_complement  \\\n",
       "  0                  0.966705     0.142321             0.857679   \n",
       "  1                  0.966705     0.130304             0.869696   \n",
       "  2                  0.966665     0.120165             0.879835   \n",
       "  3                  0.966822     0.112655             0.887345   \n",
       "  4                  0.966587     0.110777             0.889223   \n",
       "  5                  0.935642     0.137439             0.862561   \n",
       "  6                  0.966665     0.105896             0.894104   \n",
       "  7                  0.964981     0.105145             0.894855   \n",
       "  8                  0.966744     0.099887             0.900113   \n",
       "  9                  0.966626     0.099887             0.900113   \n",
       "  \n",
       "     target_share_sg  target_share_complement  target_share_dataset      lift  \n",
       "  0         0.445882                 0.092548              0.104313  4.274476  \n",
       "  1         0.408235                 0.093845              0.104313  3.913571  \n",
       "  2         0.376028                 0.094943              0.104313  3.604816  \n",
       "  3         0.354191                 0.095738              0.104313  3.395475  \n",
       "  4         0.345838                 0.095964              0.104313  3.315398  \n",
       "  5         0.222763                 0.096165              0.104313  2.135532  \n",
       "  6         0.331375                 0.096483              0.104313  3.176744  \n",
       "  7         0.313199                 0.096732              0.104313  3.002501  \n",
       "  8         0.313310                 0.097123              0.104313  3.003562  \n",
       "  9         0.312207                 0.097135              0.104313  2.992986  ,\n",
       "  '250_mu_SQF_1_BFS':     quality                                 subgroup  size_sg  size_dataset  \\\n",
       "  0  0.011373                      average_degree<2.01    850.0       25529.0   \n",
       "  1  0.010119          betweenness_centralization<0.00    850.0       25529.0   \n",
       "  2  0.009058              average_degree: [2.01:2.02[    851.0       25529.0   \n",
       "  3  0.008290                longest_path: [0.02:0.03[    847.0       25529.0   \n",
       "  4  0.008070                        021C: [0.00:0.00[    853.0       25529.0   \n",
       "  5  0.007623                         030T: [0.0:0.00[   1643.0       25529.0   \n",
       "  6  0.007569  betweenness_centralization: [0.00:0.00[    851.0       25529.0   \n",
       "  7  0.007315                        030T: [0.00:0.00[    894.0       25529.0   \n",
       "  8  0.006950                        021C: [0.00:0.00[    849.0       25529.0   \n",
       "  9  0.006938                        021C: [0.00:0.00[    852.0       25529.0   \n",
       "  \n",
       "     positives_sg  positives_dataset  size_complement  relative_size_sg  \\\n",
       "  0         379.0             2663.0          24679.0          0.033295   \n",
       "  1         347.0             2663.0          24679.0          0.033295   \n",
       "  2         320.0             2663.0          24678.0          0.033335   \n",
       "  3         300.0             2663.0          24682.0          0.033178   \n",
       "  4         295.0             2663.0          24676.0          0.033413   \n",
       "  5         366.0             2663.0          23886.0          0.064358   \n",
       "  6         282.0             2663.0          24678.0          0.033335   \n",
       "  7         280.0             2663.0          24635.0          0.035019   \n",
       "  8         266.0             2663.0          24680.0          0.033256   \n",
       "  9         266.0             2663.0          24677.0          0.033374   \n",
       "  \n",
       "     relative_size_complement  coverage_sg  coverage_complement  \\\n",
       "  0                  0.966705     0.142321             0.857679   \n",
       "  1                  0.966705     0.130304             0.869696   \n",
       "  2                  0.966665     0.120165             0.879835   \n",
       "  3                  0.966822     0.112655             0.887345   \n",
       "  4                  0.966587     0.110777             0.889223   \n",
       "  5                  0.935642     0.137439             0.862561   \n",
       "  6                  0.966665     0.105896             0.894104   \n",
       "  7                  0.964981     0.105145             0.894855   \n",
       "  8                  0.966744     0.099887             0.900113   \n",
       "  9                  0.966626     0.099887             0.900113   \n",
       "  \n",
       "     target_share_sg  target_share_complement  target_share_dataset      lift  \n",
       "  0         0.445882                 0.092548              0.104313  4.274476  \n",
       "  1         0.408235                 0.093845              0.104313  3.913571  \n",
       "  2         0.376028                 0.094943              0.104313  3.604816  \n",
       "  3         0.354191                 0.095738              0.104313  3.395475  \n",
       "  4         0.345838                 0.095964              0.104313  3.315398  \n",
       "  5         0.222763                 0.096165              0.104313  2.135532  \n",
       "  6         0.331375                 0.096483              0.104313  3.176744  \n",
       "  7         0.313199                 0.096732              0.104313  3.002501  \n",
       "  8         0.313310                 0.097123              0.104313  3.003562  \n",
       "  9         0.312207                 0.097135              0.104313  2.992986  },\n",
       " {'250_pol_GA_0_SDFS':    quality                                           subgroup  size_sg  \\\n",
       "  0  0.40123  021C: [0.00:0.00[ AND average_degree<2.01 AND ...      1.0   \n",
       "  1  0.40123  021D: [0.00:0.00[ AND average_degree<2.01 AND ...      1.0   \n",
       "  2  0.40123  021D: [0.00:0.00[ AND average_degree<2.01 AND ...      1.0   \n",
       "  3  0.40123  021U: [0.00:0.00[ AND average_degree<2.01 AND ...      1.0   \n",
       "  4  0.40123  021U: [0.00:0.00[ AND average_degree<2.01 AND ...      1.0   \n",
       "  5  0.40123  average_degree<2.01 AND density: [0.00:0.00[ A...      1.0   \n",
       "  6  0.40123  030T: [0.00:0.00[ AND average_degree<2.01 AND ...      2.0   \n",
       "  7  0.40123  average_degree<2.01 AND density<0.00 AND longe...      1.0   \n",
       "  8  0.40123  average_clustering: [0.00:0.00[ AND average_de...      2.0   \n",
       "  9  0.40123  021U: [0.00:0.00[ AND average_degree<2.01 AND ...      1.0   \n",
       "  \n",
       "     size_dataset  positives_sg  positives_dataset  size_complement  \\\n",
       "  0       25529.0           1.0            15286.0          25528.0   \n",
       "  1       25529.0           1.0            15286.0          25528.0   \n",
       "  2       25529.0           1.0            15286.0          25528.0   \n",
       "  3       25529.0           1.0            15286.0          25528.0   \n",
       "  4       25529.0           1.0            15286.0          25528.0   \n",
       "  5       25529.0           1.0            15286.0          25528.0   \n",
       "  6       25529.0           2.0            15286.0          25527.0   \n",
       "  7       25529.0           1.0            15286.0          25528.0   \n",
       "  8       25529.0           2.0            15286.0          25527.0   \n",
       "  9       25529.0           1.0            15286.0          25528.0   \n",
       "  \n",
       "     relative_size_sg  relative_size_complement  coverage_sg  \\\n",
       "  0          0.000039                  0.999961     0.000065   \n",
       "  1          0.000039                  0.999961     0.000065   \n",
       "  2          0.000039                  0.999961     0.000065   \n",
       "  3          0.000039                  0.999961     0.000065   \n",
       "  4          0.000039                  0.999961     0.000065   \n",
       "  5          0.000039                  0.999961     0.000065   \n",
       "  6          0.000078                  0.999922     0.000131   \n",
       "  7          0.000039                  0.999961     0.000065   \n",
       "  8          0.000078                  0.999922     0.000131   \n",
       "  9          0.000039                  0.999961     0.000065   \n",
       "  \n",
       "     coverage_complement  target_share_sg  target_share_complement  \\\n",
       "  0             0.999935              1.0                 0.598754   \n",
       "  1             0.999935              1.0                 0.598754   \n",
       "  2             0.999935              1.0                 0.598754   \n",
       "  3             0.999935              1.0                 0.598754   \n",
       "  4             0.999935              1.0                 0.598754   \n",
       "  5             0.999935              1.0                 0.598754   \n",
       "  6             0.999869              1.0                 0.598739   \n",
       "  7             0.999935              1.0                 0.598754   \n",
       "  8             0.999869              1.0                 0.598739   \n",
       "  9             0.999935              1.0                 0.598754   \n",
       "  \n",
       "     target_share_dataset     lift  \n",
       "  0               0.59877  1.67009  \n",
       "  1               0.59877  1.67009  \n",
       "  2               0.59877  1.67009  \n",
       "  3               0.59877  1.67009  \n",
       "  4               0.59877  1.67009  \n",
       "  5               0.59877  1.67009  \n",
       "  6               0.59877  1.67009  \n",
       "  7               0.59877  1.67009  \n",
       "  8               0.59877  1.67009  \n",
       "  9               0.59877  1.67009  ,\n",
       "  '250_pol_SQF_0_BFS':    quality                                subgroup  size_sg  size_dataset  \\\n",
       "  0  0.40123  012: [0.01:0.02[ AND 021C: [0.00:0.00[     15.0       25529.0   \n",
       "  1  0.40123  012: [0.01:0.02[ AND 021C: [0.00:0.00[      4.0       25529.0   \n",
       "  2  0.40123  012: [0.01:0.02[ AND 021D: [0.00:0.00[      7.0       25529.0   \n",
       "  3  0.40123         012: [0.01:0.02[ AND 021D>=0.00      9.0       25529.0   \n",
       "  4  0.40123  012: [0.01:0.02[ AND 021C: [0.00:0.00[      2.0       25529.0   \n",
       "  5  0.40123  012: [0.01:0.02[ AND 021U: [0.00:0.00[      1.0       25529.0   \n",
       "  6  0.40123  012: [0.01:0.02[ AND 021U: [0.00:0.00[      1.0       25529.0   \n",
       "  7  0.40123  012: [0.01:0.02[ AND 021U: [0.00:0.00[      1.0       25529.0   \n",
       "  8  0.40123          012: [0.01:0.02[ AND 021U<0.00      1.0       25529.0   \n",
       "  9  0.40123  012: [0.01:0.02[ AND 021C: [0.00:0.00[      3.0       25529.0   \n",
       "  \n",
       "     positives_sg  positives_dataset  size_complement  relative_size_sg  \\\n",
       "  0          15.0            15286.0          25514.0          0.000588   \n",
       "  1           4.0            15286.0          25525.0          0.000157   \n",
       "  2           7.0            15286.0          25522.0          0.000274   \n",
       "  3           9.0            15286.0          25520.0          0.000353   \n",
       "  4           2.0            15286.0          25527.0          0.000078   \n",
       "  5           1.0            15286.0          25528.0          0.000039   \n",
       "  6           1.0            15286.0          25528.0          0.000039   \n",
       "  7           1.0            15286.0          25528.0          0.000039   \n",
       "  8           1.0            15286.0          25528.0          0.000039   \n",
       "  9           3.0            15286.0          25526.0          0.000118   \n",
       "  \n",
       "     relative_size_complement  coverage_sg  coverage_complement  \\\n",
       "  0                  0.999412     0.000981             0.999019   \n",
       "  1                  0.999843     0.000262             0.999738   \n",
       "  2                  0.999726     0.000458             0.999542   \n",
       "  3                  0.999647     0.000589             0.999411   \n",
       "  4                  0.999922     0.000131             0.999869   \n",
       "  5                  0.999961     0.000065             0.999935   \n",
       "  6                  0.999961     0.000065             0.999935   \n",
       "  7                  0.999961     0.000065             0.999935   \n",
       "  8                  0.999961     0.000065             0.999935   \n",
       "  9                  0.999882     0.000196             0.999804   \n",
       "  \n",
       "     target_share_sg  target_share_complement  target_share_dataset     lift  \n",
       "  0              1.0                 0.598534               0.59877  1.67009  \n",
       "  1              1.0                 0.598707               0.59877  1.67009  \n",
       "  2              1.0                 0.598660               0.59877  1.67009  \n",
       "  3              1.0                 0.598629               0.59877  1.67009  \n",
       "  4              1.0                 0.598739               0.59877  1.67009  \n",
       "  5              1.0                 0.598754               0.59877  1.67009  \n",
       "  6              1.0                 0.598754               0.59877  1.67009  \n",
       "  7              1.0                 0.598754               0.59877  1.67009  \n",
       "  8              1.0                 0.598754               0.59877  1.67009  \n",
       "  9              1.0                 0.598723               0.59877  1.67009  },\n",
       " {'250_pol_GA_0.5_SDFS':     quality                            subgroup  size_sg  size_dataset  \\\n",
       "  0  0.036998         degree_centralization>=0.53    851.0       25529.0   \n",
       "  1  0.035496                   021U: [0.00:0.00[    851.0       25529.0   \n",
       "  2  0.035281                          021U>=0.00    851.0       25529.0   \n",
       "  3  0.032921                   021D: [0.00:0.00[    851.0       25529.0   \n",
       "  4  0.032854                            012<0.01    850.0       25529.0   \n",
       "  5  0.029488                   021D: [0.00:0.00[    851.0       25529.0   \n",
       "  6  0.029274  degree_centralization: [0.47:0.53[    851.0       25529.0   \n",
       "  7  0.025841    betweenness_centralization>=0.01    851.0       25529.0   \n",
       "  8  0.025412                average_degree>=2.36    851.0       25529.0   \n",
       "  9  0.024769                   021D: [0.00:0.00[    851.0       25529.0   \n",
       "  \n",
       "     positives_sg  positives_dataset  size_complement  relative_size_sg  \\\n",
       "  0         682.0            15286.0          24678.0          0.033335   \n",
       "  1         675.0            15286.0          24678.0          0.033335   \n",
       "  2         674.0            15286.0          24678.0          0.033335   \n",
       "  3         663.0            15286.0          24678.0          0.033335   \n",
       "  4         662.0            15286.0          24679.0          0.033295   \n",
       "  5         647.0            15286.0          24678.0          0.033335   \n",
       "  6         646.0            15286.0          24678.0          0.033335   \n",
       "  7         630.0            15286.0          24678.0          0.033335   \n",
       "  8         628.0            15286.0          24678.0          0.033335   \n",
       "  9         625.0            15286.0          24678.0          0.033335   \n",
       "  \n",
       "     relative_size_complement  coverage_sg  coverage_complement  \\\n",
       "  0                  0.966665     0.044616             0.955384   \n",
       "  1                  0.966665     0.044158             0.955842   \n",
       "  2                  0.966665     0.044093             0.955907   \n",
       "  3                  0.966665     0.043373             0.956627   \n",
       "  4                  0.966705     0.043308             0.956692   \n",
       "  5                  0.966665     0.042326             0.957674   \n",
       "  6                  0.966665     0.042261             0.957739   \n",
       "  7                  0.966665     0.041214             0.958786   \n",
       "  8                  0.966665     0.041083             0.958917   \n",
       "  9                  0.966665     0.040887             0.959113   \n",
       "  \n",
       "     target_share_sg  target_share_complement  target_share_dataset      lift  \n",
       "  0         0.801410                 0.591782               0.59877  1.338427  \n",
       "  1         0.793184                 0.592066               0.59877  1.324690  \n",
       "  2         0.792009                 0.592106               0.59877  1.322727  \n",
       "  3         0.779083                 0.592552               0.59877  1.301140  \n",
       "  4         0.778824                 0.592569               0.59877  1.300706  \n",
       "  5         0.760282                 0.593200               0.59877  1.269740  \n",
       "  6         0.759107                 0.593241               0.59877  1.267777  \n",
       "  7         0.740306                 0.593889               0.59877  1.236377  \n",
       "  8         0.737955                 0.593970               0.59877  1.232452  \n",
       "  9         0.734430                 0.594092               0.59877  1.226565  ,\n",
       "  '250_pol_SQF_0.5_BFS':     quality                                           subgroup  size_sg  \\\n",
       "  0  0.036998                        degree_centralization>=0.53    851.0   \n",
       "  1  0.035739         021U>=0.00 AND degree_centralization>=0.53    765.0   \n",
       "  2  0.035496                                  021U: [0.00:0.00[    851.0   \n",
       "  3  0.035281                                         021U>=0.00    851.0   \n",
       "  4  0.033841           012<0.01 AND degree_centralization>=0.53    640.0   \n",
       "  5  0.033190                            012<0.01 AND 021U>=0.00    580.0   \n",
       "  6  0.032921                                  021D: [0.00:0.00[    851.0   \n",
       "  7  0.032854                                           012<0.01    850.0   \n",
       "  8  0.032810  012<0.01 AND 021U>=0.00 AND degree_centralizat...    575.0   \n",
       "  9  0.032367          021C<0.00 AND degree_centralization>=0.53    521.0   \n",
       "  \n",
       "     size_dataset  positives_sg  positives_dataset  size_complement  \\\n",
       "  0       25529.0         682.0            15286.0          24678.0   \n",
       "  1       25529.0         616.0            15286.0          24764.0   \n",
       "  2       25529.0         675.0            15286.0          24678.0   \n",
       "  3       25529.0         674.0            15286.0          24678.0   \n",
       "  4       25529.0         520.0            15286.0          24889.0   \n",
       "  5       25529.0         475.0            15286.0          24949.0   \n",
       "  6       25529.0         663.0            15286.0          24678.0   \n",
       "  7       25529.0         662.0            15286.0          24679.0   \n",
       "  8       25529.0         470.0            15286.0          24954.0   \n",
       "  9       25529.0         430.0            15286.0          25008.0   \n",
       "  \n",
       "     relative_size_sg  relative_size_complement  coverage_sg  \\\n",
       "  0          0.033335                  0.966665     0.044616   \n",
       "  1          0.029966                  0.970034     0.040298   \n",
       "  2          0.033335                  0.966665     0.044158   \n",
       "  3          0.033335                  0.966665     0.044093   \n",
       "  4          0.025070                  0.974930     0.034018   \n",
       "  5          0.022719                  0.977281     0.031074   \n",
       "  6          0.033335                  0.966665     0.043373   \n",
       "  7          0.033295                  0.966705     0.043308   \n",
       "  8          0.022523                  0.977477     0.030747   \n",
       "  9          0.020408                  0.979592     0.028130   \n",
       "  \n",
       "     coverage_complement  target_share_sg  target_share_complement  \\\n",
       "  0             0.955384         0.801410                 0.591782   \n",
       "  1             0.959702         0.805229                 0.592392   \n",
       "  2             0.955842         0.793184                 0.592066   \n",
       "  3             0.955907         0.792009                 0.592106   \n",
       "  4             0.965982         0.812500                 0.593274   \n",
       "  5             0.968926         0.818966                 0.593651   \n",
       "  6             0.956627         0.779083                 0.592552   \n",
       "  7             0.956692         0.778824                 0.592569   \n",
       "  8             0.969253         0.817391                 0.593732   \n",
       "  9             0.971870         0.825336                 0.594050   \n",
       "  \n",
       "     target_share_dataset      lift  \n",
       "  0               0.59877  1.338427  \n",
       "  1               0.59877  1.344805  \n",
       "  2               0.59877  1.324690  \n",
       "  3               0.59877  1.322727  \n",
       "  4               0.59877  1.356948  \n",
       "  5               0.59877  1.367746  \n",
       "  6               0.59877  1.301140  \n",
       "  7               0.59877  1.300706  \n",
       "  8               0.59877  1.365117  \n",
       "  9               0.59877  1.378385  },\n",
       " {'250_pol_GA_1_SDFS':     quality                            subgroup  size_sg  size_dataset  \\\n",
       "  0  0.006755         degree_centralization>=0.53    851.0       25529.0   \n",
       "  1  0.006481                   021U: [0.00:0.00[    851.0       25529.0   \n",
       "  2  0.006442                          021U>=0.00    851.0       25529.0   \n",
       "  3  0.006011                   021D: [0.00:0.00[    851.0       25529.0   \n",
       "  4  0.005995                            012<0.01    850.0       25529.0   \n",
       "  5  0.005384                   021D: [0.00:0.00[    851.0       25529.0   \n",
       "  6  0.005345  degree_centralization: [0.47:0.53[    851.0       25529.0   \n",
       "  7  0.004718    betweenness_centralization>=0.01    851.0       25529.0   \n",
       "  8  0.004640                average_degree>=2.36    851.0       25529.0   \n",
       "  9  0.004522                   021D: [0.00:0.00[    851.0       25529.0   \n",
       "  \n",
       "     positives_sg  positives_dataset  size_complement  relative_size_sg  \\\n",
       "  0         682.0            15286.0          24678.0          0.033335   \n",
       "  1         675.0            15286.0          24678.0          0.033335   \n",
       "  2         674.0            15286.0          24678.0          0.033335   \n",
       "  3         663.0            15286.0          24678.0          0.033335   \n",
       "  4         662.0            15286.0          24679.0          0.033295   \n",
       "  5         647.0            15286.0          24678.0          0.033335   \n",
       "  6         646.0            15286.0          24678.0          0.033335   \n",
       "  7         630.0            15286.0          24678.0          0.033335   \n",
       "  8         628.0            15286.0          24678.0          0.033335   \n",
       "  9         625.0            15286.0          24678.0          0.033335   \n",
       "  \n",
       "     relative_size_complement  coverage_sg  coverage_complement  \\\n",
       "  0                  0.966665     0.044616             0.955384   \n",
       "  1                  0.966665     0.044158             0.955842   \n",
       "  2                  0.966665     0.044093             0.955907   \n",
       "  3                  0.966665     0.043373             0.956627   \n",
       "  4                  0.966705     0.043308             0.956692   \n",
       "  5                  0.966665     0.042326             0.957674   \n",
       "  6                  0.966665     0.042261             0.957739   \n",
       "  7                  0.966665     0.041214             0.958786   \n",
       "  8                  0.966665     0.041083             0.958917   \n",
       "  9                  0.966665     0.040887             0.959113   \n",
       "  \n",
       "     target_share_sg  target_share_complement  target_share_dataset      lift  \n",
       "  0         0.801410                 0.591782               0.59877  1.338427  \n",
       "  1         0.793184                 0.592066               0.59877  1.324690  \n",
       "  2         0.792009                 0.592106               0.59877  1.322727  \n",
       "  3         0.779083                 0.592552               0.59877  1.301140  \n",
       "  4         0.778824                 0.592569               0.59877  1.300706  \n",
       "  5         0.760282                 0.593200               0.59877  1.269740  \n",
       "  6         0.759107                 0.593241               0.59877  1.267777  \n",
       "  7         0.740306                 0.593889               0.59877  1.236377  \n",
       "  8         0.737955                 0.593970               0.59877  1.232452  \n",
       "  9         0.734430                 0.594092               0.59877  1.226565  ,\n",
       "  '250_pol_SQF_1_BFS':     quality                                    subgroup  size_sg  \\\n",
       "  0  0.006755                 degree_centralization>=0.53    851.0   \n",
       "  1  0.006481                           021U: [0.00:0.00[    851.0   \n",
       "  2  0.006442                                  021U>=0.00    851.0   \n",
       "  3  0.006187  021U>=0.00 AND degree_centralization>=0.53    765.0   \n",
       "  4  0.006011                           021D: [0.00:0.00[    851.0   \n",
       "  5  0.005995                                    012<0.01    850.0   \n",
       "  6  0.005384                           021D: [0.00:0.00[    851.0   \n",
       "  7  0.005358    012<0.01 AND degree_centralization>=0.53    640.0   \n",
       "  8  0.005345          degree_centralization: [0.47:0.53[    851.0   \n",
       "  9  0.005003                     012<0.01 AND 021U>=0.00    580.0   \n",
       "  \n",
       "     size_dataset  positives_sg  positives_dataset  size_complement  \\\n",
       "  0       25529.0         682.0            15286.0          24678.0   \n",
       "  1       25529.0         675.0            15286.0          24678.0   \n",
       "  2       25529.0         674.0            15286.0          24678.0   \n",
       "  3       25529.0         616.0            15286.0          24764.0   \n",
       "  4       25529.0         663.0            15286.0          24678.0   \n",
       "  5       25529.0         662.0            15286.0          24679.0   \n",
       "  6       25529.0         647.0            15286.0          24678.0   \n",
       "  7       25529.0         520.0            15286.0          24889.0   \n",
       "  8       25529.0         646.0            15286.0          24678.0   \n",
       "  9       25529.0         475.0            15286.0          24949.0   \n",
       "  \n",
       "     relative_size_sg  relative_size_complement  coverage_sg  \\\n",
       "  0          0.033335                  0.966665     0.044616   \n",
       "  1          0.033335                  0.966665     0.044158   \n",
       "  2          0.033335                  0.966665     0.044093   \n",
       "  3          0.029966                  0.970034     0.040298   \n",
       "  4          0.033335                  0.966665     0.043373   \n",
       "  5          0.033295                  0.966705     0.043308   \n",
       "  6          0.033335                  0.966665     0.042326   \n",
       "  7          0.025070                  0.974930     0.034018   \n",
       "  8          0.033335                  0.966665     0.042261   \n",
       "  9          0.022719                  0.977281     0.031074   \n",
       "  \n",
       "     coverage_complement  target_share_sg  target_share_complement  \\\n",
       "  0             0.955384         0.801410                 0.591782   \n",
       "  1             0.955842         0.793184                 0.592066   \n",
       "  2             0.955907         0.792009                 0.592106   \n",
       "  3             0.959702         0.805229                 0.592392   \n",
       "  4             0.956627         0.779083                 0.592552   \n",
       "  5             0.956692         0.778824                 0.592569   \n",
       "  6             0.957674         0.760282                 0.593200   \n",
       "  7             0.965982         0.812500                 0.593274   \n",
       "  8             0.957739         0.759107                 0.593241   \n",
       "  9             0.968926         0.818966                 0.593651   \n",
       "  \n",
       "     target_share_dataset      lift  \n",
       "  0               0.59877  1.338427  \n",
       "  1               0.59877  1.324690  \n",
       "  2               0.59877  1.322727  \n",
       "  3               0.59877  1.344805  \n",
       "  4               0.59877  1.301140  \n",
       "  5               0.59877  1.300706  \n",
       "  6               0.59877  1.269740  \n",
       "  7               0.59877  1.356948  \n",
       "  8               0.59877  1.267777  \n",
       "  9               0.59877  1.367746  },\n",
       " {'250_v_GA_0_SDFS':    quality                                           subgroup  size_sg  \\\n",
       "  0   0.9206  012: [0.02:0.02[ AND average_degree: [2.01:2.0...      1.0   \n",
       "  1   0.9206  021C: [0.00:0.00[ AND 030T: [0.00:0.00[ AND av...      1.0   \n",
       "  2   0.9206  021U: [0.00:0.00[ AND average_clustering: [0.0...      1.0   \n",
       "  3   0.9206  021C: [0.00:0.00[ AND average_clustering: [0.0...      1.0   \n",
       "  4   0.9206  021C: [0.00:0.00[ AND average_degree: [2.01:2....      1.0   \n",
       "  5   0.9206  average_degree<2.01 AND degree_centralization<...      1.0   \n",
       "  6   0.9206  average_clustering: [0.01:0.01[ AND average_de...      1.0   \n",
       "  7   0.9206  021C: [0.00:0.00[ AND average_degree: [2.01:2....      1.0   \n",
       "  8   0.9206  021C: [0.00:0.00[ AND average_degree<2.01 AND ...      1.0   \n",
       "  9   0.9206  average_clustering: [0.01:0.01[ AND average_de...      1.0   \n",
       "  \n",
       "     size_dataset  positives_sg  positives_dataset  size_complement  \\\n",
       "  0       25529.0           1.0             2027.0          25528.0   \n",
       "  1       25529.0           1.0             2027.0          25528.0   \n",
       "  2       25529.0           1.0             2027.0          25528.0   \n",
       "  3       25529.0           1.0             2027.0          25528.0   \n",
       "  4       25529.0           1.0             2027.0          25528.0   \n",
       "  5       25529.0           1.0             2027.0          25528.0   \n",
       "  6       25529.0           1.0             2027.0          25528.0   \n",
       "  7       25529.0           1.0             2027.0          25528.0   \n",
       "  8       25529.0           1.0             2027.0          25528.0   \n",
       "  9       25529.0           1.0             2027.0          25528.0   \n",
       "  \n",
       "     relative_size_sg  relative_size_complement  coverage_sg  \\\n",
       "  0          0.000039                  0.999961     0.000493   \n",
       "  1          0.000039                  0.999961     0.000493   \n",
       "  2          0.000039                  0.999961     0.000493   \n",
       "  3          0.000039                  0.999961     0.000493   \n",
       "  4          0.000039                  0.999961     0.000493   \n",
       "  5          0.000039                  0.999961     0.000493   \n",
       "  6          0.000039                  0.999961     0.000493   \n",
       "  7          0.000039                  0.999961     0.000493   \n",
       "  8          0.000039                  0.999961     0.000493   \n",
       "  9          0.000039                  0.999961     0.000493   \n",
       "  \n",
       "     coverage_complement  target_share_sg  target_share_complement  \\\n",
       "  0             0.999507              1.0                 0.079364   \n",
       "  1             0.999507              1.0                 0.079364   \n",
       "  2             0.999507              1.0                 0.079364   \n",
       "  3             0.999507              1.0                 0.079364   \n",
       "  4             0.999507              1.0                 0.079364   \n",
       "  5             0.999507              1.0                 0.079364   \n",
       "  6             0.999507              1.0                 0.079364   \n",
       "  7             0.999507              1.0                 0.079364   \n",
       "  8             0.999507              1.0                 0.079364   \n",
       "  9             0.999507              1.0                 0.079364   \n",
       "  \n",
       "     target_share_dataset       lift  \n",
       "  0                0.0794  12.594475  \n",
       "  1                0.0794  12.594475  \n",
       "  2                0.0794  12.594475  \n",
       "  3                0.0794  12.594475  \n",
       "  4                0.0794  12.594475  \n",
       "  5                0.0794  12.594475  \n",
       "  6                0.0794  12.594475  \n",
       "  7                0.0794  12.594475  \n",
       "  8                0.0794  12.594475  \n",
       "  9                0.0794  12.594475  ,\n",
       "  '250_v_SQF_0_BFS':    quality                                           subgroup  size_sg  \\\n",
       "  0   0.9206  012: [0.01:0.02[ AND 021C: [0.00:0.00[ AND 030...      1.0   \n",
       "  1   0.9206  012: [0.01:0.02[ AND 021C: [0.00:0.00[ AND 030...      1.0   \n",
       "  2   0.9206  012: [0.01:0.02[ AND 021C: [0.00:0.00[ AND 030...      1.0   \n",
       "  3   0.9206  012: [0.01:0.02[ AND 021C: [0.00:0.00[ AND 030...      1.0   \n",
       "  4   0.9206  012: [0.01:0.02[ AND 021D: [0.00:0.00[ AND 030...      1.0   \n",
       "  5   0.9206  012: [0.01:0.02[ AND 021C: [0.00:0.00[ AND 030...      1.0   \n",
       "  6   0.9206  012: [0.01:0.02[ AND 021C: [0.00:0.00[ AND 030...      1.0   \n",
       "  7   0.9206  012: [0.01:0.02[ AND 021D: [0.00:0.00[ AND 021...      1.0   \n",
       "  8   0.9206  012: [0.01:0.02[ AND 021C<0.00 AND 030T: [0.00...      1.0   \n",
       "  9   0.9206  012: [0.01:0.02[ AND 021D: [0.00:0.00[ AND 030...      1.0   \n",
       "  \n",
       "     size_dataset  positives_sg  positives_dataset  size_complement  \\\n",
       "  0       25529.0           1.0             2027.0          25528.0   \n",
       "  1       25529.0           1.0             2027.0          25528.0   \n",
       "  2       25529.0           1.0             2027.0          25528.0   \n",
       "  3       25529.0           1.0             2027.0          25528.0   \n",
       "  4       25529.0           1.0             2027.0          25528.0   \n",
       "  5       25529.0           1.0             2027.0          25528.0   \n",
       "  6       25529.0           1.0             2027.0          25528.0   \n",
       "  7       25529.0           1.0             2027.0          25528.0   \n",
       "  8       25529.0           1.0             2027.0          25528.0   \n",
       "  9       25529.0           1.0             2027.0          25528.0   \n",
       "  \n",
       "     relative_size_sg  relative_size_complement  coverage_sg  \\\n",
       "  0          0.000039                  0.999961     0.000493   \n",
       "  1          0.000039                  0.999961     0.000493   \n",
       "  2          0.000039                  0.999961     0.000493   \n",
       "  3          0.000039                  0.999961     0.000493   \n",
       "  4          0.000039                  0.999961     0.000493   \n",
       "  5          0.000039                  0.999961     0.000493   \n",
       "  6          0.000039                  0.999961     0.000493   \n",
       "  7          0.000039                  0.999961     0.000493   \n",
       "  8          0.000039                  0.999961     0.000493   \n",
       "  9          0.000039                  0.999961     0.000493   \n",
       "  \n",
       "     coverage_complement  target_share_sg  target_share_complement  \\\n",
       "  0             0.999507              1.0                 0.079364   \n",
       "  1             0.999507              1.0                 0.079364   \n",
       "  2             0.999507              1.0                 0.079364   \n",
       "  3             0.999507              1.0                 0.079364   \n",
       "  4             0.999507              1.0                 0.079364   \n",
       "  5             0.999507              1.0                 0.079364   \n",
       "  6             0.999507              1.0                 0.079364   \n",
       "  7             0.999507              1.0                 0.079364   \n",
       "  8             0.999507              1.0                 0.079364   \n",
       "  9             0.999507              1.0                 0.079364   \n",
       "  \n",
       "     target_share_dataset       lift  \n",
       "  0                0.0794  12.594475  \n",
       "  1                0.0794  12.594475  \n",
       "  2                0.0794  12.594475  \n",
       "  3                0.0794  12.594475  \n",
       "  4                0.0794  12.594475  \n",
       "  5                0.0794  12.594475  \n",
       "  6                0.0794  12.594475  \n",
       "  7                0.0794  12.594475  \n",
       "  8                0.0794  12.594475  \n",
       "  9                0.0794  12.594475  },\n",
       " {'250_v_GA_0.5_SDFS':     quality                                           subgroup  size_sg  \\\n",
       "  0  0.041624                               density: [0.00:0.00[    848.0   \n",
       "  1  0.039140                                   012: [0.02:0.02[    851.0   \n",
       "  2  0.037423                                          012>=0.02    851.0   \n",
       "  3  0.032417                               density: [0.00:0.00[    853.0   \n",
       "  4  0.029811                                      density>=0.00    854.0   \n",
       "  5  0.024551                                   012: [0.02:0.02[    851.0   \n",
       "  6  0.022006                               density: [0.00:0.00[    850.0   \n",
       "  7  0.015728                                  021C: [0.00:0.00[    852.0   \n",
       "  8  0.015111                                  021C: [0.00:0.00[    851.0   \n",
       "  9  0.014259  021U: [0.00:0.00[ AND degree_centralization: [...     33.0   \n",
       "  \n",
       "     size_dataset  positives_sg  positives_dataset  size_complement  \\\n",
       "  0       25529.0         261.0             2027.0          24681.0   \n",
       "  1       25529.0         250.0             2027.0          24678.0   \n",
       "  2       25529.0         242.0             2027.0          24678.0   \n",
       "  3       25529.0         219.0             2027.0          24676.0   \n",
       "  4       25529.0         207.0             2027.0          24675.0   \n",
       "  5       25529.0         182.0             2027.0          24678.0   \n",
       "  6       25529.0         170.0             2027.0          24679.0   \n",
       "  7       25529.0         141.0             2027.0          24677.0   \n",
       "  8       25529.0         138.0             2027.0          24678.0   \n",
       "  9       25529.0          17.0             2027.0          25496.0   \n",
       "  \n",
       "     relative_size_sg  relative_size_complement  coverage_sg  \\\n",
       "  0          0.033217                  0.966783     0.128762   \n",
       "  1          0.033335                  0.966665     0.123335   \n",
       "  2          0.033335                  0.966665     0.119388   \n",
       "  3          0.033413                  0.966587     0.108041   \n",
       "  4          0.033452                  0.966548     0.102121   \n",
       "  5          0.033335                  0.966665     0.089788   \n",
       "  6          0.033295                  0.966705     0.083868   \n",
       "  7          0.033374                  0.966626     0.069561   \n",
       "  8          0.033335                  0.966665     0.068081   \n",
       "  9          0.001293                  0.998707     0.008387   \n",
       "  \n",
       "     coverage_complement  target_share_sg  target_share_complement  \\\n",
       "  0             0.871238         0.307783                 0.071553   \n",
       "  1             0.876665         0.293772                 0.072007   \n",
       "  2             0.880612         0.284371                 0.072332   \n",
       "  3             0.891959         0.256741                 0.073270   \n",
       "  4             0.897879         0.242389                 0.073759   \n",
       "  5             0.910212         0.213866                 0.074763   \n",
       "  6             0.916132         0.200000                 0.075246   \n",
       "  7             0.930439         0.165493                 0.076427   \n",
       "  8             0.931919         0.162162                 0.076546   \n",
       "  9             0.991613         0.515152                 0.078836   \n",
       "  \n",
       "     target_share_dataset      lift  \n",
       "  0                0.0794  3.876365  \n",
       "  1                0.0794  3.699904  \n",
       "  2                0.0794  3.581507  \n",
       "  3                0.0794  3.233517  \n",
       "  4                0.0794  3.052759  \n",
       "  5                0.0794  2.693530  \n",
       "  6                0.0794  2.518895  \n",
       "  7                0.0794  2.084297  \n",
       "  8                0.0794  2.042347  \n",
       "  9                0.0794  6.488063  ,\n",
       "  '250_v_SQF_0.5_BFS':     quality                                   subgroup  size_sg  size_dataset  \\\n",
       "  0  0.041624                       density: [0.00:0.00[    848.0       25529.0   \n",
       "  1  0.039140                           012: [0.02:0.02[    851.0       25529.0   \n",
       "  2  0.037423                                  012>=0.02    851.0       25529.0   \n",
       "  3  0.032417                       density: [0.00:0.00[    853.0       25529.0   \n",
       "  4  0.030495            012>=0.02 AND 021C: [0.00:0.00[    322.0       25529.0   \n",
       "  5  0.030096  012: [0.02:0.02[ AND density: [0.00:0.00[    268.0       25529.0   \n",
       "  6  0.029811                              density>=0.00    854.0       25529.0   \n",
       "  7  0.027714     012: [0.02:0.02[ AND 021C: [0.00:0.00[    215.0       25529.0   \n",
       "  8  0.027689         012>=0.02 AND density: [0.00:0.00[    224.0       25529.0   \n",
       "  9  0.026879                012>=0.02 AND density>=0.00    595.0       25529.0   \n",
       "  \n",
       "     positives_sg  positives_dataset  size_complement  relative_size_sg  \\\n",
       "  0         261.0             2027.0          24681.0          0.033217   \n",
       "  1         250.0             2027.0          24678.0          0.033335   \n",
       "  2         242.0             2027.0          24678.0          0.033335   \n",
       "  3         219.0             2027.0          24676.0          0.033413   \n",
       "  4         113.0             2027.0          25207.0          0.012613   \n",
       "  5         100.0             2027.0          25261.0          0.010498   \n",
       "  6         207.0             2027.0          24675.0          0.033452   \n",
       "  7          82.0             2027.0          25314.0          0.008422   \n",
       "  8          84.0             2027.0          25305.0          0.008774   \n",
       "  9         152.0             2027.0          24934.0          0.023307   \n",
       "  \n",
       "     relative_size_complement  coverage_sg  coverage_complement  \\\n",
       "  0                  0.966783     0.128762             0.871238   \n",
       "  1                  0.966665     0.123335             0.876665   \n",
       "  2                  0.966665     0.119388             0.880612   \n",
       "  3                  0.966587     0.108041             0.891959   \n",
       "  4                  0.987387     0.055747             0.944253   \n",
       "  5                  0.989502     0.049334             0.950666   \n",
       "  6                  0.966548     0.102121             0.897879   \n",
       "  7                  0.991578     0.040454             0.959546   \n",
       "  8                  0.991226     0.041441             0.958559   \n",
       "  9                  0.976693     0.074988             0.925012   \n",
       "  \n",
       "     target_share_sg  target_share_complement  target_share_dataset      lift  \n",
       "  0         0.307783                 0.071553                0.0794  3.876365  \n",
       "  1         0.293772                 0.072007                0.0794  3.699904  \n",
       "  2         0.284371                 0.072332                0.0794  3.581507  \n",
       "  3         0.256741                 0.073270                0.0794  3.233517  \n",
       "  4         0.350932                 0.075931                0.0794  4.419800  \n",
       "  5         0.373134                 0.076284                0.0794  4.699431  \n",
       "  6         0.242389                 0.073759                0.0794  3.052759  \n",
       "  7         0.381395                 0.076835                0.0794  4.803474  \n",
       "  8         0.375000                 0.076783                0.0794  4.722928  \n",
       "  9         0.255462                 0.075199                0.0794  3.217412  },\n",
       " {'250_v_GA_1_SDFS':     quality              subgroup  size_sg  size_dataset  positives_sg  \\\n",
       "  0  0.007586  density: [0.00:0.00[    848.0       25529.0         261.0   \n",
       "  1  0.007146      012: [0.02:0.02[    851.0       25529.0         250.0   \n",
       "  2  0.006833             012>=0.02    851.0       25529.0         242.0   \n",
       "  3  0.005925  density: [0.00:0.00[    853.0       25529.0         219.0   \n",
       "  4  0.005452         density>=0.00    854.0       25529.0         207.0   \n",
       "  5  0.004482      012: [0.02:0.02[    851.0       25529.0         182.0   \n",
       "  6  0.004015  density: [0.00:0.00[    850.0       25529.0         170.0   \n",
       "  7  0.002873     021C: [0.00:0.00[    852.0       25529.0         141.0   \n",
       "  8  0.002759     021C: [0.00:0.00[    851.0       25529.0         138.0   \n",
       "  9  0.002563  density: [0.00:0.00[    851.0       25529.0         133.0   \n",
       "  \n",
       "     positives_dataset  size_complement  relative_size_sg  \\\n",
       "  0             2027.0          24681.0          0.033217   \n",
       "  1             2027.0          24678.0          0.033335   \n",
       "  2             2027.0          24678.0          0.033335   \n",
       "  3             2027.0          24676.0          0.033413   \n",
       "  4             2027.0          24675.0          0.033452   \n",
       "  5             2027.0          24678.0          0.033335   \n",
       "  6             2027.0          24679.0          0.033295   \n",
       "  7             2027.0          24677.0          0.033374   \n",
       "  8             2027.0          24678.0          0.033335   \n",
       "  9             2027.0          24678.0          0.033335   \n",
       "  \n",
       "     relative_size_complement  coverage_sg  coverage_complement  \\\n",
       "  0                  0.966783     0.128762             0.871238   \n",
       "  1                  0.966665     0.123335             0.876665   \n",
       "  2                  0.966665     0.119388             0.880612   \n",
       "  3                  0.966587     0.108041             0.891959   \n",
       "  4                  0.966548     0.102121             0.897879   \n",
       "  5                  0.966665     0.089788             0.910212   \n",
       "  6                  0.966705     0.083868             0.916132   \n",
       "  7                  0.966626     0.069561             0.930439   \n",
       "  8                  0.966665     0.068081             0.931919   \n",
       "  9                  0.966665     0.065614             0.934386   \n",
       "  \n",
       "     target_share_sg  target_share_complement  target_share_dataset      lift  \n",
       "  0         0.307783                 0.071553                0.0794  3.876365  \n",
       "  1         0.293772                 0.072007                0.0794  3.699904  \n",
       "  2         0.284371                 0.072332                0.0794  3.581507  \n",
       "  3         0.256741                 0.073270                0.0794  3.233517  \n",
       "  4         0.242389                 0.073759                0.0794  3.052759  \n",
       "  5         0.213866                 0.074763                0.0794  2.693530  \n",
       "  6         0.200000                 0.075246                0.0794  2.518895  \n",
       "  7         0.165493                 0.076427                0.0794  2.084297  \n",
       "  8         0.162162                 0.076546                0.0794  2.042347  \n",
       "  9         0.156287                 0.076749                0.0794  1.968349  ,\n",
       "  '250_v_SQF_1_BFS':     quality                                   subgroup  size_sg  size_dataset  \\\n",
       "  0  0.007586                       density: [0.00:0.00[    848.0       25529.0   \n",
       "  1  0.007146                           012: [0.02:0.02[    851.0       25529.0   \n",
       "  2  0.006833                                  012>=0.02    851.0       25529.0   \n",
       "  3  0.005925                       density: [0.00:0.00[    853.0       25529.0   \n",
       "  4  0.005452                              density>=0.00    854.0       25529.0   \n",
       "  5  0.004482                           012: [0.02:0.02[    851.0       25529.0   \n",
       "  6  0.004103                012>=0.02 AND density>=0.00    595.0       25529.0   \n",
       "  7  0.004015                       density: [0.00:0.00[    850.0       25529.0   \n",
       "  8  0.003425            012>=0.02 AND 021C: [0.00:0.00[    322.0       25529.0   \n",
       "  9  0.003084  012: [0.02:0.02[ AND density: [0.00:0.00[    268.0       25529.0   \n",
       "  \n",
       "     positives_sg  positives_dataset  size_complement  relative_size_sg  \\\n",
       "  0         261.0             2027.0          24681.0          0.033217   \n",
       "  1         250.0             2027.0          24678.0          0.033335   \n",
       "  2         242.0             2027.0          24678.0          0.033335   \n",
       "  3         219.0             2027.0          24676.0          0.033413   \n",
       "  4         207.0             2027.0          24675.0          0.033452   \n",
       "  5         182.0             2027.0          24678.0          0.033335   \n",
       "  6         152.0             2027.0          24934.0          0.023307   \n",
       "  7         170.0             2027.0          24679.0          0.033295   \n",
       "  8         113.0             2027.0          25207.0          0.012613   \n",
       "  9         100.0             2027.0          25261.0          0.010498   \n",
       "  \n",
       "     relative_size_complement  coverage_sg  coverage_complement  \\\n",
       "  0                  0.966783     0.128762             0.871238   \n",
       "  1                  0.966665     0.123335             0.876665   \n",
       "  2                  0.966665     0.119388             0.880612   \n",
       "  3                  0.966587     0.108041             0.891959   \n",
       "  4                  0.966548     0.102121             0.897879   \n",
       "  5                  0.966665     0.089788             0.910212   \n",
       "  6                  0.976693     0.074988             0.925012   \n",
       "  7                  0.966705     0.083868             0.916132   \n",
       "  8                  0.987387     0.055747             0.944253   \n",
       "  9                  0.989502     0.049334             0.950666   \n",
       "  \n",
       "     target_share_sg  target_share_complement  target_share_dataset      lift  \n",
       "  0         0.307783                 0.071553                0.0794  3.876365  \n",
       "  1         0.293772                 0.072007                0.0794  3.699904  \n",
       "  2         0.284371                 0.072332                0.0794  3.581507  \n",
       "  3         0.256741                 0.073270                0.0794  3.233517  \n",
       "  4         0.242389                 0.073759                0.0794  3.052759  \n",
       "  5         0.213866                 0.074763                0.0794  2.693530  \n",
       "  6         0.255462                 0.075199                0.0794  3.217412  \n",
       "  7         0.200000                 0.075246                0.0794  2.518895  \n",
       "  8         0.350932                 0.075931                0.0794  4.419800  \n",
       "  9         0.373134                 0.076284                0.0794  4.699431  }]"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "subgroup_archive_results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['10', 'pol', 'GA', '0', 'SDFS']"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "list(subgroup_archive_results[0].keys())[0].split(\"_\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "from pprint import pprint"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "250_pol_GA_0.5_SDFS\n",
      "    quality                            subgroup  size_sg  positives_sg  \\\n",
      "0  0.036998         degree_centralization>=0.53    851.0         682.0   \n",
      "1  0.035496                   021U: [0.00:0.00[    851.0         675.0   \n",
      "2  0.035281                          021U>=0.00    851.0         674.0   \n",
      "3  0.032921                   021D: [0.00:0.00[    851.0         663.0   \n",
      "4  0.032854                            012<0.01    850.0         662.0   \n",
      "5  0.029488                   021D: [0.00:0.00[    851.0         647.0   \n",
      "6  0.029274  degree_centralization: [0.47:0.53[    851.0         646.0   \n",
      "7  0.025841    betweenness_centralization>=0.01    851.0         630.0   \n",
      "8  0.025412                average_degree>=2.36    851.0         628.0   \n",
      "9  0.024769                   021D: [0.00:0.00[    851.0         625.0   \n",
      "\n",
      "   positives_dataset  target_share_sg      lift  \n",
      "0            15286.0         0.801410  1.338427  \n",
      "1            15286.0         0.793184  1.324690  \n",
      "2            15286.0         0.792009  1.322727  \n",
      "3            15286.0         0.779083  1.301140  \n",
      "4            15286.0         0.778824  1.300706  \n",
      "5            15286.0         0.760282  1.269740  \n",
      "6            15286.0         0.759107  1.267777  \n",
      "7            15286.0         0.740306  1.236377  \n",
      "8            15286.0         0.737955  1.232452  \n",
      "9            15286.0         0.734430  1.226565  \n",
      "250_pol_SQF_0.5_BFS\n",
      "    quality                                           subgroup  size_sg  \\\n",
      "0  0.036998                        degree_centralization>=0.53    851.0   \n",
      "1  0.035739         021U>=0.00 AND degree_centralization>=0.53    765.0   \n",
      "2  0.035496                                  021U: [0.00:0.00[    851.0   \n",
      "3  0.035281                                         021U>=0.00    851.0   \n",
      "4  0.033841           012<0.01 AND degree_centralization>=0.53    640.0   \n",
      "5  0.033190                            012<0.01 AND 021U>=0.00    580.0   \n",
      "6  0.032921                                  021D: [0.00:0.00[    851.0   \n",
      "7  0.032854                                           012<0.01    850.0   \n",
      "8  0.032810  012<0.01 AND 021U>=0.00 AND degree_centralizat...    575.0   \n",
      "9  0.032367          021C<0.00 AND degree_centralization>=0.53    521.0   \n",
      "\n",
      "   positives_sg  positives_dataset  target_share_sg      lift  \n",
      "0         682.0            15286.0         0.801410  1.338427  \n",
      "1         616.0            15286.0         0.805229  1.344805  \n",
      "2         675.0            15286.0         0.793184  1.324690  \n",
      "3         674.0            15286.0         0.792009  1.322727  \n",
      "4         520.0            15286.0         0.812500  1.356948  \n",
      "5         475.0            15286.0         0.818966  1.367746  \n",
      "6         663.0            15286.0         0.779083  1.301140  \n",
      "7         662.0            15286.0         0.778824  1.300706  \n",
      "8         470.0            15286.0         0.817391  1.365117  \n",
      "9         430.0            15286.0         0.825336  1.378385  \n",
      "250_pol_GA_1_SDFS\n",
      "    quality                            subgroup  size_sg  positives_sg  \\\n",
      "0  0.006755         degree_centralization>=0.53    851.0         682.0   \n",
      "1  0.006481                   021U: [0.00:0.00[    851.0         675.0   \n",
      "2  0.006442                          021U>=0.00    851.0         674.0   \n",
      "3  0.006011                   021D: [0.00:0.00[    851.0         663.0   \n",
      "4  0.005995                            012<0.01    850.0         662.0   \n",
      "5  0.005384                   021D: [0.00:0.00[    851.0         647.0   \n",
      "6  0.005345  degree_centralization: [0.47:0.53[    851.0         646.0   \n",
      "7  0.004718    betweenness_centralization>=0.01    851.0         630.0   \n",
      "8  0.004640                average_degree>=2.36    851.0         628.0   \n",
      "9  0.004522                   021D: [0.00:0.00[    851.0         625.0   \n",
      "\n",
      "   positives_dataset  target_share_sg      lift  \n",
      "0            15286.0         0.801410  1.338427  \n",
      "1            15286.0         0.793184  1.324690  \n",
      "2            15286.0         0.792009  1.322727  \n",
      "3            15286.0         0.779083  1.301140  \n",
      "4            15286.0         0.778824  1.300706  \n",
      "5            15286.0         0.760282  1.269740  \n",
      "6            15286.0         0.759107  1.267777  \n",
      "7            15286.0         0.740306  1.236377  \n",
      "8            15286.0         0.737955  1.232452  \n",
      "9            15286.0         0.734430  1.226565  \n",
      "250_pol_SQF_1_BFS\n",
      "    quality                                    subgroup  size_sg  \\\n",
      "0  0.006755                 degree_centralization>=0.53    851.0   \n",
      "1  0.006481                           021U: [0.00:0.00[    851.0   \n",
      "2  0.006442                                  021U>=0.00    851.0   \n",
      "3  0.006187  021U>=0.00 AND degree_centralization>=0.53    765.0   \n",
      "4  0.006011                           021D: [0.00:0.00[    851.0   \n",
      "5  0.005995                                    012<0.01    850.0   \n",
      "6  0.005384                           021D: [0.00:0.00[    851.0   \n",
      "7  0.005358    012<0.01 AND degree_centralization>=0.53    640.0   \n",
      "8  0.005345          degree_centralization: [0.47:0.53[    851.0   \n",
      "9  0.005003                     012<0.01 AND 021U>=0.00    580.0   \n",
      "\n",
      "   positives_sg  positives_dataset  target_share_sg      lift  \n",
      "0         682.0            15286.0         0.801410  1.338427  \n",
      "1         675.0            15286.0         0.793184  1.324690  \n",
      "2         674.0            15286.0         0.792009  1.322727  \n",
      "3         616.0            15286.0         0.805229  1.344805  \n",
      "4         663.0            15286.0         0.779083  1.301140  \n",
      "5         662.0            15286.0         0.778824  1.300706  \n",
      "6         647.0            15286.0         0.760282  1.269740  \n",
      "7         520.0            15286.0         0.812500  1.356948  \n",
      "8         646.0            15286.0         0.759107  1.267777  \n",
      "9         475.0            15286.0         0.818966  1.367746  \n"
     ]
    }
   ],
   "source": [
    "for sg_res in subgroup_archive_results:\n",
    "    for res in sg_res.keys():\n",
    "        if res.split(\"_\")[3] != \"0\":\n",
    "            if res.split(\"_\")[0] == \"250\":\n",
    "                if res.split(\"_\")[1] in ['pol']:\n",
    "                    print(res)\n",
    "                    pprint(sg_res[res][[\"quality\", \"subgroup\", \"size_sg\", \"positives_sg\", \"positives_dataset\", \"target_share_sg\", \"lift\"]])\n"
   ]
  }
 ],
 "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": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}