fraud_detection.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Installation des librairies nécessaires"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "#import tensorflow as tf\n",
    "#from tensorflow import keras\n",
    "\n",
    "#import os\n",
    "#import tempfile\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "from sklearn.preprocessing import StandardScaler"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Chargement des données"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/folders/mk/hc7kkglj5fncc3271_yfcf6r0000gn/T/ipykernel_23198/2286139514.py:1: DtypeWarning: Columns (1,2,3,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22) have mixed types. Specify dtype option on import or set low_memory=False.\n",
      "  data = pd.read_csv('/Users/annabellenarsama/Desktop/FouillesDonneesMassives/guillaume.txt', sep=';', decimal=',')\n"
     ]
    },
    {
     "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>ZIBZIN</th>\n",
       "      <th>IDAvisAutorisationCheque</th>\n",
       "      <th>FlagImpaye</th>\n",
       "      <th>Montant</th>\n",
       "      <th>DateTransaction</th>\n",
       "      <th>CodeDecision</th>\n",
       "      <th>VerifianceCPT1</th>\n",
       "      <th>VerifianceCPT2</th>\n",
       "      <th>VerifianceCPT3</th>\n",
       "      <th>D2CB</th>\n",
       "      <th>...</th>\n",
       "      <th>TauxImpNb_RB</th>\n",
       "      <th>TauxImpNB_CPM</th>\n",
       "      <th>EcartNumCheq</th>\n",
       "      <th>NbrMagasin3J</th>\n",
       "      <th>DiffDateTr1</th>\n",
       "      <th>DiffDateTr2</th>\n",
       "      <th>DiffDateTr3</th>\n",
       "      <th>CA3TRetMtt</th>\n",
       "      <th>CA3TR</th>\n",
       "      <th>Heure</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>A013010004908126703060931</td>\n",
       "      <td>78643044</td>\n",
       "      <td>0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>2017-02-01 07:32:14</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>551</td>\n",
       "      <td>...</td>\n",
       "      <td>37.186668</td>\n",
       "      <td>52.076034</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>27134</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>A013011306908024927155000</td>\n",
       "      <td>78643045</td>\n",
       "      <td>0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>2017-02-01 07:43:37</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>551</td>\n",
       "      <td>...</td>\n",
       "      <td>48.844716</td>\n",
       "      <td>52.076034</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>1.797685</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>28.61</td>\n",
       "      <td>8.61</td>\n",
       "      <td>27817</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>A013010002908283134592527</td>\n",
       "      <td>78643046</td>\n",
       "      <td>0</td>\n",
       "      <td>57.64</td>\n",
       "      <td>2017-02-01 07:47:38</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>549</td>\n",
       "      <td>...</td>\n",
       "      <td>73.11828</td>\n",
       "      <td>52.076034</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>57.64</td>\n",
       "      <td>0.0</td>\n",
       "      <td>28058</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>A011010002908105209831316</td>\n",
       "      <td>78643047</td>\n",
       "      <td>0</td>\n",
       "      <td>54.29</td>\n",
       "      <td>2017-02-01 07:48:48</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>267</td>\n",
       "      <td>...</td>\n",
       "      <td>110.056926</td>\n",
       "      <td>53.554234</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>54.29</td>\n",
       "      <td>0.0</td>\n",
       "      <td>28128</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>A013010041908000125652029</td>\n",
       "      <td>78643048</td>\n",
       "      <td>0</td>\n",
       "      <td>26.9</td>\n",
       "      <td>2017-02-01 08:13:27</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>549</td>\n",
       "      <td>...</td>\n",
       "      <td>45.368313</td>\n",
       "      <td>52.076034</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1.997106</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>59.15</td>\n",
       "      <td>32.25</td>\n",
       "      <td>29607</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 23 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                      ZIBZIN IDAvisAutorisationCheque FlagImpaye Montant  \\\n",
       "0  A013010004908126703060931                 78643044          0    20.0   \n",
       "1  A013011306908024927155000                 78643045          0    20.0   \n",
       "2  A013010002908283134592527                 78643046          0   57.64   \n",
       "3  A011010002908105209831316                 78643047          0   54.29   \n",
       "4  A013010041908000125652029                 78643048          0    26.9   \n",
       "\n",
       "       DateTransaction CodeDecision VerifianceCPT1 VerifianceCPT2  \\\n",
       "0  2017-02-01 07:32:14            1              0              0   \n",
       "1  2017-02-01 07:43:37            1              0              0   \n",
       "2  2017-02-01 07:47:38            1              0              0   \n",
       "3  2017-02-01 07:48:48            0              1              1   \n",
       "4  2017-02-01 08:13:27            1              0              0   \n",
       "\n",
       "  VerifianceCPT3 D2CB  ... TauxImpNb_RB TauxImpNB_CPM EcartNumCheq  \\\n",
       "0              0  551  ...    37.186668     52.076034            0   \n",
       "1              0  551  ...    48.844716     52.076034            1   \n",
       "2              0  549  ...     73.11828     52.076034            0   \n",
       "3              1  267  ...   110.056926     53.554234            0   \n",
       "4              0  549  ...    45.368313     52.076034            1   \n",
       "\n",
       "  NbrMagasin3J DiffDateTr1 DiffDateTr2 DiffDateTr3 CA3TRetMtt  CA3TR  Heure  \n",
       "0            1         4.0         4.0         4.0       20.0    0.0  27134  \n",
       "1            2    1.797685         4.0         4.0      28.61   8.61  27817  \n",
       "2            1         4.0         4.0         4.0      57.64    0.0  28058  \n",
       "3            1         4.0         4.0         4.0      54.29    0.0  28128  \n",
       "4            1    1.997106         4.0         4.0      59.15  32.25  29607  \n",
       "\n",
       "[5 rows x 23 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = pd.read_csv('/Users/annabellenarsama/Desktop/FouillesDonneesMassives/guillaume.txt', sep=';', decimal=',')\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Description des données"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 4646774 entries, 0 to 4646773\n",
      "Data columns (total 23 columns):\n",
      " #   Column                    Dtype \n",
      "---  ------                    ----- \n",
      " 0   ZIBZIN                    object\n",
      " 1   IDAvisAutorisationCheque  object\n",
      " 2   FlagImpaye                object\n",
      " 3   Montant                   object\n",
      " 4   DateTransaction           object\n",
      " 5   CodeDecision              object\n",
      " 6   VerifianceCPT1            object\n",
      " 7   VerifianceCPT2            object\n",
      " 8   VerifianceCPT3            object\n",
      " 9   D2CB                      object\n",
      " 10  ScoringFP1                object\n",
      " 11  ScoringFP2                object\n",
      " 12  ScoringFP3                object\n",
      " 13  TauxImpNb_RB              object\n",
      " 14  TauxImpNB_CPM             object\n",
      " 15  EcartNumCheq              object\n",
      " 16  NbrMagasin3J              object\n",
      " 17  DiffDateTr1               object\n",
      " 18  DiffDateTr2               object\n",
      " 19  DiffDateTr3               object\n",
      " 20  CA3TRetMtt                object\n",
      " 21  CA3TR                     object\n",
      " 22  Heure                     object\n",
      "dtypes: object(23)\n",
      "memory usage: 815.4+ MB\n"
     ]
    }
   ],
   "source": [
    "data.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dimensions : (4646774, 23) \n",
      "\n",
      "Classe : <class 'pandas.core.frame.DataFrame'> \n",
      "\n",
      "Variables : \n",
      " Index(['ZIBZIN', 'IDAvisAutorisationCheque', 'FlagImpaye', 'Montant',\n",
      "       'DateTransaction', 'CodeDecision', 'VerifianceCPT1', 'VerifianceCPT2',\n",
      "       'VerifianceCPT3', 'D2CB', 'ScoringFP1', 'ScoringFP2', 'ScoringFP3',\n",
      "       'TauxImpNb_RB', 'TauxImpNB_CPM', 'EcartNumCheq', 'NbrMagasin3J',\n",
      "       'DiffDateTr1', 'DiffDateTr2', 'DiffDateTr3', 'CA3TRetMtt', 'CA3TR',\n",
      "       'Heure'],\n",
      "      dtype='object')\n"
     ]
    }
   ],
   "source": [
    "print(\"Dimensions :\", data.shape, '\\n') # 4,646,774 observations, 23 variables\n",
    "print(\"Classe :\", type(data), '\\n') # df\n",
    "print(\"Variables : \\n\", data.columns)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Suppresion des variables inutiles"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "ZIBZIN                                        ZIBZIN\n",
       "IDAvisAutorisationCheque    IDAvisAutorisationCheque\n",
       "FlagImpaye                                FlagImpaye\n",
       "Montant                                      Montant\n",
       "DateTransaction                      DateTransaction\n",
       "CodeDecision                            CodeDecision\n",
       "VerifianceCPT1                        VerifianceCPT1\n",
       "VerifianceCPT2                        VerifianceCPT2\n",
       "VerifianceCPT3                        VerifianceCPT3\n",
       "D2CB                                            D2CB\n",
       "ScoringFP1                                ScoringFP1\n",
       "ScoringFP2                                ScoringFP2\n",
       "ScoringFP3                                ScoringFP3\n",
       "TauxImpNb_RB                            TauxImpNb_RB\n",
       "TauxImpNB_CPM                          TauxImpNB_CPM\n",
       "EcartNumCheq                            EcartNumCheq\n",
       "NbrMagasin3J                            NbrMagasin3J\n",
       "DiffDateTr1                              DiffDateTr1\n",
       "DiffDateTr2                              DiffDateTr2\n",
       "DiffDateTr3                              DiffDateTr3\n",
       "CA3TRetMtt                                CA3TRetMtt\n",
       "CA3TR                                          CA3TR\n",
       "Heure                                          Heure\n",
       "Name: 1956360, dtype: object"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.iloc[1956360]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "data.drop(1956360,inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "data.drop(columns=['CodeDecision', 'ZIBZIN'], inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(4646773, 21)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Reformatage des variables restantes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "object\n",
      "float64\n"
     ]
    }
   ],
   "source": [
    "# Reformatage de la variable 'Montant' en variable quantitative :\n",
    "print(data['Montant'].dtypes) # object\n",
    "data['Montant'] = data['Montant'].astype(str).str.replace(\",\", \".\").astype(float)\n",
    "\n",
    "# Vérification du type de la variable :\n",
    "print(data['Montant'].dtypes) # float64"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ScoringFP1 : object\n",
      "ScoringFP2 : object\n",
      "ScoringFP3 : object\n",
      "ScoringFP1 : float64\n",
      "ScoringFP2 : float64\n",
      "ScoringFP3 : float64\n"
     ]
    }
   ],
   "source": [
    "# Reformatage des variables 'Scoring' en variables quantitatives :\n",
    "\n",
    "print(\"ScoringFP1 :\", data['ScoringFP1'].dtypes) # object\n",
    "data['ScoringFP1'] = data['ScoringFP1'].astype(str).str.replace(\",\", \".\").astype(float)\n",
    "\n",
    "print(\"ScoringFP2 :\", data['ScoringFP2'].dtypes) # object\n",
    "data['ScoringFP2'] = data['ScoringFP2'].astype(str).str.replace(\",\", \".\").astype(float)\n",
    "\n",
    "print(\"ScoringFP3 :\", data['ScoringFP3'].dtypes) # object\n",
    "data['ScoringFP3'] = data['ScoringFP3'].astype(str).str.replace(\",\", \".\").astype(float)\n",
    "\n",
    "# Vérification du type des variables 'Scoring' :\n",
    "\n",
    "print(\"ScoringFP1 :\", data['ScoringFP1'].dtypes) # float64\n",
    "\n",
    "print(\"ScoringFP2 :\", data['ScoringFP2'].dtypes) # float64\n",
    "\n",
    "print(\"ScoringFP3 :\", data['ScoringFP3'].dtypes) # float64"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "TauxImpNb_RB : object\n",
      "TauxImpNB_CPM : object\n",
      "TauxImpNb_RB : float64\n",
      "TauxImpNB_CPM : float64\n"
     ]
    }
   ],
   "source": [
    "# Reformatage des variables 'TauxImp' en variables quantitatives :\n",
    "\n",
    "print('TauxImpNb_RB :', data['TauxImpNb_RB'].dtypes) # object\n",
    "data['TauxImpNb_RB'] = data['TauxImpNb_RB'].astype(str).str.replace(\",\", \".\").astype(float)\n",
    "\n",
    "print('TauxImpNB_CPM :', data['TauxImpNB_CPM'].dtypes) # object\n",
    "data['TauxImpNB_CPM'] = data['TauxImpNB_CPM'].astype(str).str.replace(\",\", \".\").astype(float)\n",
    "\n",
    "# Vérification du type des variables :\n",
    "\n",
    "print('TauxImpNb_RB :', data['TauxImpNb_RB'].dtypes) # float64\n",
    "\n",
    "print('TauxImpNB_CPM :', data['TauxImpNB_CPM'].dtypes) # float64"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DiffDateTr1 : object\n",
      "DiffDateTr2 : object\n",
      "DiffDateTr3 : object\n",
      "DiffDateTr1 : float64\n",
      "DiffDateTr2 : float64\n",
      "DiffDateTr3 : float64\n"
     ]
    }
   ],
   "source": [
    "# Reformatage des variables 'DiffDate' en variables quantitatives :\n",
    "\n",
    "print('DiffDateTr1 :', data['DiffDateTr1'].dtypes) # object\n",
    "data['DiffDateTr1'] = data['DiffDateTr1'].astype(str).str.replace(\",\", \".\").astype(float)\n",
    "\n",
    "print('DiffDateTr2 :', data['DiffDateTr2'].dtypes) # object\n",
    "data['DiffDateTr2'] = data['DiffDateTr2'].astype(str).str.replace(\",\", \".\").astype(float)\n",
    "\n",
    "print('DiffDateTr3 :', data['DiffDateTr3'].dtypes) # object\n",
    "data['DiffDateTr3'] = data['DiffDateTr3'].astype(str).str.replace(\",\", \".\").astype(float)\n",
    "\n",
    "# Vérification du type des variables :\n",
    "\n",
    "print('DiffDateTr1 :', data['DiffDateTr1'].dtypes) # float64\n",
    "\n",
    "print('DiffDateTr2 :', data['DiffDateTr2'].dtypes) # float64\n",
    "\n",
    "print('DiffDateTr3 :', data['DiffDateTr3'].dtypes) # float64"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CA3TRetMtt : object\n",
      "CA3TR : object\n",
      "CA3TRetMtt : float64\n",
      "CA3TR : float64\n"
     ]
    }
   ],
   "source": [
    "# Reformatage des variables 'CAT3TR' en variable quantitative :\n",
    "\n",
    "print('CA3TRetMtt :', data['CA3TRetMtt'].dtypes) # object\n",
    "data['CA3TRetMtt'] = data['CA3TRetMtt'].astype(str).str.replace(\",\", \".\").astype(float)\n",
    "\n",
    "print('CA3TR :', data['CA3TR'].dtypes) # object\n",
    "data['CA3TR'] = data['CA3TR'].astype(str).str.replace(\",\", \".\").astype(float)\n",
    "\n",
    "# Vérification du type de Valeur foncière :\n",
    "\n",
    "print('CA3TRetMtt :', data['CA3TRetMtt'].dtypes) # float64\n",
    "\n",
    "print('CA3TR :', data['CA3TR'].dtypes) # float64"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DateTransaction : object\n",
      "Date : \n",
      " Année      int32\n",
      "Mois       int32\n",
      "Jour       int32\n",
      "Heure      int32\n",
      "Minute     int32\n",
      "Seconde    int32\n",
      "dtype: object\n"
     ]
    }
   ],
   "source": [
    "print('DateTransaction :', data['DateTransaction'].dtypes)\n",
    "data['DateTransaction'] = pd.to_datetime(data['DateTransaction'], yearfirst=True)\n",
    "\n",
    "# Création de trois nouvelles colonnes : Année, Mois, Jour\n",
    "data['Année'] = data['DateTransaction'].dt.year\n",
    "data['Mois'] = data['DateTransaction'].dt.month\n",
    "data['Jour'] = data['DateTransaction'].dt.day\n",
    "\n",
    "# Création de trois nouvelles colonnes : Heure, Minute, Seconde\n",
    "data['Heure'] = data['DateTransaction'].dt.hour\n",
    "data['Minute'] = data['DateTransaction'].dt.minute\n",
    "data['Seconde'] = data['DateTransaction'].dt.second\n",
    "\n",
    "print('Date : \\n', data[['Année', 'Mois', 'Jour', 'Heure', 'Minute', 'Seconde']].dtypes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "data[\"VerifianceCPT1\"] = data[\"VerifianceCPT1\"].astype(float)\n",
    "data[\"VerifianceCPT2\"] = data[\"VerifianceCPT2\"].astype(float)\n",
    "data[\"VerifianceCPT3\"] = data[\"VerifianceCPT3\"].astype(float)\n",
    "data[\"IDAvisAutorisationCheque\"] = data[\"IDAvisAutorisationCheque\"].astype(float)\n",
    "data[\"EcartNumCheq\"] = data[\"EcartNumCheq\"].astype(float)\n",
    "data[\"NbrMagasin3J\"] = data[\"NbrMagasin3J\"].astype(float)\n",
    "data[\"D2CB\"] = data[\"D2CB\"].astype(float)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "FlagImpaye : object\n",
      "FlagImpaye : bool \n",
      "\n",
      "Modalités de la cible : [False  True]\n"
     ]
    }
   ],
   "source": [
    "# Reformatage de 'FlagImpaye' en booléen :\n",
    "print('FlagImpaye :', data['FlagImpaye'].dtypes)\n",
    "data['FlagImpaye'] = data['FlagImpaye'].astype(bool)\n",
    "\n",
    "# Vérification du reformatage :\n",
    "print('FlagImpaye :', data['FlagImpaye'].dtypes, \"\\n\")\n",
    "\n",
    "# Modalités de la variable :\n",
    "print('Modalités de la cible :', data['FlagImpaye'].unique()) # vérification des valeurs booléennes de la variable FlagImpaye"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "IDAvisAutorisationCheque           float64\n",
       "FlagImpaye                            bool\n",
       "Montant                            float64\n",
       "DateTransaction             datetime64[ns]\n",
       "VerifianceCPT1                     float64\n",
       "VerifianceCPT2                     float64\n",
       "VerifianceCPT3                     float64\n",
       "D2CB                               float64\n",
       "ScoringFP1                         float64\n",
       "ScoringFP2                         float64\n",
       "ScoringFP3                         float64\n",
       "TauxImpNb_RB                       float64\n",
       "TauxImpNB_CPM                      float64\n",
       "EcartNumCheq                       float64\n",
       "NbrMagasin3J                       float64\n",
       "DiffDateTr1                        float64\n",
       "DiffDateTr2                        float64\n",
       "DiffDateTr3                        float64\n",
       "CA3TRetMtt                         float64\n",
       "CA3TR                              float64\n",
       "Heure                                int32\n",
       "Année                                int32\n",
       "Mois                                 int32\n",
       "Jour                                 int32\n",
       "Minute                               int32\n",
       "Seconde                              int32\n",
       "dtype: object"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.dtypes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Classe majoritaire (négative) vs. Classe minoritaire (positive)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Examples:\n",
      "    Total: 4646773\n",
      "   Positive: 62622 (1.35% of total)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "neg, pos = np.bincount(data['FlagImpaye'])\n",
    "total = neg + pos\n",
    "print('Examples:\\n    Total: {}\\n   Positive: {} ({:.2f}% of total)\\n'.format(\n",
    "    total, pos, 100 * pos / total))\n",
    "\n",
    "# https://www.tensorflow.org/tutorials/structured_data/imbalanced_data?hl=fr\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Imbalance Ratio: 73.20352272364345\n"
     ]
    }
   ],
   "source": [
    "IR = neg/pos\n",
    "print(f'Imbalance Ratio: {IR}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Plot des classes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: xlabel='count', ylabel='FlagImpaye'>"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAkQAAAGwCAYAAABIC3rIAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8g+/7EAAAACXBIWXMAAA9hAAAPYQGoP6dpAAAghUlEQVR4nO3dfVSUdf7/8degMiICaqJGoAiI6SkMb+q4at4nqW20a7m7uKLW2Uopicqb+ubNboVuauZqyu528CbN1DS3WrNA09VqQwzvStO8odSEREEwUZn5/eE2vwhTGIe5wM/zcc6cw1xzzcV7mjo8u+YzMzan0+kUAACAwXysHgAAAMBqBBEAADAeQQQAAIxHEAEAAOMRRAAAwHgEEQAAMB5BBAAAjFfX6gFqA4fDoWPHjikgIEA2m83qcQAAQCU4nU6dOXNGISEh8vG58jkggqgSjh07prCwMKvHAAAAbvjmm28UGhp6xX0IokoICAiQdOkfaGBgoMXTAACAyigqKlJYWJjr7/iVEESV8OPLZIGBgQQRAAC1TGWWu7CoGgAAGI8gAgAAxiOIAACA8QgiAABgPIIIAAAYjyACAADGI4gAAIDxCCIAAGA8gggAABiPIAIAAMYjiAAAgPEIIgAAYDyCCAAAGI8gAgAAxiOIAACA8QgiAABgvLpWD1CbtJzQUja7zeoxAAC4rpx6+ZTVI3CGCAAAgCACAADGI4gAAIDxCCIAAGA8gggAABiPIAIAAMYjiAAAgPEIIgAAYDyCCAAAGI8gAgAAxiOIAACA8QgiAABgPIIIAAAYjyACAADGI4gAAIDxCCIAAGA8gggAABiPIAIAAMYjiAAAgPEIIgAAYDyCCAAAGI8gAgAAxiOIAACA8QgiAABgPIIIAAAYjyACAADGI4gAAIDxCCIAAGA8gggAABiPIAIAAMYjiAAAgPEIIgAAYDyCCAAAGI8gAgAAxiOIAACA8QgiAABgPIIIAAAYjyACAADGI4gAAIDxCCIAAGA8gggAABiPIAIAAMYjiAAAgPEIIgAAYDyCCAAAGI8gAgAAxiOIAACA8QgiAABgPIIIAAAYjyACAADGI4gAAIDxCCIAAGA8gggAABiPIAIAAMYjiAAAgPEIIgAAYDyCCAAAGI8gAgAAxiOIAACA8QgiAABgPIIIAAAYjyACAADGI4gAAIDxCCIAAGA8gggAABiPIAIAAMYjiAAAgPEIIgAAYLxaGUQLFy5Uo0aNrB4DAABcJywNohEjRshms1W4HDhwwMqxAACAYepaPUBcXJzS09PLbQsODrZoGgAAYCLLXzKz2+1q0aJFucsrr7yiW2+9Vf7+/goLC9Po0aNVXFz8i8fYsWOHevfurYCAAAUGBqpTp07atm2b6/YtW7aoR48e8vPzU1hYmB5//HGVlJR44+EBAIBawPIguhwfHx/NmTNHe/bs0aJFi7RhwwaNGzfuF/dPSEhQaGiosrKylJ2drQkTJqhevXqSpK+//lpxcXH67W9/q507d+rNN9/Uli1blJSU9IvHKy0tVVFRUbkLAAC4fln+ktm7776rhg0buq7ffffdWrlypet6eHi4nn/+eT3yyCN69dVXL3uM3NxcPf3007r55pslSW3atHHdlpqaqoSEBCUnJ7tumzNnjnr27Kn58+erfv36FY6XmpqqqVOneuLhAQCAWsDyIOrdu7fmz5/vuu7v76+MjAylpqZq7969Kioq0sWLF3Xu3DmdPXtWDRo0qHCMlJQUPfTQQ1qyZIn69eun+++/X5GRkZIuvZy2c+dOLV261LW/0+mUw+HQoUOH1K5duwrHmzhxolJSUlzXi4qKFBYW5smHDQAAahDLXzLz9/dXVFSU61JaWqrBgwcrJiZGb731lrKzszVv3jxJ0vnz5y97jClTpmjPnj0aNGiQNmzYoPbt22vNmjWSpOLiYj388MPKyclxXXbs2KH9+/e7ounn7Ha7AgMDy10AAMD1y/IzRD+XnZ0th8OhmTNnysfnUq+tWLHiqveLjo5WdHS0nnjiCf3+979Xenq67rvvPnXs2FFffPGFoqKiqnt0AABQS1l+hujnoqKidOHCBf3tb3/TwYMHtWTJEi1YsOAX9//hhx+UlJSkjz76SEeOHNHWrVuVlZXleils/Pjx+vjjj5WUlKScnBzt379fa9euveKiagAAYJYaF0QdOnTQrFmzNH36dN1yyy1aunSpUlNTf3H/OnXq6OTJkxo+fLiio6P1wAMP6O6773Ytio6JidGmTZv01VdfqUePHoqNjdWkSZMUEhLirYcEAABqOJvT6XRaPURNV1RUpKCgIAU9GiSb3Wb1OAAAXFdOvXyqWo7749/vwsLCq64HrnFniAAAALyNIAIAAMYjiAAAgPEIIgAAYDyCCAAAGI8gAgAAxiOIAACA8QgiAABgPIIIAAAYjyACAADGI4gAAIDxCCIAAGA8gggAABiPIAIAAMYjiAAAgPEIIgAAYDyCCAAAGI8gAgAAxiOIAACA8QgiAABgPIIIAAAYjyACAADGI4gAAIDxCCIAAGA8gggAABiPIAIAAMYjiAAAgPEIIgAAYDyCCAAAGI8gAgAAxiOIAACA8QgiAABgPIIIAAAYjyACAADGI4gAAIDxCCIAAGA8gggAABiPIAIAAMYjiAAAgPEIIgAAYDyCCAAAGI8gAgAAxiOIAACA8QgiAABgPIIIAAAYjyACAADGI4gAAIDxCCIAAGA8t4Po4sWLysjIUFpams6cOSNJOnbsmIqLiz02HAAAgDfUdedOR44cUVxcnHJzc1VaWqr+/fsrICBA06dPV2lpqRYsWODpOQEAAKqNW2eIxo4dq86dO+vUqVPy8/Nzbb/vvvuUmZnpseEAAAC8wa0zRP/5z3/08ccfy9fXt9z28PBwHT161CODAQAAeItbZ4gcDofKysoqbP/2228VEBBwzUMBAAB4k1tBdNddd2n27Nmu6zabTcXFxZo8ebIGDhzoqdkAAAC8wq2XzGbOnKkBAwaoffv2OnfunP7whz9o//79atq0qd544w1PzwgAAFCt3Aqi0NBQ7dixQ8uXL9fOnTtVXFysBx98UAkJCeUWWQMAANQGbgVRSUmJ/P39NWzYME/PAwAA4HVurSFq3ry5Ro0apS1btnh6HgAAAK9zK4hef/11FRQUqE+fPoqOjta0adN07NgxT88GAADgFW4FUXx8vN5++20dPXpUjzzyiJYtW6ZWrVpp8ODBWr16tS5evOjpOQEAAKrNNX25a3BwsFJSUrRz507NmjVLGRkZGjJkiEJCQjRp0iSdPXvWU3MCAABUG7cWVf/oxIkTWrRokRYuXKgjR45oyJAhevDBB/Xtt99q+vTp+vTTT/XBBx94alYAAIBq4VYQrV69Wunp6Vq/fr3at2+v0aNHa9iwYWrUqJFrn1/96ldq166dp+YEAACoNm4F0ciRI/W73/1OW7duVZcuXS67T0hIiJ599tlrGg4AAMAb3Aqi48ePq0GDBlfcx8/PT5MnT3ZrKAAAAG9yK4h+GkPnzp3T+fPny90eGBh4bVMBAAB4kdufVD1+/HitWLFCJ0+erHB7WVnZNQ9WE+VOyyX2AAC4Drn1tvtx48Zpw4YNmj9/vux2u/75z39q6tSpCgkJ0eLFiz09IwAAQLVy6wzRO++8o8WLF6tXr14aOXKkevTooaioKLVq1UpLly5VQkKCp+cEAACoNm6dISooKFBERISkS+uFCgoKJEndu3fX5s2bPTcdAACAF7gVRBERETp06JAk6eabb9aKFSskXTpz9NPPIgIAAKgN3AqikSNHaseOHZKkCRMmaN68eapfv76eeOIJPf300x4dEAAAoLrZnE6n81oPcvjwYW3fvl1RUVGKiYnxxFw1SlFRkYKCglRYWMi7zAAAqCWq8vf7mr7L7Efh4eEKDw/3xKEAAAC8zu1vu8/MzNTgwYMVGRmpyMhIDR48WBkZGZ6cDQAAwCvcCqJXX31VcXFxCggI0NixYzV27FgFBgZq4MCBmjdvnqdnBAAAqFZurSEKDQ3VhAkTlJSUVG77vHnz9OKLL+ro0aMeG7AmYA0RAAC1T1X+frt1huj06dOKi4ursP2uu+5SYWGhO4cEAACwjFtB9Otf/1pr1qypsH3t2rUaPHjwNQ8FAADgTW69y6x9+/Z64YUX9NFHH6lr166SpE8//VRbt27Vk08+qTlz5rj2ffzxxz0zKQAAQDVxaw1R69atK3dwm00HDx6s8lA1DWuIAACofar9c4h+/NoOAACA64Hbn0MEAABwvXDrDJHT6dSqVau0ceNG5eXlyeFwlLt99erVHhkOAADAG9wKouTkZKWlpal3795q3ry5bDabp+cCAADwGreCaMmSJVq9erUGDhzo6XkAAAC8zq01REFBQYqIiPD0LAAAAJZwK4imTJmiqVOn6ocffvD0PAAAAF7n1ktmDzzwgN544w01a9ZM4eHhqlevXrnbt2/f7pHhAAAAvMGtIEpMTFR2draGDRvGomoAAFDruRVE7733ntavX6/u3bt7eh4AAACvc2sNUVhYGF9hAQAArhtuBdHMmTM1btw4HT582MPjAAAAeJ9bL5kNGzZMZ8+eVWRkpBo0aFBhUXVBQYFHhgMAAPAGt4Jo9uzZHh4DAADAOm6/ywwAAOB6UaUgKioqqtR+LLgGAAC1SZWCqFGjRlf8zCGn0ymbzaaysrJrHgwAAMBbqhREGzdurK45AAAALFOlIOrZs2d1zQEAAGAZtxZV/9JaIpvNJrvdLl9f32saCgAAwJvcCqKrrSUKDQ3ViBEjNHnyZPn4uPXZjwAAAF7jVhAtXLhQzz77rEaMGKHbb79dkvTZZ59p0aJF+r//+z/l5+drxowZstvteuaZZzw6MAAAgKfZnE6ns6p36tu3rx5++GE98MAD5bavWLFCaWlpyszM1JIlS/TCCy9o7969HhvWKkVFRQoKClLQo0Gy2SueGTv18ikLpgIAAFfy49/vwsLCq34kkFuvZ3388ceKjY2tsD02NlaffPKJJKl79+7Kzc115/AAAABe5fa33b/22msVtr/22msKCwuTJJ08eVKNGze+tukAAAC8wK01RDNmzND999+vdevWqUuXLpKkbdu2ae/evVq1apUkKSsrS0OHDvXcpAAAANXErTVEknT48GGlpaVp3759kqS2bdvq4YcfVnh4uCfnqxFYQwQAQO1TlTVEbp0hkqTw8HClpqa6e3cAAIAaw+0gkqSzZ88qNzdX58+fL7c9JibmmoYCAADwJreCKD8/XyNHjtS6desueztf7goAAGoTt95llpycrNOnT+u///2v/Pz89P7772vRokVq06aN/vWvf3l6RgAAgGrl1hmiDRs2aO3atercubN8fHzUqlUr9e/fX4GBgUpNTdWgQYM8PScAAEC1cesMUUlJiZo1ayZJaty4sfLz8yVJt956q7Zv3+656QAAALzArSBq27at6+32HTp0UFpamo4ePaoFCxboxhtv9OiAAAAA1c2tl8zGjh2r48ePS5ImT56suLg4LV26VL6+vlq4cKEn5wMAAKh2bgXRsGHDXD936tRJR44c0d69e9WyZUs1bdrUY8MBAAB4wzV9DtGPGjRooI4dO3riUAAAAF5X6SBKSUmp9EFnzZrl1jAAAABWqHQQff7555Xaz2ar+F1fAAAANVmlg2jjxo06ePCgwsPD5ePj1pvTAAAAaqQqlU2bNm30/fffu64PHTpUJ06c8PhQAAAA3lSlIHI6neWu//vf/1ZJSYlHBwIAAPA2XvsCAADGq1IQ2Wy2CoumWUQNAABquyp9DpHT6dSIESNkt9slSefOndMjjzwif3//cvutXr3acxMCAABUsyoFUWJiYrnrP/3EagAAgNqqSkGUnp5eXXMAAABYhkXVAADAeAQRAAAwHkEEAACMRxABAADjEUQAAMB4BBEAADAeQQQAAIxHEAEAAOMRRAAAwHgEEQAAMB5BBAAAjEcQAQAA4xFEAADAeAQRAAAwHkEEAACMRxABAADjEUQAAMB4BBEAADAeQQQAAIxHEAEAAOMRRAAAwHgEEQAAMB5BBAAAjEcQAQAA4xFEAADAeAQRAAAwHkEEAACMRxABAADjEUQAAMB4BBEAADAeQQQAAIxHEAEAAOMRRAAAwHgEEQAAMB5BBAAAjEcQAQAA4xFEAADAeAQRAAAwHkEEAACMRxABAADjEUQAAMB4BBEAADAeQQQAAIxHEAEAAOMRRAAAwHgEEQAAMF6NCiKbzXbFy5QpU6weEQAAXIfqWj3ATx0/ftz185tvvqlJkyZp3759rm0NGzZ0/ex0OlVWVqa6dWvUQwAAALVQjTpD1KJFC9clKChINpvNdX3v3r0KCAjQunXr1KlTJ9ntdm3ZskUjRoxQfHx8ueMkJyerV69erusOh0Opqalq3bq1/Pz81KFDB61ateoX5ygtLVVRUVG5CwAAuH7VutMrEyZM0IwZMxQREaHGjRtX6j6pqal6/fXXtWDBArVp00abN2/WsGHDFBwcrJ49e152/6lTp3p6dAAAUEPVuiD685//rP79+1d6/9LSUr344ovKyMhQ165dJUkRERHasmWL0tLSLhtEEydOVEpKiut6UVGRwsLCrn14AABQI9W6IOrcuXOV9j9w4IDOnj1bIaLOnz+v2NjYy97HbrfLbre7PSMAAKhdal0Q+fv7l7vu4+Mjp9NZbtuFCxdcPxcXF0uS3nvvPd10003l9iN6AACAVAuD6OeCg4O1e/fucttycnJUr149SVL79u1lt9uVm5t72ZfHAAAAan0Q9enTRy+99JIWL16srl276vXXX9fu3btdL4cFBAToqaee0hNPPCGHw6Hu3bursLBQW7duVWBgoBITEy1+BAAAwGq1PogGDBig5557TuPGjdO5c+c0atQoDR8+XLt27XLt85e//EXBwcFKTU3VwYMH1ahRI3Xs2FHPPPOMhZMDAICawub8+QIcVFBUVKSgoCAFPRokm91W4fZTL5+yYCoAAHAlP/79LiwsVGBg4BX3rVEfzAgAAGAFgggAABiPIAIAAMYjiAAAgPEIIgAAYDyCCAAAGI8gAgAAxiOIAACA8QgiAABgPIIIAAAYjyACAADGI4gAAIDxCCIAAGA8gggAABiPIAIAAMYjiAAAgPEIIgAAYDyCCAAAGI8gAgAAxiOIAACA8QgiAABgPIIIAAAYjyACAADGI4gAAIDxCCIAAGA8gggAABiPIAIAAMYjiAAAgPEIIgAAYDyCCAAAGI8gAgAAxiOIAACA8QgiAABgPIIIAAAYjyACAADGI4gAAIDxCCIAAGA8gggAABiPIAIAAMYjiAAAgPEIIgAAYDyCCAAAGI8gAgAAxiOIAACA8QgiAABgPIIIAAAYjyACAADGI4gAAIDxCCIAAGA8gggAABiPIAIAAMYjiAAAgPEIIgAAYDyCCAAAGI8gAgAAxiOIAACA8QgiAABgPIIIAAAYjyACAADGI4gAAIDxCCIAAGA8gggAABiPIAIAAMYjiAAAgPEIIgAAYDyCCAAAGK+u1QPUJrnTchUYGGj1GAAAwMM4QwQAAIxHEAEAAOMRRAAAwHgEEQAAMB5BBAAAjEcQAQAA4xFEAADAeAQRAAAwHkEEAACMRxABAADjEUQAAMB4BBEAADAeQQQAAIxHEAEAAOMRRAAAwHgEEQAAMB5BBAAAjFfX6gFqA6fTKUkqKiqyeBIAAFBZP/7d/vHv+JUQRJVw8uRJSVJYWJjFkwAAgKo6c+aMgoKCrrgPQVQJTZo0kSTl5uZe9R8ovKuoqEhhYWH65ptvFBgYaPU4+Amem5qL56bm4rnxLKfTqTNnzigkJOSq+xJEleDjc2mpVVBQEP+C1lCBgYE8NzUUz03NxXNTc/HceE5lT2SwqBoAABiPIAIAAMYjiCrBbrdr8uTJstvtVo+Cn+G5qbl4bmounpuai+fGOjZnZd6LBgAAcB3jDBEAADAeQQQAAIxHEAEAAOMRRAAAwHgEUSXMmzdP4eHhql+/vu644w599tlnVo9kvM2bN+uee+5RSEiIbDab3n77batHwv+kpqaqS5cuCggIULNmzRQfH699+/ZZPRYkzZ8/XzExMa4P/evatavWrVtn9Vi4jGnTpslmsyk5OdnqUYxBEF3Fm2++qZSUFE2ePFnbt29Xhw4dNGDAAOXl5Vk9mtFKSkrUoUMHzZs3z+pR8DObNm3SmDFj9Omnn+rDDz/UhQsXdNddd6mkpMTq0YwXGhqqadOmKTs7W9u2bVOfPn107733as+ePVaPhp/IyspSWlqaYmJirB7FKLzt/iruuOMOdenSRXPnzpUkORwOhYWF6bHHHtOECRMsng6SZLPZtGbNGsXHx1s9Ci4jPz9fzZo106ZNm3TnnXdaPQ5+pkmTJnrppZf04IMPWj0KJBUXF6tjx4569dVX9fzzz+u2227T7NmzrR7LCJwhuoLz588rOztb/fr1c23z8fFRv3799Mknn1g4GVB7FBYWSvr/X5KMmqGsrEzLly9XSUmJunbtavU4+J8xY8Zo0KBB5f7uwDv4ctcr+P7771VWVqbmzZuX2968eXPt3bvXoqmA2sPhcCg5OVndunXTLbfcYvU4kLRr1y517dpV586dU8OGDbVmzRq1b9/e6rEgafny5dq+fbuysrKsHsVIBBGAajNmzBjt3r1bW7ZssXoU/E/btm2Vk5OjwsJCrVq1SomJidq0aRNRZLFvvvlGY8eO1Ycffqj69etbPY6RCKIraNq0qerUqaMTJ06U237ixAm1aNHCoqmA2iEpKUnvvvuuNm/erNDQUKvHwf/4+voqKipKktSpUydlZWXplVdeUVpamsWTmS07O1t5eXnq2LGja1tZWZk2b96suXPnqrS0VHXq1LFwwusfa4iuwNfXV506dVJmZqZrm8PhUGZmJq+5A7/A6XQqKSlJa9as0YYNG9S6dWurR8IVOBwOlZaWWj2G8fr27atdu3YpJyfHdencubMSEhKUk5NDDHkBZ4iuIiUlRYmJiercubNuv/12zZ49WyUlJRo5cqTVoxmtuLhYBw4ccF0/dOiQcnJy1KRJE7Vs2dLCyTBmzBgtW7ZMa9euVUBAgL777jtJUlBQkPz8/CyezmwTJ07U3XffrZYtW+rMmTNatmyZPvroI61fv97q0YwXEBBQYZ2dv7+/brjhBtbfeQlBdBVDhw5Vfn6+Jk2apO+++0633Xab3n///QoLreFd27ZtU+/evV3XU1JSJEmJiYlauHChRVNBuvThf5LUq1evctvT09M1YsQI7w8El7y8PA0fPlzHjx9XUFCQYmJitH79evXv39/q0QDL8TlEAADAeKwhAgAAxiOIAACA8QgiAABgPIIIAAAYjyACAADGI4gAAIDxCCIAAGA8gggAAFhi8+bNuueeexQSEiKbzaa33367ysdwOp2aMWOGoqOjZbfbddNNN+mFF16o8nH4pGoAAGCJkpISdejQQaNGjdJvfvMbt44xduxYffDBB5oxY4ZuvfVWFRQUqKCgoMrH4ZOqAcBNhw8fVuvWrfX555/rtttus3ocoFaz2Wxas2aN4uPjXdtKS0v17LPP6o033tDp06d1yy23aPr06a6vBvryyy8VExOj3bt3q23bttf0+3nJDAAA1EhJSUn65JNPtHz5cu3cuVP333+/4uLitH//fknSO++8o4iICL377rtq3bq1wsPD9dBDD7l1hoggAlBrORwO/fWvf1VUVJTsdrtatmzpWjuwa9cu9enTR35+frrhhhv0pz/9ScXFxa779urVS8nJyeWOFx8fX+4LaMPDw/Xiiy9q1KhRCggIUMuWLfX3v//ddXvr1q0lSbGxsbLZbBW+0BaA+3Jzc5Wenq6VK1eqR48eioyM1FNPPaXu3bsrPT1dknTw4EEdOXJEK1eu1OLFi7Vw4UJlZ2dryJAhVf59BBGAWmvixImaNm2annvuOX3xxRdatmyZmjdvrpKSEg0YMECNGzdWVlaWVq5cqYyMDCUlJVX5d8ycOVOdO3fW559/rtGjR+vRRx/Vvn37JEmfffaZJCkjI0PHjx/X6tWrPfr4AJPt2rVLZWVlio6OVsOGDV2XTZs26euvv5Z06X+KSktLtXjxYvXo0UO9evXSa6+9po0bN7r+O60sFlUDqJXOnDmjV155RXPnzlViYqIkKTIyUt27d9c//vEPnTt3TosXL5a/v78kae7cubrnnns0ffp0NW/evNK/Z+DAgRo9erQkafz48Xr55Ze1ceNGtW3bVsHBwZKkG264QS1atPDwIwTMVlxcrDp16ig7O1t16tQpd1vDhg0lSTfeeKPq1q2r6Oho123t2rWTdOkMU1XWFRFEAGqlL7/8UqWlperbt+9lb+vQoYMrhiSpW7ducjgc2rdvX5WCKCYmxvWzzWZTixYtlJeXd23DA7iq2NhYlZWVKS8vTz169LjsPt26ddPFixf19ddfKzIyUpL01VdfSZJatWpVpd9HEAGolfz8/K7p/j4+Pvr5m2wvXLhQYb969eqVu26z2eRwOK7pdwO4pLi4WAcOHHBdP3TokHJyctSkSRNFR0crISFBw4cP18yZMxUbG6v8/HxlZmYqJiZGgwYNUr9+/dSxY0eNGjVKs2fPlsPh0JgxY9S/f/9yZ40qgzVEAGqlNm3ayM/PT5mZmRVua9eunXbs2KGSkhLXtq1bt8rHx8d1Cj04OFjHjx933V5WVqbdu3dXaQZfX1/XfQFU3bZt2xQbG6vY2FhJUkpKimJjYzVp0iRJUnp6uoYPH64nn3xSbdu2VXx8vLKystSyZUtJl/7H5p133lHTpk115513atCgQWrXrp2WL19e5Vk4QwSgVqpfv77Gjx+vcePGydfXV926dVN+fr727NmjhIQETZ48WYmJiZoyZYry8/P12GOP6Y9//KPr5bI+ffooJSVF7733niIjIzVr1iydPn26SjM0a9ZMfn5+ev/99xUaGqr69esrKCioGh4tcH3q1atXhTO1P1WvXj1NnTpVU6dO/cV9QkJC9NZbb13zLJwhAlBrPffcc3ryySc1adIktWvXTkOHDlVeXp4aNGig9evXq6CgQF26dNGQIUPUt29fzZ0713XfUaNGKTExUcOHD1fPnj0VERGh3r17V+n3161bV3PmzFFaWppCQkJ07733evohAvASPqkaAAAYjzNEAADAeAQRAAAwHkEEAACMRxABAADjEUQAAMB4BBEAADAeQQQAAIxHEAEAAOMRRAAAwHgEEQAAMB5BBAAAjPf/AAZT1IgCqRTBAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.barplot(data['FlagImpaye'].value_counts(), orient='h', color='green')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train-Test Split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "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>IDAvisAutorisationCheque</th>\n",
       "      <th>FlagImpaye</th>\n",
       "      <th>Montant</th>\n",
       "      <th>DateTransaction</th>\n",
       "      <th>VerifianceCPT1</th>\n",
       "      <th>VerifianceCPT2</th>\n",
       "      <th>VerifianceCPT3</th>\n",
       "      <th>D2CB</th>\n",
       "      <th>ScoringFP1</th>\n",
       "      <th>ScoringFP2</th>\n",
       "      <th>...</th>\n",
       "      <th>DiffDateTr2</th>\n",
       "      <th>DiffDateTr3</th>\n",
       "      <th>CA3TRetMtt</th>\n",
       "      <th>CA3TR</th>\n",
       "      <th>Heure</th>\n",
       "      <th>Année</th>\n",
       "      <th>Mois</th>\n",
       "      <th>Jour</th>\n",
       "      <th>Minute</th>\n",
       "      <th>Seconde</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>78643044.0</td>\n",
       "      <td>False</td>\n",
       "      <td>20.00</td>\n",
       "      <td>2017-02-01 07:32:14</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>551.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>20.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>7</td>\n",
       "      <td>2017</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>32</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>78643045.0</td>\n",
       "      <td>False</td>\n",
       "      <td>20.00</td>\n",
       "      <td>2017-02-01 07:43:37</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>551.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>28.61</td>\n",
       "      <td>8.61</td>\n",
       "      <td>7</td>\n",
       "      <td>2017</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>43</td>\n",
       "      <td>37</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>78643046.0</td>\n",
       "      <td>False</td>\n",
       "      <td>57.64</td>\n",
       "      <td>2017-02-01 07:47:38</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>549.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>57.64</td>\n",
       "      <td>0.00</td>\n",
       "      <td>7</td>\n",
       "      <td>2017</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>47</td>\n",
       "      <td>38</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1956361</th>\n",
       "      <td>78643047.0</td>\n",
       "      <td>True</td>\n",
       "      <td>54.29</td>\n",
       "      <td>2017-02-01 07:48:48</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>267.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>54.29</td>\n",
       "      <td>0.00</td>\n",
       "      <td>7</td>\n",
       "      <td>2017</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>48</td>\n",
       "      <td>48</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>78643047.0</td>\n",
       "      <td>False</td>\n",
       "      <td>54.29</td>\n",
       "      <td>2017-02-01 07:48:48</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>267.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>54.29</td>\n",
       "      <td>0.00</td>\n",
       "      <td>7</td>\n",
       "      <td>2017</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>48</td>\n",
       "      <td>48</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4646769</th>\n",
       "      <td>84966399.0</td>\n",
       "      <td>False</td>\n",
       "      <td>23.94</td>\n",
       "      <td>2017-11-30 21:20:15</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>4.788000</td>\n",
       "      <td>-22.542614</td>\n",
       "      <td>...</td>\n",
       "      <td>5.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>23.94</td>\n",
       "      <td>0.00</td>\n",
       "      <td>21</td>\n",
       "      <td>2017</td>\n",
       "      <td>11</td>\n",
       "      <td>30</td>\n",
       "      <td>20</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4646770</th>\n",
       "      <td>84966400.0</td>\n",
       "      <td>False</td>\n",
       "      <td>92.60</td>\n",
       "      <td>2017-11-30 21:22:49</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>514.0</td>\n",
       "      <td>0.060052</td>\n",
       "      <td>-2.236379</td>\n",
       "      <td>...</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>262.45</td>\n",
       "      <td>169.85</td>\n",
       "      <td>21</td>\n",
       "      <td>2017</td>\n",
       "      <td>11</td>\n",
       "      <td>30</td>\n",
       "      <td>22</td>\n",
       "      <td>49</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4646771</th>\n",
       "      <td>84966401.0</td>\n",
       "      <td>False</td>\n",
       "      <td>69.00</td>\n",
       "      <td>2017-11-30 21:24:37</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.631579</td>\n",
       "      <td>-1.062954</td>\n",
       "      <td>...</td>\n",
       "      <td>5.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>69.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>21</td>\n",
       "      <td>2017</td>\n",
       "      <td>11</td>\n",
       "      <td>30</td>\n",
       "      <td>24</td>\n",
       "      <td>37</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4646772</th>\n",
       "      <td>84966402.0</td>\n",
       "      <td>False</td>\n",
       "      <td>57.80</td>\n",
       "      <td>2017-11-30 21:57:32</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>28.900000</td>\n",
       "      <td>0.567666</td>\n",
       "      <td>...</td>\n",
       "      <td>5.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>57.80</td>\n",
       "      <td>0.00</td>\n",
       "      <td>21</td>\n",
       "      <td>2017</td>\n",
       "      <td>11</td>\n",
       "      <td>30</td>\n",
       "      <td>57</td>\n",
       "      <td>32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4646773</th>\n",
       "      <td>84966405.0</td>\n",
       "      <td>False</td>\n",
       "      <td>42.21</td>\n",
       "      <td>2017-11-30 22:07:13</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>461.0</td>\n",
       "      <td>0.004578</td>\n",
       "      <td>-26.658799</td>\n",
       "      <td>...</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>42.21</td>\n",
       "      <td>0.00</td>\n",
       "      <td>22</td>\n",
       "      <td>2017</td>\n",
       "      <td>11</td>\n",
       "      <td>30</td>\n",
       "      <td>7</td>\n",
       "      <td>13</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>4646773 rows × 26 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         IDAvisAutorisationCheque  FlagImpaye  Montant     DateTransaction  \\\n",
       "0                      78643044.0       False    20.00 2017-02-01 07:32:14   \n",
       "1                      78643045.0       False    20.00 2017-02-01 07:43:37   \n",
       "2                      78643046.0       False    57.64 2017-02-01 07:47:38   \n",
       "1956361                78643047.0        True    54.29 2017-02-01 07:48:48   \n",
       "3                      78643047.0       False    54.29 2017-02-01 07:48:48   \n",
       "...                           ...         ...      ...                 ...   \n",
       "4646769                84966399.0       False    23.94 2017-11-30 21:20:15   \n",
       "4646770                84966400.0       False    92.60 2017-11-30 21:22:49   \n",
       "4646771                84966401.0       False    69.00 2017-11-30 21:24:37   \n",
       "4646772                84966402.0       False    57.80 2017-11-30 21:57:32   \n",
       "4646773                84966405.0       False    42.21 2017-11-30 22:07:13   \n",
       "\n",
       "         VerifianceCPT1  VerifianceCPT2  VerifianceCPT3   D2CB  ScoringFP1  \\\n",
       "0                   0.0             0.0             0.0  551.0    0.000000   \n",
       "1                   0.0             0.0             0.0  551.0    0.000000   \n",
       "2                   0.0             0.0             0.0  549.0    0.000000   \n",
       "1956361             1.0             1.0             1.0  267.0    0.000000   \n",
       "3                   1.0             1.0             1.0  267.0    0.000000   \n",
       "...                 ...             ...             ...    ...         ...   \n",
       "4646769             0.0             0.0             0.0    1.0    4.788000   \n",
       "4646770             0.0             0.0             0.0  514.0    0.060052   \n",
       "4646771             0.0             0.0             0.0    1.0    3.631579   \n",
       "4646772             0.0             0.0             0.0    1.0   28.900000   \n",
       "4646773             0.0             0.0             0.0  461.0    0.004578   \n",
       "\n",
       "         ScoringFP2  ...  DiffDateTr2  DiffDateTr3  CA3TRetMtt   CA3TR  Heure  \\\n",
       "0          0.000000  ...          4.0          4.0       20.00    0.00      7   \n",
       "1          0.000000  ...          4.0          4.0       28.61    8.61      7   \n",
       "2          0.000000  ...          4.0          4.0       57.64    0.00      7   \n",
       "1956361    0.000000  ...          4.0          4.0       54.29    0.00      7   \n",
       "3          0.000000  ...          4.0          4.0       54.29    0.00      7   \n",
       "...             ...  ...          ...          ...         ...     ...    ...   \n",
       "4646769  -22.542614  ...          5.0          5.0       23.94    0.00     21   \n",
       "4646770   -2.236379  ...          4.0          4.0      262.45  169.85     21   \n",
       "4646771   -1.062954  ...          5.0          5.0       69.00    0.00     21   \n",
       "4646772    0.567666  ...          5.0          5.0       57.80    0.00     21   \n",
       "4646773  -26.658799  ...          4.0          4.0       42.21    0.00     22   \n",
       "\n",
       "         Année  Mois  Jour  Minute  Seconde  \n",
       "0         2017     2     1      32       14  \n",
       "1         2017     2     1      43       37  \n",
       "2         2017     2     1      47       38  \n",
       "1956361   2017     2     1      48       48  \n",
       "3         2017     2     1      48       48  \n",
       "...        ...   ...   ...     ...      ...  \n",
       "4646769   2017    11    30      20       15  \n",
       "4646770   2017    11    30      22       49  \n",
       "4646771   2017    11    30      24       37  \n",
       "4646772   2017    11    30      57       32  \n",
       "4646773   2017    11    30       7       13  \n",
       "\n",
       "[4646773 rows x 26 columns]"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Trier les données en fonction de la date :\n",
    "df = data.sort_values(by='DateTransaction')\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.drop(columns=['DateTransaction'], inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(4646773, 25)"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Split train/test sets selon la date de transaction :\n",
    "train_size = 0.7\n",
    "split_index = int(len(df) * train_size)\n",
    "\n",
    "train_data = df[:split_index]\n",
    "test_data = df[split_index:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Split des features et du label :\n",
    "\n",
    "X_train = train_data.drop(columns=['FlagImpaye'])\n",
    "X_test = test_data.drop(columns=['FlagImpaye'])\n",
    "\n",
    "y_train = train_data['FlagImpaye']\n",
    "y_test = test_data['FlagImpaye']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(3252741, 24)"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(3252741,)"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1394032, 24)"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1394032,)"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Examples:\n",
      "    Total: 3252741\n",
      "   Positive: 50682 (1.56% of total)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Imbalance Ratio du train set :\n",
    "\n",
    "neg, pos = np.bincount(y_train)\n",
    "total = neg + pos\n",
    "print('Examples:\\n    Total: {}\\n   Positive: {} ({:.2f}% of total)\\n'.format(\n",
    "    total, pos, 100 * pos / total))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "63.1794128092814"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "neg/pos"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Tomek Link (undersampling)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "#from imblearn.under_sampling import TomekLinks\n",
    "\n",
    "#tomek = TomekLinks(sampling_strategy='majority', n_jobs=8)\n",
    "\n",
    "#X_rus, y_rus = tomek.fit_resample(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "#predictor_names = X_train.columns\n",
    "\n",
    "#X_tl = pd.DataFrame(data=X_rus, columns=predictor_names)\n",
    "#y_tl = y_rus\n",
    "\n",
    "# Enregistrement du DataFrame dans un fichier CSV\n",
    "#X_tl.to_csv(\"X_tomek.csv\", index=False)\n",
    "#y_tl.to_csv(\"y_tomek.csv\", index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "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>IDAvisAutorisationCheque</th>\n",
       "      <th>Montant</th>\n",
       "      <th>VerifianceCPT1</th>\n",
       "      <th>VerifianceCPT2</th>\n",
       "      <th>VerifianceCPT3</th>\n",
       "      <th>D2CB</th>\n",
       "      <th>ScoringFP1</th>\n",
       "      <th>ScoringFP2</th>\n",
       "      <th>ScoringFP3</th>\n",
       "      <th>TauxImpNb_RB</th>\n",
       "      <th>...</th>\n",
       "      <th>DiffDateTr2</th>\n",
       "      <th>DiffDateTr3</th>\n",
       "      <th>CA3TRetMtt</th>\n",
       "      <th>CA3TR</th>\n",
       "      <th>Heure</th>\n",
       "      <th>Année</th>\n",
       "      <th>Mois</th>\n",
       "      <th>Jour</th>\n",
       "      <th>Minute</th>\n",
       "      <th>Seconde</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>78643044.0</td>\n",
       "      <td>20.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>551.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>37.186668</td>\n",
       "      <td>...</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>20.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>7</td>\n",
       "      <td>2017</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>32</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>78643045.0</td>\n",
       "      <td>20.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>551.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>48.844716</td>\n",
       "      <td>...</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>28.61</td>\n",
       "      <td>8.61</td>\n",
       "      <td>7</td>\n",
       "      <td>2017</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>43</td>\n",
       "      <td>37</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>78643046.0</td>\n",
       "      <td>57.64</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>549.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>73.118280</td>\n",
       "      <td>...</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>57.64</td>\n",
       "      <td>0.00</td>\n",
       "      <td>7</td>\n",
       "      <td>2017</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>47</td>\n",
       "      <td>38</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>78643047.0</td>\n",
       "      <td>54.29</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>267.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>110.056926</td>\n",
       "      <td>...</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>54.29</td>\n",
       "      <td>0.00</td>\n",
       "      <td>7</td>\n",
       "      <td>2017</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>48</td>\n",
       "      <td>48</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>78643047.0</td>\n",
       "      <td>54.29</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>267.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>110.056926</td>\n",
       "      <td>...</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>54.29</td>\n",
       "      <td>0.00</td>\n",
       "      <td>7</td>\n",
       "      <td>2017</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>48</td>\n",
       "      <td>48</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 24 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   IDAvisAutorisationCheque  Montant  VerifianceCPT1  VerifianceCPT2  \\\n",
       "0                78643044.0    20.00             0.0             0.0   \n",
       "1                78643045.0    20.00             0.0             0.0   \n",
       "2                78643046.0    57.64             0.0             0.0   \n",
       "3                78643047.0    54.29             1.0             1.0   \n",
       "4                78643047.0    54.29             1.0             1.0   \n",
       "\n",
       "   VerifianceCPT3   D2CB  ScoringFP1  ScoringFP2  ScoringFP3  TauxImpNb_RB  \\\n",
       "0             0.0  551.0         0.0         0.0         0.0     37.186668   \n",
       "1             0.0  551.0         0.0         0.0         0.0     48.844716   \n",
       "2             0.0  549.0         0.0         0.0         0.0     73.118280   \n",
       "3             1.0  267.0         0.0         0.0         0.0    110.056926   \n",
       "4             1.0  267.0         0.0         0.0         0.0    110.056926   \n",
       "\n",
       "   ...  DiffDateTr2  DiffDateTr3  CA3TRetMtt  CA3TR  Heure  Année  Mois  Jour  \\\n",
       "0  ...          4.0          4.0       20.00   0.00      7   2017     2     1   \n",
       "1  ...          4.0          4.0       28.61   8.61      7   2017     2     1   \n",
       "2  ...          4.0          4.0       57.64   0.00      7   2017     2     1   \n",
       "3  ...          4.0          4.0       54.29   0.00      7   2017     2     1   \n",
       "4  ...          4.0          4.0       54.29   0.00      7   2017     2     1   \n",
       "\n",
       "   Minute  Seconde  \n",
       "0      32       14  \n",
       "1      43       37  \n",
       "2      47       38  \n",
       "3      48       48  \n",
       "4      48       48  \n",
       "\n",
       "[5 rows x 24 columns]"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_tomek = pd.read_csv('/Users/annabellenarsama/Desktop/FouillesDonneesMassives/X_tomek.csv')\n",
    "X_tomek.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "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>FlagImpaye</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   FlagImpaye\n",
       "0       False\n",
       "1       False\n",
       "2       False\n",
       "3        True\n",
       "4       False"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_tomek = pd.read_csv('/Users/annabellenarsama/Desktop/FouillesDonneesMassives/y_tomek.csv')\n",
    "y_tomek.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(3251078, 24)"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_tomek.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(3251078, 1)"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_tomek.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Examples:\n",
      "    Total: 3251078\n",
      "   Positive: 50682 (1.56% of total)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Imbalance Ratio du Tomek Link :\n",
    "\n",
    "neg, pos = np.bincount(y_tomek['FlagImpaye'])\n",
    "total = neg + pos\n",
    "print('Examples:\\n    Total: {}\\n   Positive: {} ({:.2f}% of total)\\n'.format(\n",
    "    total, pos, 100 * pos / total))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Chargement des données train ré-échantillonnées par SMOTEEN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = pd.read_csv('/Users/annabellenarsama/Desktop/FouillesDonneesMassives/X.csv', sep=',')\n",
    "y = pd.read_csv('/Users/annabellenarsama/Desktop/FouillesDonneesMassives/y.csv', sep=',')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "X.drop(columns=['Unnamed: 0'], inplace=True)\n",
    "y.drop(columns=['Unnamed: 0'], inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 3794420 entries, 0 to 3794419\n",
      "Data columns (total 24 columns):\n",
      " #   Column                    Dtype  \n",
      "---  ------                    -----  \n",
      " 0   IDAvisAutorisationCheque  float64\n",
      " 1   Montant                   float64\n",
      " 2   VerifianceCPT1            float64\n",
      " 3   VerifianceCPT2            float64\n",
      " 4   VerifianceCPT3            float64\n",
      " 5   D2CB                      float64\n",
      " 6   ScoringFP1                float64\n",
      " 7   ScoringFP2                float64\n",
      " 8   ScoringFP3                float64\n",
      " 9   TauxImpNb_RB              float64\n",
      " 10  TauxImpNB_CPM             float64\n",
      " 11  EcartNumCheq              float64\n",
      " 12  NbrMagasin3J              float64\n",
      " 13  DiffDateTr1               float64\n",
      " 14  DiffDateTr2               float64\n",
      " 15  DiffDateTr3               float64\n",
      " 16  CA3TRetMtt                float64\n",
      " 17  CA3TR                     float64\n",
      " 18  Heure                     int64  \n",
      " 19  Année                     int64  \n",
      " 20  Mois                      int64  \n",
      " 21  Jour                      int64  \n",
      " 22  Minute                    int64  \n",
      " 23  Seconde                   int64  \n",
      "dtypes: float64(18), int64(6)\n",
      "memory usage: 694.8 MB\n"
     ]
    }
   ],
   "source": [
    "X.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 3794420 entries, 0 to 3794419\n",
      "Data columns (total 1 columns):\n",
      " #   Column      Dtype\n",
      "---  ------      -----\n",
      " 0   FlagImpaye  bool \n",
      "dtypes: bool(1)\n",
      "memory usage: 3.6 MB\n"
     ]
    }
   ],
   "source": [
    "y.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dimensions de X : (3794420, 24)\n",
      "Noms des variables de X : \n",
      " Index(['IDAvisAutorisationCheque', 'Montant', 'VerifianceCPT1',\n",
      "       'VerifianceCPT2', 'VerifianceCPT3', 'D2CB', 'ScoringFP1', 'ScoringFP2',\n",
      "       'ScoringFP3', 'TauxImpNb_RB', 'TauxImpNB_CPM', 'EcartNumCheq',\n",
      "       'NbrMagasin3J', 'DiffDateTr1', 'DiffDateTr2', 'DiffDateTr3',\n",
      "       'CA3TRetMtt', 'CA3TR', 'Heure', 'Année', 'Mois', 'Jour', 'Minute',\n",
      "       'Seconde'],\n",
      "      dtype='object') \n",
      "\n",
      "\n",
      "Dimensions de y : (3794420, 1)\n",
      "Noms des variables de y : \n",
      " Index(['FlagImpaye'], dtype='object') \n"
     ]
    }
   ],
   "source": [
    "print(f\"Dimensions de X : {X.shape}\")\n",
    "print(f\"Noms des variables de X : \\n {X.columns} \")\n",
    "\n",
    "print(\"\\n\")\n",
    "\n",
    "print(f\"Dimensions de y : {y.shape}\")\n",
    "print(f\"Noms des variables de y : \\n {y.columns} \")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Examples:\n",
      "    Total: 3794420\n",
      "   Positive: 21925 (0.58% of total)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "neg, pos = np.bincount(y['FlagImpaye'])\n",
    "total = neg + pos\n",
    "print('Examples:\\n    Total: {}\\n   Positive: {} ({:.2f}% of total)\\n'.format(\n",
    "    total, pos, 100 * pos / total))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# K-Means"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/annabellenarsama/anaconda3/envs/bigdataenv/lib/python3.11/site-packages/sklearn/cluster/_kmeans.py:1416: FutureWarning: The default value of `n_init` will change from 10 to 'auto' in 1.4. Set the value of `n_init` explicitly to suppress the warning\n",
      "  super()._check_params_vs_input(X, default_n_init=10)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-2 {color: black;}#sk-container-id-2 pre{padding: 0;}#sk-container-id-2 div.sk-toggleable {background-color: white;}#sk-container-id-2 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-2 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-2 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-2 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-2 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-2 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-2 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-2 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-2 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-2 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-2 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-2 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-2 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-2 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-2 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-2 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-2 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-2 div.sk-item {position: relative;z-index: 1;}#sk-container-id-2 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-2 div.sk-item::before, #sk-container-id-2 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-2 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-2 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-2 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-2 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-2 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-2 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-2 div.sk-label-container {text-align: center;}#sk-container-id-2 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-2 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-2\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>KMeans(n_clusters=2, random_state=1)</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-2\" type=\"checkbox\" checked><label for=\"sk-estimator-id-2\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">KMeans</label><div class=\"sk-toggleable__content\"><pre>KMeans(n_clusters=2, random_state=1)</pre></div></div></div></div></div>"
      ],
      "text/plain": [
       "KMeans(n_clusters=2, random_state=1)"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.cluster import KMeans\n",
    "\n",
    "# Instanciation de l'algorithme :\n",
    "kmeans = KMeans(n_clusters=2, random_state=1)\n",
    "\n",
    "# Fit du modèle sur les données rééchantillonnées :\n",
    "kmeans.fit(X_tomek, y_tomek)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Centroïdes de chaque exemple : \n",
      " [[7.92740779e+07 5.84940266e+01 2.35443238e-01 2.50560623e-01\n",
      "  2.75855087e-01 3.58179406e+02 1.61660103e+00 3.13205984e+00\n",
      "  9.76480648e-02 3.86652718e+01 3.84445858e+01 9.82945064e+03\n",
      "  1.03912755e+00 3.81793035e+00 4.11874252e+00 4.15314213e+00\n",
      "  6.56583118e+01 7.16428513e+00 1.39178170e+01 2.01700000e+03\n",
      "  2.55157033e+00 1.53370731e+01 2.94907070e+01 2.95049328e+01]\n",
      " [8.05514080e+07 6.00595895e+01 2.95838615e-01 3.15998046e-01\n",
      "  3.49942426e-01 3.25744074e+02 2.03353109e+00 2.30295178e+00\n",
      "  1.24389114e-01 3.89213792e+01 3.84397053e+01 8.61779714e+03\n",
      "  1.03693481e+00 3.86129865e+00 4.15656107e+00 4.18950884e+00\n",
      "  6.72614297e+01 7.20184012e+00 1.40411047e+01 2.01700000e+03\n",
      "  4.53827456e+00 1.53448903e+01 2.94977671e+01 2.95047211e+01]] \n",
      "\n",
      "Étiquetage des exemples : [0 0 0 ... 1 1 1]\n"
     ]
    }
   ],
   "source": [
    "# Paramètres du modèle :\n",
    "centers = kmeans.cluster_centers_\n",
    "labels = kmeans.labels_\n",
    "\n",
    "# Affichage des paramètres du modèle :\n",
    "print(\"Centroïdes de chaque exemple : \\n\", centers, \"\\n\")\n",
    "print(\"Étiquetage des exemples :\", labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Prédiction :\n",
    "y_pred_kmeans = kmeans.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "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>col_0</th>\n",
       "      <th>1</th>\n",
       "      <th>All</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FlagImpaye</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>False</th>\n",
       "      <td>1382092</td>\n",
       "      <td>1382092</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>True</th>\n",
       "      <td>11940</td>\n",
       "      <td>11940</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>All</th>\n",
       "      <td>1394032</td>\n",
       "      <td>1394032</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "col_0             1      All\n",
       "FlagImpaye                  \n",
       "False       1382092  1382092\n",
       "True          11940    11940\n",
       "All         1394032  1394032"
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Évaluation :\n",
    "from sklearn.metrics import f1_score, precision_score, recall_score, classification_report, confusion_matrix\n",
    "\n",
    "# Matrice de confusion :\n",
    "conf_m = pd.crosstab(y_test, y_pred_kmeans, margins=True)\n",
    "conf_m\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[      0, 1382092],\n",
       "       [      0,   11940]])"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mc = confusion_matrix(y_test, y_pred_kmeans)\n",
    "mc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/annabellenarsama/anaconda3/envs/bigdataenv/lib/python3.11/site-packages/sklearn/metrics/_classification.py:1471: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/Users/annabellenarsama/anaconda3/envs/bigdataenv/lib/python3.11/site-packages/sklearn/metrics/_classification.py:1471: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "       False       0.00      0.00      0.00   1382092\n",
      "        True       0.01      1.00      0.02     11940\n",
      "\n",
      "    accuracy                           0.01   1394032\n",
      "   macro avg       0.00      0.50      0.01   1394032\n",
      "weighted avg       0.00      0.01      0.00   1394032\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/annabellenarsama/anaconda3/envs/bigdataenv/lib/python3.11/site-packages/sklearn/metrics/_classification.py:1471: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    }
   ],
   "source": [
    "# Rapport complet des métriques du modèle (se concentre sur la classe majoritaire):\n",
    "metriques = classification_report(y_test, y_pred_kmeans)\n",
    "print(metriques)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Métriques du modèle sur la classe minoritaire"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "La F1-mesure avec un weighted average est de : 0.00014547529092249242.\n"
     ]
    }
   ],
   "source": [
    "# Binary :\n",
    "#f1_bin = f1_score(y_test, y_pred, average='binary', pos_label=1)\n",
    "#print(f\"La F1-mesure avec un binary average est de : {f1_bin}.\")\n",
    "\n",
    "# Weighted (par défaut) :\n",
    "f1_weigh = f1_score(y_test, y_pred_kmeans, average='weighted', pos_label=1)\n",
    "print(f\"La F1-mesure avec un weighted average est de : {f1_weigh}.\")\n",
    "\n",
    "# Macro :\n",
    "#f1_mac = f1_score(y_test, y_pred, average='macro', pos_label=1)\n",
    "#print(f\"La F1-mesure avec un macro average est de : {f1_mac}.\")\n",
    "\n",
    "# Micro :\n",
    "#f1_mic = f1_score(y_test, y_pred, average='micro', pos_label=1)\n",
    "#print(f\"La F1-mesure avec un micro average est de : {f1_mic}.\")\n",
    "\n",
    "# None (f1 score pour chaque classe comme le ferait le binaire):\n",
    "#f1_none = f1_score(y_test, y_pred, average=None)\n",
    "#print(f\"La F1-mesure pour chaque classe est de : {f1_none}.\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Régression logistique"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/annabellenarsama/anaconda3/envs/bigdataenv/lib/python3.11/site-packages/sklearn/utils/validation.py:1183: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
      "  y = column_or_1d(y, warn=True)\n",
      "/Users/annabellenarsama/anaconda3/envs/bigdataenv/lib/python3.11/site-packages/sklearn/metrics/_classification.py:1471: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Precision : [0.99143492 0.        ]\n",
      "Recall : [1. 0.]\n",
      "F1 Score : [0.99569904 0.        ]\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn import metrics\n",
    "\n",
    "# Création du modèle :\n",
    "reglog = LogisticRegression()\n",
    "\n",
    "# Entraînement :\n",
    "reglog.fit(X_tomek, y_tomek)\n",
    "\n",
    "# Prédictions :\n",
    "y_pred_reg = reglog.predict(X_test)\n",
    "\n",
    "# Évaluation :\n",
    "precision = metrics.precision_score(y_test, y_pred_reg, average=None, pos_label=1)\n",
    "recall = metrics.recall_score(y_test, y_pred_reg, average=None, pos_label=1)\n",
    "f1_score = metrics.f1_score(y_test, y_pred_reg,average=None, pos_label=1)\n",
    "\n",
    "# Affichage des performances :\n",
    "print(f\"Precision : {precision}\")\n",
    "print(f\"Recall : {recall}\")\n",
    "print(f\"F1 Score : {f1_score}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Auto-Encodeur"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "25400/25400 [==============================] - 10s 401us/step - loss: -702941560576147456.0000\n",
      "Epoch 2/10\n",
      "25400/25400 [==============================] - 10s 390us/step - loss: -4307110302417158144.0000\n",
      "Epoch 3/10\n",
      "25400/25400 [==============================] - 10s 392us/step - loss: -11081184948470677504.0000\n",
      "Epoch 4/10\n",
      "25400/25400 [==============================] - 10s 390us/step - loss: -20998264160056770560.0000\n",
      "Epoch 5/10\n",
      "25400/25400 [==============================] - 15s 585us/step - loss: -34003606551763353600.0000\n",
      "Epoch 6/10\n",
      "25400/25400 [==============================] - 11s 417us/step - loss: -50129901703795834880.0000\n",
      "Epoch 7/10\n",
      "25400/25400 [==============================] - 9s 360us/step - loss: -69386009480850833408.0000\n",
      "Epoch 8/10\n",
      "25400/25400 [==============================] - 9s 355us/step - loss: -91771503272416772096.0000\n",
      "Epoch 9/10\n",
      "25400/25400 [==============================] - 9s 374us/step - loss: -117287077969842405376.0000\n",
      "Epoch 10/10\n",
      "25400/25400 [==============================] - 9s 361us/step - loss: -145932284972383600640.0000\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/annabellenarsama/anaconda3/envs/bigdataenv/lib/python3.11/site-packages/keras/src/engine/training.py:3103: UserWarning: You are saving your model as an HDF5 file via `model.save()`. This file format is considered legacy. We recommend using instead the native Keras format, e.g. `model.save('my_model.keras')`.\n",
      "  saving_api.save_model(\n"
     ]
    }
   ],
   "source": [
    "from keras.layers import Input, Dense\n",
    "from keras.models import Model\n",
    "import tensorflow as tf\n",
    "\n",
    "# Dimension des données d'entrée :\n",
    "input_dim = X_tomek.shape[1]  # nombre de caractéristiques dans nos données\n",
    "\n",
    "# Architecture de l'auto-encodeur :\n",
    "input_layer = Input(shape=(input_dim))\n",
    "encoded = Dense(16, activation='relu')(input_layer)  # couche d'encodage\n",
    "decoded = Dense(input_dim, activation='sigmoid')(encoded)  # couche de décodage\n",
    "\n",
    "# Créer le modèle :\n",
    "autoencoder = Model(input_layer, decoded)\n",
    "\n",
    "# Compiler le modèle :\n",
    "autoencoder.compile(optimizer='adam', loss='binary_crossentropy')\n",
    "\n",
    "# Entraîner le modèle :\n",
    "with tf.device('/device:GPU:0'): # utiliser le GPU pour les calculs\n",
    "    history = autoencoder.fit(X_tomek, X_tomek, epochs=10, batch_size=128, shuffle=True)\n",
    "\n",
    "# Visualisation de l'évolution de la perte :\n",
    "plt.plot(history.history['loss'])\n",
    "plt.xlabel('Époch')\n",
    "plt.ylabel('Perte')\n",
    "plt.show()\n",
    "\n",
    "# Sauvegarde du modèle :\n",
    "autoencoder.save('autoencoder_model.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "43564/43564 [==============================] - 9s 208us/step\n"
     ]
    }
   ],
   "source": [
    "# Prédiction :\n",
    "\n",
    "# Charger le modèle entraîné :\n",
    "loaded_autoencoder = tf.keras.models.load_model('autoencoder_model.h5')\n",
    "\n",
    "# Faire des prédictions sur l'ensemble de test :\n",
    "y_pred_nn = loaded_autoencoder.predict(X_test)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "F1-score: 0.016984691018028807\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import f1_score\n",
    "import numpy as np\n",
    "\n",
    "# Calculer la perte de reconstruction (MSE)\n",
    "mse = np.mean(np.power(X_test - y_pred_nn, 2), axis=1)\n",
    "\n",
    "# Supposer un seuil pour définir les anomalies\n",
    "seuil = 0.1  # Vous devez ajuster ce seuil en fonction de votre problème\n",
    "\n",
    "# Binariser les prédictions en fonction du seuil\n",
    "anomalies = (mse > seuil).astype(int)\n",
    "\n",
    "# Calculer le F1-score\n",
    "f1 = f1_score(y_test, anomalies)\n",
    "\n",
    "print(\"F1-score:\", f1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [],
   "source": [
    "def EtudeGainPerte(X_test, y_test, y_pred):\n",
    "\n",
    "    poids = 0.05\n",
    "\n",
    "    tab = []\n",
    "\n",
    "#Passage des y en liste\n",
    "    y_test=list(y_test)\n",
    "    y_pred=list(y_pred)\n",
    "\n",
    "    for i in range(len(X_test)):\n",
    "        TN=0\n",
    "        FP=0\n",
    "        FN=0\n",
    "        TP=0\n",
    "\n",
    "        M = X_test.iloc[i]['Montant']\n",
    "\n",
    "        if (y_test[i]==0 & y_pred[i]==0):\n",
    "            TN = 0.05 * M\n",
    "        elif(y_test[i]==0 & y_pred[i]==1):\n",
    "            FP = 0.70 * M * 0.05\n",
    "        elif(y_test[i]==1 & y_pred[i]==1):\n",
    "            TP = poids * M \n",
    "        else:\n",
    "            if M <= 20:\n",
    "                FN = 0\n",
    "            elif M <= 50:\n",
    "                FN = 0.2 * M\n",
    "            elif M <= 100:\n",
    "                FN = 0.3 * M\n",
    "            elif M <= 200:\n",
    "                FN = 0.5 * M\n",
    "            else:\n",
    "                FN = 0.8 * M\n",
    "        rt = TN + FP - FN + TP\n",
    "\n",
    "\n",
    "        tab.append(rt)\n",
    "    return tab"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "4343895.025502005"
      ]
     },
     "execution_count": 94,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "loss_kmeans = EtudeGainPerte(X_test, y_test, y_pred_kmeans)\n",
    "sum(loss_kmeans)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3659124.154501696"
      ]
     },
     "execution_count": 92,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "loss_reg = EtudeGainPerte(X_test, y_test, y_pred_reg)\n",
    "sum(loss_reg)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "venv",
   "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": "3.11.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}