Projet_AD_Ybeggazene.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "8a2760f2",
   "metadata": {},
   "source": [
    "# Projet Analyse de Données L3 IM2D - Université Paris Dauphine\n",
    "\n",
    "**Réalisé par : Zakaria YBEGGAZENE**\n",
    "\n",
    "## **Thème : Création d'un classifier des crimes commis à Chicago**"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "32f27c96",
   "metadata": {},
   "source": [
    "## Introduction :\n",
    "\n",
    "Au début, on avait en tête de travailler sur des thématiques avancées comme la détection d'objets en utilisant des réseaux de neuronnes convolutionnels avec du transfer learning. Mais après une bonne réflexion, on s'est rendu compte que ça ne correspondait pas trop aux notions vues dans l'UE PAD, et qu'il serait plus judicieux de choisir un thème dans lequel on pourrait appliquer des modèles linéaires en choisissant notamment un jeu de données numériques plutôt que du dégital.\n",
    "\n",
    "C'est la raison pour laquelle on a basculé environ une semaine avant le rendu du projet vers la thématique : **classification des crimes commis dans la ville de Chicago**."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "155ecc1a",
   "metadata": {},
   "source": [
    "On commence par importer quelques bibliothèques de base"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "b96d3c9a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Populating the interactive namespace from numpy and matplotlib\n"
     ]
    }
   ],
   "source": [
    "%pylab inline\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a76dc05c",
   "metadata": {},
   "source": [
    "## 1. Importation des données :\n",
    "\n",
    "Le dataset des crimes commis à Chicago depuis 2001 jusqu'à aujourd'hui (moins les 7 derniers jours) est disponible sur le site : https://data.cityofchicago.org/Public-Safety/Crimes-2001-to-Present/ijzp-q8t2 avec des descriptions détaillées de tous les attributs et la possibilité d'effectuer des recherches et des filtres directement sur le site.\n",
    "\n",
    "On a décidé de télécharger que les crimes depuis 01/01/2015 jusqu'au 31/12/2020 sous format csv. Le fichier .csv est de taille 360Mo et contient plus de 1.5 million d'exemples.\n",
    "\n",
    "On charge ce dataset dans un `DataFrame pandas` en utilisant la fonction `read_csv()` (veillez à bien mettre le fichier csv dans le même dossier que ce notebook avec le même nom que ci-dessous)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "388208ee",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv(\"chicago_crimes_2015_2020.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "811499f0",
   "metadata": {},
   "source": [
    "On affiche les 5 premières lignes du dataframe :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "cbb2abfc",
   "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>ID</th>\n",
       "      <th>Case Number</th>\n",
       "      <th>Date</th>\n",
       "      <th>Block</th>\n",
       "      <th>IUCR</th>\n",
       "      <th>Primary Type</th>\n",
       "      <th>Description</th>\n",
       "      <th>Location Description</th>\n",
       "      <th>Arrest</th>\n",
       "      <th>Domestic</th>\n",
       "      <th>...</th>\n",
       "      <th>Ward</th>\n",
       "      <th>Community Area</th>\n",
       "      <th>FBI Code</th>\n",
       "      <th>X Coordinate</th>\n",
       "      <th>Y Coordinate</th>\n",
       "      <th>Year</th>\n",
       "      <th>Updated On</th>\n",
       "      <th>Latitude</th>\n",
       "      <th>Longitude</th>\n",
       "      <th>Location</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>11227503</td>\n",
       "      <td>JB146383</td>\n",
       "      <td>01/01/2015 12:01:00 AM</td>\n",
       "      <td>061XX S KILBOURN AVE</td>\n",
       "      <td>1751</td>\n",
       "      <td>OFFENSE INVOLVING CHILDREN</td>\n",
       "      <td>CRIM SEX ABUSE BY FAM MEMBER</td>\n",
       "      <td>RESIDENCE</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>...</td>\n",
       "      <td>13.0</td>\n",
       "      <td>65.0</td>\n",
       "      <td>17</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2015</td>\n",
       "      <td>04/12/2019 04:00:15 PM</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>11042867</td>\n",
       "      <td>JA374764</td>\n",
       "      <td>01/01/2015 12:01:00 AM</td>\n",
       "      <td>086XX S MANISTEE AVE</td>\n",
       "      <td>0266</td>\n",
       "      <td>CRIM SEXUAL ASSAULT</td>\n",
       "      <td>PREDATORY</td>\n",
       "      <td>RESIDENCE</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>7.0</td>\n",
       "      <td>46.0</td>\n",
       "      <td>02</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2015</td>\n",
       "      <td>08/05/2017 03:50:08 PM</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>11042931</td>\n",
       "      <td>JA374193</td>\n",
       "      <td>01/01/2015 12:01:00 AM</td>\n",
       "      <td>069XX S STEWART AVE</td>\n",
       "      <td>1562</td>\n",
       "      <td>SEX OFFENSE</td>\n",
       "      <td>AGG CRIMINAL SEXUAL ABUSE</td>\n",
       "      <td>RESIDENCE</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>6.0</td>\n",
       "      <td>68.0</td>\n",
       "      <td>17</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2015</td>\n",
       "      <td>08/05/2017 03:50:08 PM</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>11042927</td>\n",
       "      <td>JA374449</td>\n",
       "      <td>01/01/2015 12:01:00 AM</td>\n",
       "      <td>022XX N AVERS AVE</td>\n",
       "      <td>1750</td>\n",
       "      <td>OFFENSE INVOLVING CHILDREN</td>\n",
       "      <td>CHILD ABUSE</td>\n",
       "      <td>RESIDENCE</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>...</td>\n",
       "      <td>30.0</td>\n",
       "      <td>22.0</td>\n",
       "      <td>20</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2015</td>\n",
       "      <td>08/25/2017 03:48:04 PM</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>11042939</td>\n",
       "      <td>JA376124</td>\n",
       "      <td>01/01/2015 12:01:00 AM</td>\n",
       "      <td>049XX S LECLAIRE AVE</td>\n",
       "      <td>0281</td>\n",
       "      <td>CRIM SEXUAL ASSAULT</td>\n",
       "      <td>NON-AGGRAVATED</td>\n",
       "      <td>RESIDENCE</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>23.0</td>\n",
       "      <td>56.0</td>\n",
       "      <td>02</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2015</td>\n",
       "      <td>08/05/2017 03:50:08 PM</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 22 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         ID Case Number                    Date                 Block  IUCR  \\\n",
       "0  11227503    JB146383  01/01/2015 12:01:00 AM  061XX S KILBOURN AVE  1751   \n",
       "1  11042867    JA374764  01/01/2015 12:01:00 AM  086XX S MANISTEE AVE  0266   \n",
       "2  11042931    JA374193  01/01/2015 12:01:00 AM   069XX S STEWART AVE  1562   \n",
       "3  11042927    JA374449  01/01/2015 12:01:00 AM     022XX N AVERS AVE  1750   \n",
       "4  11042939    JA376124  01/01/2015 12:01:00 AM  049XX S LECLAIRE AVE  0281   \n",
       "\n",
       "                 Primary Type                   Description  \\\n",
       "0  OFFENSE INVOLVING CHILDREN  CRIM SEX ABUSE BY FAM MEMBER   \n",
       "1         CRIM SEXUAL ASSAULT                     PREDATORY   \n",
       "2                 SEX OFFENSE     AGG CRIMINAL SEXUAL ABUSE   \n",
       "3  OFFENSE INVOLVING CHILDREN                   CHILD ABUSE   \n",
       "4         CRIM SEXUAL ASSAULT                NON-AGGRAVATED   \n",
       "\n",
       "  Location Description  Arrest  Domestic  ...  Ward  Community Area  FBI Code  \\\n",
       "0            RESIDENCE   False      True  ...  13.0            65.0        17   \n",
       "1            RESIDENCE   False     False  ...   7.0            46.0        02   \n",
       "2            RESIDENCE   False     False  ...   6.0            68.0        17   \n",
       "3            RESIDENCE   False      True  ...  30.0            22.0        20   \n",
       "4            RESIDENCE   False     False  ...  23.0            56.0        02   \n",
       "\n",
       "   X Coordinate Y Coordinate  Year              Updated On  Latitude  \\\n",
       "0           NaN          NaN  2015  04/12/2019 04:00:15 PM       NaN   \n",
       "1           NaN          NaN  2015  08/05/2017 03:50:08 PM       NaN   \n",
       "2           NaN          NaN  2015  08/05/2017 03:50:08 PM       NaN   \n",
       "3           NaN          NaN  2015  08/25/2017 03:48:04 PM       NaN   \n",
       "4           NaN          NaN  2015  08/05/2017 03:50:08 PM       NaN   \n",
       "\n",
       "  Longitude  Location  \n",
       "0       NaN       NaN  \n",
       "1       NaN       NaN  \n",
       "2       NaN       NaN  \n",
       "3       NaN       NaN  \n",
       "4       NaN       NaN  \n",
       "\n",
       "[5 rows x 22 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "46423c2c",
   "metadata": {},
   "source": [
    "On appelle `info()` pour avoir plus de détails sur le dataframe notamment le nombre de lignes (1584347) et le nombre de colonnes (22) ainsi que le type de données dans chaque colonne et le nombre de valeurs non nulles"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "d86c9de9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 1584347 entries, 0 to 1584346\n",
      "Data columns (total 22 columns):\n",
      " #   Column                Non-Null Count    Dtype  \n",
      "---  ------                --------------    -----  \n",
      " 0   ID                    1584347 non-null  int64  \n",
      " 1   Case Number           1584347 non-null  object \n",
      " 2   Date                  1584347 non-null  object \n",
      " 3   Block                 1584347 non-null  object \n",
      " 4   IUCR                  1584347 non-null  object \n",
      " 5   Primary Type          1584347 non-null  object \n",
      " 6   Description           1584347 non-null  object \n",
      " 7   Location Description  1577736 non-null  object \n",
      " 8   Arrest                1584347 non-null  bool   \n",
      " 9   Domestic              1584347 non-null  bool   \n",
      " 10  Beat                  1584347 non-null  int64  \n",
      " 11  District              1584346 non-null  float64\n",
      " 12  Ward                  1584314 non-null  float64\n",
      " 13  Community Area        1584346 non-null  float64\n",
      " 14  FBI Code              1584347 non-null  object \n",
      " 15  X Coordinate          1562871 non-null  float64\n",
      " 16  Y Coordinate          1562871 non-null  float64\n",
      " 17  Year                  1584347 non-null  int64  \n",
      " 18  Updated On            1584347 non-null  object \n",
      " 19  Latitude              1562871 non-null  float64\n",
      " 20  Longitude             1562871 non-null  float64\n",
      " 21  Location              1562871 non-null  object \n",
      "dtypes: bool(2), float64(7), int64(3), object(10)\n",
      "memory usage: 244.8+ MB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "48ed9d6c",
   "metadata": {},
   "source": [
    "## 2. Exploration et Preprocessing :\n",
    "\n",
    "Nous allons maintenant explorer nos données et effectuer le preprocessing nécessaire pour préparer le dataset pour l'étape de la modélisation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "fa1f109e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "001XX N STATE ST         5168\n",
       "0000X W TERMINAL ST      2682\n",
       "008XX N MICHIGAN AVE     2639\n",
       "0000X N STATE ST         2045\n",
       "0000X S STATE ST         1783\n",
       "                         ... \n",
       "016XX W Morse Ave           1\n",
       "0000X W CHICAGO RIVER       1\n",
       "013XX W Grand ave           1\n",
       "007XX W Montrose Ave        1\n",
       "027XX W 109TH ST            1\n",
       "Name: Block, Length: 35290, dtype: int64"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['Block'].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "da6b8e76",
   "metadata": {},
   "source": [
    "Un bon nombre de colonnes comme les `ID` des crimes et leur `Case Number` ne sont pas exploitables puisque ce sont des données séquentielles qui servent juste à représenter les crimes.\n",
    "\n",
    "Le `IUCR` qui est le Illinois Uniform Crime Reporting code est lié à l'attribut `Primary Type`, ces codes et leurs significations se trouvent via ce [lien](https://data.cityofchicago.org/d/c7ck-438e). Le `FBI Code` est aussi un code attribué par le FBI au crime :\n",
    ">Indicates the crime classification as outlined in the FBI's National Incident-Based Reporting System (NIBRS). See the Chicago Police Department listing of these classifications at http://gis.chicagopolice.org/clearmap_crime_sums/crime_types.html.\n",
    "\n",
    "La `Description` détaillée du crime n'est pas vraiment exploitatble dans notre contexte aussi, pareil pour la date de la mise à jour des informations concernant le crime `Updated On`.\n",
    "\n",
    "L'attribut `Location` est un tuple de deux éléments qui sont la `Latitude` et la `Longitude` qu'on possède déjà comme deux attributs séparés, on peut alors le supprimer.\n",
    "\n",
    "`X Coordinate` et `Y Coordinate` représentent l'endroit où le crime s'est passé (tout comme la `Latitude` et la `Longitude`) selon le *State Plane Illinois East NAD 1983 projection*. On peut s'en passer alors puisqu'ils circulent la même information que d'autres attributs.\n",
    "\n",
    "Enfin, l'attribut `Block` contient l'adresse partielle de l'endroit où le crime s'est déroulé sous forme de chaîne de caractères. L'attribut circule beaucoup de valeurs différentes (même en gardant que les noms des rues, chose qu'on a essayé puis effacé pour ne pas rajouter beaucoup de cellules au notebook), il est donc difficile de les encoder en quelque chose d'exploitatble. Par conséquent, on a décidé de supprimer cet attribut.\n",
    "\n",
    "Dans la cellule suivante, on supprime les attributs cités ci-dessus. Mais on supprime aussi les lignes qui comptent des valeus nulles avec la fonction `dropna()` puisque la proportion de ces lignes n'est pas grande par rapport à la taille du jeu de données en entier.\n",
    "\n",
    "On affiche alors le DataFrame obtenu."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "b6e222d0",
   "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>Date</th>\n",
       "      <th>Primary Type</th>\n",
       "      <th>Location Description</th>\n",
       "      <th>Arrest</th>\n",
       "      <th>Domestic</th>\n",
       "      <th>Beat</th>\n",
       "      <th>District</th>\n",
       "      <th>Ward</th>\n",
       "      <th>Community Area</th>\n",
       "      <th>Year</th>\n",
       "      <th>Latitude</th>\n",
       "      <th>Longitude</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>01/01/2015 12:01:00 AM</td>\n",
       "      <td>OBSCENITY</td>\n",
       "      <td>RESIDENCE</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>221</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>2015</td>\n",
       "      <td>41.811212</td>\n",
       "      <td>-87.615663</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>01/01/2015 12:01:00 AM</td>\n",
       "      <td>DECEPTIVE PRACTICE</td>\n",
       "      <td>ATHLETIC CLUB</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>1814</td>\n",
       "      <td>18.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>2015</td>\n",
       "      <td>41.911151</td>\n",
       "      <td>-87.635982</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>01/01/2015 12:01:00 AM</td>\n",
       "      <td>CRIM SEXUAL ASSAULT</td>\n",
       "      <td>RESIDENCE</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>2532</td>\n",
       "      <td>25.0</td>\n",
       "      <td>37.0</td>\n",
       "      <td>25.0</td>\n",
       "      <td>2015</td>\n",
       "      <td>41.902203</td>\n",
       "      <td>-87.760110</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>01/01/2015 12:01:00 AM</td>\n",
       "      <td>CRIM SEXUAL ASSAULT</td>\n",
       "      <td>RESIDENCE</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>925</td>\n",
       "      <td>9.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>37.0</td>\n",
       "      <td>2015</td>\n",
       "      <td>41.814283</td>\n",
       "      <td>-87.633512</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>01/01/2015 12:01:00 AM</td>\n",
       "      <td>CRIM SEXUAL ASSAULT</td>\n",
       "      <td>RESIDENCE</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>1512</td>\n",
       "      <td>15.0</td>\n",
       "      <td>29.0</td>\n",
       "      <td>25.0</td>\n",
       "      <td>2015</td>\n",
       "      <td>41.888997</td>\n",
       "      <td>-87.770622</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1584342</th>\n",
       "      <td>03/29/2021 11:35:00 PM</td>\n",
       "      <td>WEAPONS VIOLATION</td>\n",
       "      <td>STREET</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>2533</td>\n",
       "      <td>25.0</td>\n",
       "      <td>37.0</td>\n",
       "      <td>25.0</td>\n",
       "      <td>2021</td>\n",
       "      <td>41.912013</td>\n",
       "      <td>-87.745929</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1584343</th>\n",
       "      <td>03/29/2021 11:45:00 PM</td>\n",
       "      <td>BATTERY</td>\n",
       "      <td>STREET</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>822</td>\n",
       "      <td>8.0</td>\n",
       "      <td>14.0</td>\n",
       "      <td>63.0</td>\n",
       "      <td>2021</td>\n",
       "      <td>41.795251</td>\n",
       "      <td>-87.709642</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1584344</th>\n",
       "      <td>03/29/2021 11:45:00 PM</td>\n",
       "      <td>BATTERY</td>\n",
       "      <td>SIDEWALK</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>1831</td>\n",
       "      <td>18.0</td>\n",
       "      <td>42.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>2021</td>\n",
       "      <td>41.889761</td>\n",
       "      <td>-87.627999</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1584345</th>\n",
       "      <td>03/29/2021 11:45:00 PM</td>\n",
       "      <td>BATTERY</td>\n",
       "      <td>APARTMENT</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>1931</td>\n",
       "      <td>19.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>2021</td>\n",
       "      <td>41.929029</td>\n",
       "      <td>-87.672011</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1584346</th>\n",
       "      <td>03/29/2021 11:55:00 PM</td>\n",
       "      <td>ASSAULT</td>\n",
       "      <td>STREET</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>712</td>\n",
       "      <td>7.0</td>\n",
       "      <td>16.0</td>\n",
       "      <td>68.0</td>\n",
       "      <td>2021</td>\n",
       "      <td>41.784366</td>\n",
       "      <td>-87.650975</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1558281 rows × 12 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                           Date         Primary Type Location Description  \\\n",
       "13       01/01/2015 12:01:00 AM            OBSCENITY            RESIDENCE   \n",
       "15       01/01/2015 12:01:00 AM   DECEPTIVE PRACTICE        ATHLETIC CLUB   \n",
       "16       01/01/2015 12:01:00 AM  CRIM SEXUAL ASSAULT            RESIDENCE   \n",
       "17       01/01/2015 12:01:00 AM  CRIM SEXUAL ASSAULT            RESIDENCE   \n",
       "18       01/01/2015 12:01:00 AM  CRIM SEXUAL ASSAULT            RESIDENCE   \n",
       "...                         ...                  ...                  ...   \n",
       "1584342  03/29/2021 11:35:00 PM    WEAPONS VIOLATION               STREET   \n",
       "1584343  03/29/2021 11:45:00 PM              BATTERY               STREET   \n",
       "1584344  03/29/2021 11:45:00 PM              BATTERY             SIDEWALK   \n",
       "1584345  03/29/2021 11:45:00 PM              BATTERY            APARTMENT   \n",
       "1584346  03/29/2021 11:55:00 PM              ASSAULT               STREET   \n",
       "\n",
       "         Arrest  Domestic  Beat  District  Ward  Community Area  Year  \\\n",
       "13        False     False   221       2.0   3.0            38.0  2015   \n",
       "15        False     False  1814      18.0   2.0             7.0  2015   \n",
       "16        False     False  2532      25.0  37.0            25.0  2015   \n",
       "17        False     False   925       9.0   3.0            37.0  2015   \n",
       "18        False     False  1512      15.0  29.0            25.0  2015   \n",
       "...         ...       ...   ...       ...   ...             ...   ...   \n",
       "1584342    True     False  2533      25.0  37.0            25.0  2021   \n",
       "1584343    True      True   822       8.0  14.0            63.0  2021   \n",
       "1584344    True     False  1831      18.0  42.0             8.0  2021   \n",
       "1584345   False      True  1931      19.0  32.0             7.0  2021   \n",
       "1584346   False     False   712       7.0  16.0            68.0  2021   \n",
       "\n",
       "          Latitude  Longitude  \n",
       "13       41.811212 -87.615663  \n",
       "15       41.911151 -87.635982  \n",
       "16       41.902203 -87.760110  \n",
       "17       41.814283 -87.633512  \n",
       "18       41.888997 -87.770622  \n",
       "...            ...        ...  \n",
       "1584342  41.912013 -87.745929  \n",
       "1584343  41.795251 -87.709642  \n",
       "1584344  41.889761 -87.627999  \n",
       "1584345  41.929029 -87.672011  \n",
       "1584346  41.784366 -87.650975  \n",
       "\n",
       "[1558281 rows x 12 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.drop(columns = ['ID', 'Case Number', 'Block', 'IUCR', 'Description', 'FBI Code',\n",
    "                   'X Coordinate', 'Y Coordinate', 'Updated On', 'Location'], inplace = True)\n",
    "df.dropna(inplace = True)\n",
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2b7f2051",
   "metadata": {},
   "source": [
    "On voit ci-dessous qu'on n'a plus d'enregistrements qui contiennent des valeurs nulles et que nous avons réduit le nombre de colonnes jusqu'à 12."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "babb4874",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 1558281 entries, 13 to 1584346\n",
      "Data columns (total 12 columns):\n",
      " #   Column                Non-Null Count    Dtype  \n",
      "---  ------                --------------    -----  \n",
      " 0   Date                  1558281 non-null  object \n",
      " 1   Primary Type          1558281 non-null  object \n",
      " 2   Location Description  1558281 non-null  object \n",
      " 3   Arrest                1558281 non-null  bool   \n",
      " 4   Domestic              1558281 non-null  bool   \n",
      " 5   Beat                  1558281 non-null  int64  \n",
      " 6   District              1558281 non-null  float64\n",
      " 7   Ward                  1558281 non-null  float64\n",
      " 8   Community Area        1558281 non-null  float64\n",
      " 9   Year                  1558281 non-null  int64  \n",
      " 10  Latitude              1558281 non-null  float64\n",
      " 11  Longitude             1558281 non-null  float64\n",
      "dtypes: bool(2), float64(5), int64(2), object(3)\n",
      "memory usage: 133.7+ MB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "79fa0294",
   "metadata": {},
   "source": [
    "L'attribut `Primary Type` qui représente le type primaire du crime est l'attribut qu'on aimerait prédire avec nos modèles.\n",
    "\n",
    "Cependant, on remarque ci-dessous avec la fonction `value_counts()` que cet attribut sous format `String` prend un bon nombre de valeurs et que le nombre d'exemples dont on dispose pour chaque type n'est pas du tout équilibré puisque les vols (THEFT) sont beaucoup plus commun que d'autres crimes comme la Traite des êtres humains (HUMAN TRAFFICKING). Et si on veut garder tous ces types comme des classes pour nos modèles ça nous fera beaucoup de classes et avec des nombres d'exemples (d'entraînement et de test) totalement déséquilibrés."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "15ccaa46",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "THEFT                                354048\n",
       "BATTERY                              296870\n",
       "CRIMINAL DAMAGE                      172321\n",
       "ASSAULT                              118037\n",
       "DECEPTIVE PRACTICE                   103402\n",
       "OTHER OFFENSE                        100414\n",
       "NARCOTICS                             82613\n",
       "BURGLARY                              71536\n",
       "MOTOR VEHICLE THEFT                   63563\n",
       "ROBBERY                               60653\n",
       "CRIMINAL TRESPASS                     38033\n",
       "WEAPONS VIOLATION                     33531\n",
       "OFFENSE INVOLVING CHILDREN            13163\n",
       "PUBLIC PEACE VIOLATION                 9769\n",
       "INTERFERENCE WITH PUBLIC OFFICER       6911\n",
       "CRIM SEXUAL ASSAULT                    6508\n",
       "SEX OFFENSE                            6188\n",
       "PROSTITUTION                           4545\n",
       "HOMICIDE                               3986\n",
       "ARSON                                  2845\n",
       "CRIMINAL SEXUAL ASSAULT                2339\n",
       "LIQUOR LAW VIOLATION                   1374\n",
       "STALKING                               1185\n",
       "KIDNAPPING                             1063\n",
       "GAMBLING                               1059\n",
       "INTIMIDATION                            913\n",
       "CONCEALED CARRY LICENSE VIOLATION       691\n",
       "OBSCENITY                               385\n",
       "NON-CRIMINAL                            137\n",
       "PUBLIC INDECENCY                         70\n",
       "HUMAN TRAFFICKING                        63\n",
       "OTHER NARCOTIC VIOLATION                 34\n",
       "NON - CRIMINAL                           25\n",
       "NON-CRIMINAL (SUBJECT SPECIFIED)          6\n",
       "RITUALISM                                 1\n",
       "Name: Primary Type, dtype: int64"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['Primary Type'].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e5743ac4",
   "metadata": {},
   "source": [
    "On affiche les **histogrammes** des autres attributs numériques pour voir les tendances qu'il prennent :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "ab39c403",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x864 with 9 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df.hist(figsize=(12,12))\n",
    "print('-')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ec30cc61",
   "metadata": {},
   "source": [
    "Ici on affiche un **barplot** du nombre d'exemples qu'on a pour chaque type de crime (`Primary Type`). On remarque que c'est assez déséquilibré."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "3e0f5307",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(15, 5))\n",
    "plt.bar(df['Primary Type'].value_counts().index.tolist(),df['Primary Type'].value_counts())\n",
    "plt.xticks(rotation=60, ha=\"right\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e07bf665",
   "metadata": {},
   "source": [
    "Etant données les difficultés rencontrées pour utiliser `Primary Type` directement pour nos modèles, on a décidé de faire du *Hand Engineering* d'un nouvel attribut qui serait plus équilibré et plus représentatif des crimes.\n",
    "\n",
    "Après quelques recherches, notamment dans le [*Criminal Code*](https://www.ilga.gov/legislation/ilcs/ilcs3.asp?ChapterID=53&ActID=1876) du *Illinois Compiled Statutes (ILCS)*, on a trouvé qu'on pouvait mettre ces types de crimes dans 3 catégories majeures :\n",
    "* Offenses directed against the person (ODATP)\n",
    "* Offenses directed against property (ODAP)\n",
    "* Offenses affecting public health, safety and decency/affecting governmental functions (OAPHSD)\n",
    "\n",
    "Nous avons mis ces types de crimes dans chacune des catégories comme suit :\n",
    "\n",
    "**Offenses directed against the person** : Battery, Assault, Offense involving children, Sex Offense, Crim Sexual Assault, Homicide, Prostitution, Stalking, Kidnapping, Intimidation, Human Trafficking.\n",
    "\n",
    "**Offenses directed against property** : Theft, Deceptive Practice, Burglary, Motor Vehicle theft, Robbery, Arson.\n",
    "\n",
    "**Offenses affecting public health, safety and decency/Gov func** : Narcotics, Weapons Violation, Public Peace Violation, Interference with Public Officer, Liquor Law Violation, Concealed Carry License Violation, Gambling, Obscenity, Public Indecency, Other Narcotic Violation.\n",
    "\n",
    "Les crimes de type : **Other Offense, Criminal Damage et Criminal Trespass** sont supprimés puisqu'ils représentent un mix des catégories.\n",
    "Les lignes avec comme type : **Non-Criminal, Non-Criminal (Subject Specified) et Non - Criminal** ne sont pas pris en compte.\n",
    "Enfin, la seule ligne avec comme type de crime **Ritualism** sera supprimée aussi par manque d'exemples.\n",
    "\n",
    "Tout cela est réalisé dans les prochaines cellules :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "17bc0605",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df[~df['Primary Type'].isin([\"OTHER OFFENSE\", \"CRIMINAL DAMAGE\", \"CRIMINAL TRESPASS\", \"NON-CRIMINAL\",\n",
    "                                 \"NON-CRIMINAL (SUBJECT SPECIFIED)\", \"NON - CRIMINAL\", \"RITUALISM\"])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "c311f870",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of crimes in the category 'Offenses directed against the person' : 452521\n",
      "Number of crimes in the category 'Offenses directed against property' : 656047\n",
      "Number of crimes in the category 'Offenses affecting public health, safety and decency/affecting governmental functions' : 136437\n"
     ]
    }
   ],
   "source": [
    "print(\"Number of crimes in the category 'Offenses directed against the person' :\",\n",
    "      len(df[df['Primary Type'].isin([\"BATTERY\", \"ASSAULT\", \"OFFENSE INVOLVING CHILDREN\", \"SEX OFFENSE\", \"HOMICIDE\",\n",
    "                                      \"CRIM SEXUAL ASSAULT\", \"PROSTITUTION\", \"STALKING\", \"KIDNAPPING\",\n",
    "                                      \"INTIMIDATION\", \"HUMAN TRAFFICKING\"])]))\n",
    "print(\"Number of crimes in the category 'Offenses directed against property' :\",\n",
    "      len(df[df['Primary Type'].isin([\"THEFT\", \"DECEPTIVE PRACTICE\", \"BURGLARY\", \"MOTOR VEHICLE THEFT\", \"ROBBERY\",\n",
    "                                      \"ARSON\"])]))\n",
    "print(\"Number of crimes in the category 'Offenses affecting public health, safety and decency/affecting governmental functions' :\",\n",
    "      len(df[df['Primary Type'].isin([\"NARCOTICS\", \"WEAPONS VIOLATION\", \"PUBLIC PEACE VIOLATION\", \"LIQUOR LAW VIOLATION\",\n",
    "                                      \"CONCEALED CARRY LICENSE VIOLATION\", \"INTERFERENCE WITH PUBLIC OFFICER\",\n",
    "                                      \"GAMBLING\", \"OBSCENITY\", \"PUBLIC INDECENCY\", \"OTHER NARCOTIC VIOLATION\"])]))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "41011601",
   "metadata": {},
   "source": [
    "On crée la colonne `Category` que nos modèles essayeront de prédire."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "e414d9cf",
   "metadata": {},
   "outputs": [],
   "source": [
    "df.insert(1, \"Category\", -1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "776db928",
   "metadata": {},
   "source": [
    "On associe 0, 1 ou 2 aux catégories ODATP, ODAP ou OAPHSD respectivement.\n",
    "\n",
    "*Le Warning 'SettingWithCopyWarning' s'est affiché car pandas a confondu l'utilisation de `loc` dans la cellule ci-dessous avec d'autres utilisation qui peuvent être problématiques ce qui n'est pas le cas ici. Pour plus de détails : lire la documentation*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "7b160a2e",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\lenovo\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\pandas\\core\\indexing.py:1720: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  self._setitem_single_column(loc, value, pi)\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>Date</th>\n",
       "      <th>Category</th>\n",
       "      <th>Primary Type</th>\n",
       "      <th>Location Description</th>\n",
       "      <th>Arrest</th>\n",
       "      <th>Domestic</th>\n",
       "      <th>Beat</th>\n",
       "      <th>District</th>\n",
       "      <th>Ward</th>\n",
       "      <th>Community Area</th>\n",
       "      <th>Year</th>\n",
       "      <th>Latitude</th>\n",
       "      <th>Longitude</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>01/01/2015 12:01:00 AM</td>\n",
       "      <td>2</td>\n",
       "      <td>OBSCENITY</td>\n",
       "      <td>RESIDENCE</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>221</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>2015</td>\n",
       "      <td>41.811212</td>\n",
       "      <td>-87.615663</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>01/01/2015 12:01:00 AM</td>\n",
       "      <td>1</td>\n",
       "      <td>DECEPTIVE PRACTICE</td>\n",
       "      <td>ATHLETIC CLUB</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>1814</td>\n",
       "      <td>18.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>2015</td>\n",
       "      <td>41.911151</td>\n",
       "      <td>-87.635982</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>01/01/2015 12:01:00 AM</td>\n",
       "      <td>0</td>\n",
       "      <td>CRIM SEXUAL ASSAULT</td>\n",
       "      <td>RESIDENCE</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>2532</td>\n",
       "      <td>25.0</td>\n",
       "      <td>37.0</td>\n",
       "      <td>25.0</td>\n",
       "      <td>2015</td>\n",
       "      <td>41.902203</td>\n",
       "      <td>-87.760110</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>01/01/2015 12:01:00 AM</td>\n",
       "      <td>0</td>\n",
       "      <td>CRIM SEXUAL ASSAULT</td>\n",
       "      <td>RESIDENCE</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>925</td>\n",
       "      <td>9.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>37.0</td>\n",
       "      <td>2015</td>\n",
       "      <td>41.814283</td>\n",
       "      <td>-87.633512</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>01/01/2015 12:01:00 AM</td>\n",
       "      <td>0</td>\n",
       "      <td>CRIM SEXUAL ASSAULT</td>\n",
       "      <td>RESIDENCE</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>1512</td>\n",
       "      <td>15.0</td>\n",
       "      <td>29.0</td>\n",
       "      <td>25.0</td>\n",
       "      <td>2015</td>\n",
       "      <td>41.888997</td>\n",
       "      <td>-87.770622</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1584342</th>\n",
       "      <td>03/29/2021 11:35:00 PM</td>\n",
       "      <td>2</td>\n",
       "      <td>WEAPONS VIOLATION</td>\n",
       "      <td>STREET</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>2533</td>\n",
       "      <td>25.0</td>\n",
       "      <td>37.0</td>\n",
       "      <td>25.0</td>\n",
       "      <td>2021</td>\n",
       "      <td>41.912013</td>\n",
       "      <td>-87.745929</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1584343</th>\n",
       "      <td>03/29/2021 11:45:00 PM</td>\n",
       "      <td>0</td>\n",
       "      <td>BATTERY</td>\n",
       "      <td>STREET</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>822</td>\n",
       "      <td>8.0</td>\n",
       "      <td>14.0</td>\n",
       "      <td>63.0</td>\n",
       "      <td>2021</td>\n",
       "      <td>41.795251</td>\n",
       "      <td>-87.709642</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1584344</th>\n",
       "      <td>03/29/2021 11:45:00 PM</td>\n",
       "      <td>0</td>\n",
       "      <td>BATTERY</td>\n",
       "      <td>SIDEWALK</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>1831</td>\n",
       "      <td>18.0</td>\n",
       "      <td>42.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>2021</td>\n",
       "      <td>41.889761</td>\n",
       "      <td>-87.627999</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1584345</th>\n",
       "      <td>03/29/2021 11:45:00 PM</td>\n",
       "      <td>0</td>\n",
       "      <td>BATTERY</td>\n",
       "      <td>APARTMENT</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>1931</td>\n",
       "      <td>19.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>2021</td>\n",
       "      <td>41.929029</td>\n",
       "      <td>-87.672011</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1584346</th>\n",
       "      <td>03/29/2021 11:55:00 PM</td>\n",
       "      <td>0</td>\n",
       "      <td>ASSAULT</td>\n",
       "      <td>STREET</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>712</td>\n",
       "      <td>7.0</td>\n",
       "      <td>16.0</td>\n",
       "      <td>68.0</td>\n",
       "      <td>2021</td>\n",
       "      <td>41.784366</td>\n",
       "      <td>-87.650975</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1247344 rows × 13 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                           Date  Category         Primary Type  \\\n",
       "13       01/01/2015 12:01:00 AM         2            OBSCENITY   \n",
       "15       01/01/2015 12:01:00 AM         1   DECEPTIVE PRACTICE   \n",
       "16       01/01/2015 12:01:00 AM         0  CRIM SEXUAL ASSAULT   \n",
       "17       01/01/2015 12:01:00 AM         0  CRIM SEXUAL ASSAULT   \n",
       "18       01/01/2015 12:01:00 AM         0  CRIM SEXUAL ASSAULT   \n",
       "...                         ...       ...                  ...   \n",
       "1584342  03/29/2021 11:35:00 PM         2    WEAPONS VIOLATION   \n",
       "1584343  03/29/2021 11:45:00 PM         0              BATTERY   \n",
       "1584344  03/29/2021 11:45:00 PM         0              BATTERY   \n",
       "1584345  03/29/2021 11:45:00 PM         0              BATTERY   \n",
       "1584346  03/29/2021 11:55:00 PM         0              ASSAULT   \n",
       "\n",
       "        Location Description  Arrest  Domestic  Beat  District  Ward  \\\n",
       "13                 RESIDENCE   False     False   221       2.0   3.0   \n",
       "15             ATHLETIC CLUB   False     False  1814      18.0   2.0   \n",
       "16                 RESIDENCE   False     False  2532      25.0  37.0   \n",
       "17                 RESIDENCE   False     False   925       9.0   3.0   \n",
       "18                 RESIDENCE   False     False  1512      15.0  29.0   \n",
       "...                      ...     ...       ...   ...       ...   ...   \n",
       "1584342               STREET    True     False  2533      25.0  37.0   \n",
       "1584343               STREET    True      True   822       8.0  14.0   \n",
       "1584344             SIDEWALK    True     False  1831      18.0  42.0   \n",
       "1584345            APARTMENT   False      True  1931      19.0  32.0   \n",
       "1584346               STREET   False     False   712       7.0  16.0   \n",
       "\n",
       "         Community Area  Year   Latitude  Longitude  \n",
       "13                 38.0  2015  41.811212 -87.615663  \n",
       "15                  7.0  2015  41.911151 -87.635982  \n",
       "16                 25.0  2015  41.902203 -87.760110  \n",
       "17                 37.0  2015  41.814283 -87.633512  \n",
       "18                 25.0  2015  41.888997 -87.770622  \n",
       "...                 ...   ...        ...        ...  \n",
       "1584342            25.0  2021  41.912013 -87.745929  \n",
       "1584343            63.0  2021  41.795251 -87.709642  \n",
       "1584344             8.0  2021  41.889761 -87.627999  \n",
       "1584345             7.0  2021  41.929029 -87.672011  \n",
       "1584346            68.0  2021  41.784366 -87.650975  \n",
       "\n",
       "[1247344 rows x 13 columns]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.loc[df['Primary Type'].isin([\"BATTERY\", \"ASSAULT\", \"OFFENSE INVOLVING CHILDREN\", \"SEX OFFENSE\", \"HOMICIDE\",\n",
    "                \"CRIM SEXUAL ASSAULT\", \"CRIMINAL SEXUAL ASSAULT\", \"PROSTITUTION\", \"STALKING\", \"KIDNAPPING\",\n",
    "                \"INTIMIDATION\", \"HUMAN TRAFFICKING\"]), 'Category'] = 0\n",
    "df.loc[df['Primary Type'].isin([\"THEFT\", \"DECEPTIVE PRACTICE\", \"BURGLARY\", \"MOTOR VEHICLE THEFT\", \"ROBBERY\",\n",
    "                \"ARSON\"]), 'Category'] = 1\n",
    "df.loc[df['Primary Type'].isin([\"NARCOTICS\", \"WEAPONS VIOLATION\", \"PUBLIC PEACE VIOLATION\", \"LIQUOR LAW VIOLATION\",\n",
    "                \"CONCEALED CARRY LICENSE VIOLATION\", \"INTERFERENCE WITH PUBLIC OFFICER\", \"GAMBLING\", \"OBSCENITY\",\n",
    "                \"PUBLIC INDECENCY\", \"OTHER NARCOTIC VIOLATION\"]), 'Category'] = 2\n",
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dbf0678e",
   "metadata": {},
   "source": [
    "On supprime la colonne `Primary Type` et on vérifie bien qu'on a 3 catégories de crimes dorénavant."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "d9b62f79",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1    656047\n",
       "0    454860\n",
       "2    136437\n",
       "Name: Category, dtype: int64"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = df.drop('Primary Type', axis=1)\n",
    "df['Category'].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e9ce28be",
   "metadata": {},
   "source": [
    "Bien que les exemples pour ces catégories ne soit pas très équilibré (voir **barplot** ci-dessous), ça reste plutôt acceptable et meilleur que ce qu'on avait avant. En plus, cela peut être tout à fait normal que notre modèle soit biaisé vers la catégorie de crimes qui se produient le plus."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "37f4bf57",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(15, 5))\n",
    "plt.bar(df['Category'].value_counts().index.tolist(),df['Category'].value_counts())\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3a3fbdec",
   "metadata": {},
   "source": [
    "On découvre l'attribut `Location Description` sous forme de `String` qui indique dans quel endroit s'est passé le crime.\n",
    "\n",
    "On trouve 193 *Locations* différentes dans notre dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "f64a5b25",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "STREET                                                   273469\n",
      "RESIDENCE                                                189416\n",
      "APARTMENT                                                169633\n",
      "SIDEWALK                                                 122477\n",
      "OTHER                                                     42118\n",
      "SMALL RETAIL STORE                                        34233\n",
      "RESTAURANT                                                30136\n",
      "PARKING LOT/GARAGE(NON.RESID.)                            30033\n",
      "ALLEY                                                     27792\n",
      "DEPARTMENT STORE                                          24198\n",
      "VEHICLE NON-COMMERCIAL                                    22031\n",
      "RESIDENCE PORCH/HALLWAY                                   20222\n",
      "RESIDENTIAL YARD (FRONT/BACK)                             19885\n",
      "GROCERY FOOD STORE                                        17495\n",
      "GAS STATION                                               16366\n",
      "SCHOOL, PUBLIC, BUILDING                                  16273\n",
      "RESIDENCE-GARAGE                                          15204\n",
      "BAR OR TAVERN                                              9804\n",
      "PARK PROPERTY                                              9594\n",
      "CONVENIENCE STORE                                          9394\n",
      "CTA TRAIN                                                  8374\n",
      "COMMERCIAL / BUSINESS OFFICE                               7325\n",
      "DRUG STORE                                                 6259\n",
      "HOTEL/MOTEL                                                5954\n",
      "BANK                                                       5395\n",
      "HOSPITAL BUILDING/GROUNDS                                  5159\n",
      "CTA BUS                                                    4919\n",
      "SCHOOL, PUBLIC, GROUNDS                                    4615\n",
      "PARKING LOT / GARAGE (NON RESIDENTIAL)                     4599\n",
      "OTHER (SPECIFY)                                            4462\n",
      "CTA PLATFORM                                               4389\n",
      "CTA STATION                                                4367\n",
      "NURSING HOME/RETIREMENT HOME                               3756\n",
      "RESIDENCE - PORCH / HALLWAY                                3430\n",
      "VACANT LOT/LAND                                            3401\n",
      "CHA APARTMENT                                              3394\n",
      "CTA BUS STOP                                               2966\n",
      "RESIDENCE - YARD (FRONT / BACK)                            2924\n",
      "DRIVEWAY - RESIDENTIAL                                     2918\n",
      "POLICE FACILITY/VEH PARKING LOT                            2873\n",
      "ATHLETIC CLUB                                              2592\n",
      "ATM (AUTOMATIC TELLER MACHINE)                             2441\n",
      "SCHOOL, PRIVATE, BUILDING                                  2436\n",
      "TAVERN/LIQUOR STORE                                        2232\n",
      "RESIDENCE - GARAGE                                         2155\n",
      "CHA PARKING LOT/GROUNDS                                    2087\n",
      "CURRENCY EXCHANGE                                          1995\n",
      "GOVERNMENT BUILDING/PROPERTY                               1928\n",
      "TAXICAB                                                    1886\n",
      "CHURCH/SYNAGOGUE/PLACE OF WORSHIP                          1835\n",
      "CONSTRUCTION SITE                                          1597\n",
      "ABANDONED BUILDING                                         1551\n",
      "WAREHOUSE                                                  1352\n",
      "AIRPORT TERMINAL UPPER LEVEL - SECURE AREA                 1106\n",
      "SPORTS ARENA/STADIUM                                       1090\n",
      "VEHICLE-COMMERCIAL                                         1027\n",
      "MEDICAL/DENTAL OFFICE                                      1010\n",
      "LIBRARY                                                     993\n",
      "BARBERSHOP                                                  992\n",
      "HOSPITAL BUILDING / GROUNDS                                 944\n",
      "HOTEL / MOTEL                                               936\n",
      "DAY CARE CENTER                                             758\n",
      "SCHOOL, PRIVATE, GROUNDS                                    683\n",
      "CTA GARAGE / OTHER PROPERTY                                 659\n",
      "POLICE FACILITY / VEHICLE PARKING LOT                       643\n",
      "AUTO / BOAT / RV DEALERSHIP                                 643\n",
      "NURSING / RETIREMENT HOME                                   638\n",
      "CAR WASH                                                    625\n",
      "CHA HALLWAY/STAIRWELL/ELEVATOR                              623\n",
      "OTHER COMMERCIAL TRANSPORTATION                             610\n",
      "OTHER RAILROAD PROP / TRAIN DEPOT                           606\n",
      "COLLEGE/UNIVERSITY GROUNDS                                  596\n",
      "AIRPORT BUILDING NON-TERMINAL - NON-SECURE AREA             579\n",
      "AIRPORT TERMINAL LOWER LEVEL - NON-SECURE AREA              559\n",
      "MOVIE HOUSE/THEATER                                         515\n",
      "AIRPORT VENDING ESTABLISHMENT                               491\n",
      "APPLIANCE STORE                                             487\n",
      "VACANT LOT / LAND                                           483\n",
      "FACTORY/MANUFACTURING BUILDING                              478\n",
      "VEHICLE - OTHER RIDE SHARE SERVICE (E.G., UBER, LYFT)       446\n",
      "CLEANING STORE                                              425\n",
      "AIRPORT EXTERIOR - NON-SECURE AREA                          421\n",
      "AUTO                                                        414\n",
      "AIRCRAFT                                                    397\n",
      "AIRPORT BUILDING NON-TERMINAL - SECURE AREA                 387\n",
      "SCHOOL - PUBLIC BUILDING                                    384\n",
      "AIRPORT PARKING LOT                                         377\n",
      "TAVERN / LIQUOR STORE                                       360\n",
      "CHA PARKING LOT / GROUNDS                                   353\n",
      "AIRPORT TERMINAL LOWER LEVEL - SECURE AREA                  338\n",
      "AIRPORT TERMINAL UPPER LEVEL - NON-SECURE AREA              330\n",
      "LAKEFRONT/WATERFRONT/RIVERBANK                              325\n",
      "GOVERNMENT BUILDING / PROPERTY                              304\n",
      "VEHICLE - OTHER RIDE SERVICE                                304\n",
      "JAIL / LOCK-UP FACILITY                                     296\n",
      "PAWN SHOP                                                   277\n",
      "AIRPORT/AIRCRAFT                                            272\n",
      "POOL ROOM                                                   219\n",
      "MEDICAL / DENTAL OFFICE                                     217\n",
      "CHURCH / SYNAGOGUE / PLACE OF WORSHIP                       210\n",
      "SCHOOL - PUBLIC GROUNDS                                     207\n",
      "VEHICLE - DELIVERY TRUCK                                    202\n",
      "ANIMAL HOSPITAL                                             195\n",
      "FIRE STATION                                                188\n",
      "HOUSE                                                       182\n",
      "HIGHWAY/EXPRESSWAY                                          175\n",
      "FEDERAL BUILDING                                            173\n",
      "COLLEGE/UNIVERSITY RESIDENCE HALL                           162\n",
      "BOWLING ALLEY                                               152\n",
      "AIRPORT EXTERIOR - SECURE AREA                              147\n",
      "VEHICLE - OTHER RIDE SHARE SERVICE (LYFT, UBER, ETC.)       146\n",
      "CHA HALLWAY / STAIRWELL / ELEVATOR                          140\n",
      "COIN OPERATED MACHINE                                       136\n",
      "BOAT/WATERCRAFT                                             130\n",
      "VEHICLE - COMMERCIAL                                        129\n",
      "SCHOOL - PRIVATE BUILDING                                   122\n",
      "PORCH                                                       114\n",
      "YARD                                                        109\n",
      "BRIDGE                                                      101\n",
      "CREDIT UNION                                                 99\n",
      "FACTORY / MANUFACTURING BUILDING                             97\n",
      "OTHER RAILROAD PROPERTY / TRAIN DEPOT                        89\n",
      "PARKING LOT                                                  68\n",
      "CTA PARKING LOT / GARAGE / OTHER PROPERTY                    62\n",
      "CEMETARY                                                     54\n",
      "SAVINGS AND LOAN                                             54\n",
      "FOREST PRESERVE                                              52\n",
      "HIGHWAY / EXPRESSWAY                                         49\n",
      "SCHOOL - PRIVATE GROUNDS                                     48\n",
      "AIRPORT TERMINAL MEZZANINE - NON-SECURE AREA                 48\n",
      "SPORTS ARENA / STADIUM                                       44\n",
      "AIRPORT TRANSPORTATION SYSTEM (ATS)                          41\n",
      "CTA TRACKS - RIGHT OF WAY                                    41\n",
      "COLLEGE / UNIVERSITY - GROUNDS                               39\n",
      "VACANT LOT                                                   33\n",
      "LAKEFRONT / WATERFRONT / RIVERBANK                           31\n",
      "NEWSSTAND                                                    25\n",
      "BOAT / WATERCRAFT                                            21\n",
      "CHA PARKING LOT                                              20\n",
      "RETAIL STORE                                                 20\n",
      "HALLWAY                                                      18\n",
      "GARAGE                                                       16\n",
      "GAS STATION DRIVE/PROP.                                      14\n",
      "COLLEGE / UNIVERSITY - RESIDENCE HALL                        13\n",
      "MOVIE HOUSE / THEATER                                        13\n",
      "GANGWAY                                                      11\n",
      "HOTEL                                                         9\n",
      "VEHICLE-COMMERCIAL - TROLLEY BUS                              8\n",
      "STAIRWELL                                                     8\n",
      "HOSPITAL                                                      8\n",
      "VEHICLE-COMMERCIAL - ENTERTAINMENT/PARTY BUS                  8\n",
      "DRIVEWAY                                                      7\n",
      "VESTIBULE                                                     7\n",
      "TAVERN                                                        6\n",
      "VEHICLE - COMMERCIAL: ENTERTAINMENT / PARTY BUS               6\n",
      "BARBER SHOP/BEAUTY SALON                                      6\n",
      "CHA GROUNDS                                                   5\n",
      "SCHOOL YARD                                                   4\n",
      "CTA \"L\" PLATFORM                                              4\n",
      "OFFICE                                                        4\n",
      "NURSING HOME                                                  3\n",
      "LIQUOR STORE                                                  3\n",
      "KENNEL                                                        3\n",
      "HORSE STABLE                                                  3\n",
      "FARM                                                          3\n",
      "CTA PROPERTY                                                  3\n",
      "BASEMENT                                                      2\n",
      "CLUB                                                          2\n",
      "WOODED AREA                                                   2\n",
      "CHA HALLWAY                                                   2\n",
      "RIVER BANK                                                    2\n",
      "VEHICLE - COMMERCIAL: TROLLEY BUS                             2\n",
      "ELEVATOR                                                      2\n",
      "GARAGE/AUTO REPAIR                                            2\n",
      "CHURCH                                                        2\n",
      "RAILROAD PROPERTY                                             2\n",
      "LAKE                                                          1\n",
      "CTA \"L\" TRAIN                                                 1\n",
      "CHA PLAY LOT                                                  1\n",
      "CTA SUBWAY STATION                                            1\n",
      "YMCA                                                          1\n",
      "LAUNDRY ROOM                                                  1\n",
      "CHA ELEVATOR                                                  1\n",
      "CLEANERS/LAUNDROMAT                                           1\n",
      "GOVERNMENT BUILDING                                           1\n",
      "MOTEL                                                         1\n",
      "POOLROOM                                                      1\n",
      "EXPRESSWAY EMBANKMENT                                         1\n",
      "TRUCK                                                         1\n",
      "TRAILER                                                       1\n",
      "ROOMING HOUSE                                                 1\n",
      "DELIVERY TRUCK                                                1\n",
      "LAGOON                                                        1\n"
     ]
    }
   ],
   "source": [
    "#Il y a 193 Location possibles, listées ci-dessous\n",
    "print(df['Location Description'].value_counts().to_string())"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2b819835",
   "metadata": {},
   "source": [
    "On essaye de voir le nombre d'arrestation effectuées par catégorie :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "f0416666",
   "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>Arrest</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Category</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>91221</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>53552</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>122627</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          Arrest\n",
       "Category        \n",
       "0          91221\n",
       "1          53552\n",
       "2         122627"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[['Arrest', 'Category']].groupby('Category').sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "88393eee",
   "metadata": {},
   "source": [
    "La probabilité qu'une arrestation soit effectuée par catégorie :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "a0f14ef8",
   "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>Arrest</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Category</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.200547</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.081628</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.898781</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            Arrest\n",
       "Category          \n",
       "0         0.200547\n",
       "1         0.081628\n",
       "2         0.898781"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[['Arrest', 'Category']].groupby('Category').mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5884505d",
   "metadata": {},
   "source": [
    "Les différentes combinaisons pour `Beat`, `District`, `Ward` et `Community Area` :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "e805bb73",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Beat  District  Ward  Community Area\n",
       "1834  18.0      42.0  8.0               14667\n",
       "111   1.0       42.0  32.0              10711\n",
       "112   1.0       42.0  32.0              10271\n",
       "1831  18.0      42.0  8.0               10264\n",
       "421   4.0       7.0   43.0              10036\n",
       "                                        ...  \n",
       "1923  25.0      46.0  6.0                   1\n",
       "735   6.0       17.0  67.0                  1\n",
       "734   6.0       17.0  67.0                  1\n",
       "412   22.0      8.0   48.0                  1\n",
       "814   31.0      23.0  56.0                  1\n",
       "Length: 1119, dtype: int64"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[['Beat', 'District', 'Ward', 'Community Area']].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "c5e873c7",
   "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>Date</th>\n",
       "      <th>Category</th>\n",
       "      <th>Location Description</th>\n",
       "      <th>Arrest</th>\n",
       "      <th>Domestic</th>\n",
       "      <th>Beat</th>\n",
       "      <th>District</th>\n",
       "      <th>Ward</th>\n",
       "      <th>Community Area</th>\n",
       "      <th>Year</th>\n",
       "      <th>Latitude</th>\n",
       "      <th>Longitude</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>01/01/2015 12:01:00 AM</td>\n",
       "      <td>2</td>\n",
       "      <td>RESIDENCE</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>221</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>2015</td>\n",
       "      <td>41.811212</td>\n",
       "      <td>-87.615663</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>01/01/2015 12:01:00 AM</td>\n",
       "      <td>1</td>\n",
       "      <td>ATHLETIC CLUB</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>1814</td>\n",
       "      <td>18.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>2015</td>\n",
       "      <td>41.911151</td>\n",
       "      <td>-87.635982</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>01/01/2015 12:01:00 AM</td>\n",
       "      <td>0</td>\n",
       "      <td>RESIDENCE</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>2532</td>\n",
       "      <td>25.0</td>\n",
       "      <td>37.0</td>\n",
       "      <td>25.0</td>\n",
       "      <td>2015</td>\n",
       "      <td>41.902203</td>\n",
       "      <td>-87.760110</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>01/01/2015 12:01:00 AM</td>\n",
       "      <td>0</td>\n",
       "      <td>RESIDENCE</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>925</td>\n",
       "      <td>9.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>37.0</td>\n",
       "      <td>2015</td>\n",
       "      <td>41.814283</td>\n",
       "      <td>-87.633512</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>01/01/2015 12:01:00 AM</td>\n",
       "      <td>0</td>\n",
       "      <td>RESIDENCE</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>1512</td>\n",
       "      <td>15.0</td>\n",
       "      <td>29.0</td>\n",
       "      <td>25.0</td>\n",
       "      <td>2015</td>\n",
       "      <td>41.888997</td>\n",
       "      <td>-87.770622</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                      Date  Category Location Description  Arrest  Domestic  \\\n",
       "13  01/01/2015 12:01:00 AM         2            RESIDENCE   False     False   \n",
       "15  01/01/2015 12:01:00 AM         1        ATHLETIC CLUB   False     False   \n",
       "16  01/01/2015 12:01:00 AM         0            RESIDENCE   False     False   \n",
       "17  01/01/2015 12:01:00 AM         0            RESIDENCE   False     False   \n",
       "18  01/01/2015 12:01:00 AM         0            RESIDENCE   False     False   \n",
       "\n",
       "    Beat  District  Ward  Community Area  Year   Latitude  Longitude  \n",
       "13   221       2.0   3.0            38.0  2015  41.811212 -87.615663  \n",
       "15  1814      18.0   2.0             7.0  2015  41.911151 -87.635982  \n",
       "16  2532      25.0  37.0            25.0  2015  41.902203 -87.760110  \n",
       "17   925       9.0   3.0            37.0  2015  41.814283 -87.633512  \n",
       "18  1512      15.0  29.0            25.0  2015  41.888997 -87.770622  "
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "840dd46b",
   "metadata": {},
   "source": [
    "On aimerais ensuite nous occupé de l'attribut `Date` et de le rendre plus exploitable.\n",
    "\n",
    "On en extrait alors les attributs `Day`, `Month` et `Hour` qui peuvent jouer un rôle important dans la prédiction de la catégorie du crime qui aura lieu selon le temps."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "9a86bffb",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "df.insert(9, \"Day\", df['Date'].str.split('/', 2, expand=True)[1])\n",
    "df.loc[:,'Day'] = pd.to_numeric(df.loc[:,'Day'])\n",
    "df.insert(10, \"Month\", df['Date'].str.split('/', 1, expand=True)[0])\n",
    "df.loc[:,'Month'] = pd.to_numeric(df.loc[:,'Month'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "ea9e709f",
   "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>Date</th>\n",
       "      <th>Category</th>\n",
       "      <th>Location Description</th>\n",
       "      <th>Arrest</th>\n",
       "      <th>Domestic</th>\n",
       "      <th>Beat</th>\n",
       "      <th>District</th>\n",
       "      <th>Ward</th>\n",
       "      <th>Community Area</th>\n",
       "      <th>Hour</th>\n",
       "      <th>Day</th>\n",
       "      <th>Month</th>\n",
       "      <th>Year</th>\n",
       "      <th>Latitude</th>\n",
       "      <th>Longitude</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>01/01/2015 12:01:00 AM</td>\n",
       "      <td>2</td>\n",
       "      <td>RESIDENCE</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>221</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>12</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2015</td>\n",
       "      <td>41.811212</td>\n",
       "      <td>-87.615663</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>01/01/2015 12:01:00 AM</td>\n",
       "      <td>1</td>\n",
       "      <td>ATHLETIC CLUB</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>1814</td>\n",
       "      <td>18.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>12</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2015</td>\n",
       "      <td>41.911151</td>\n",
       "      <td>-87.635982</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>01/01/2015 12:01:00 AM</td>\n",
       "      <td>0</td>\n",
       "      <td>RESIDENCE</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>2532</td>\n",
       "      <td>25.0</td>\n",
       "      <td>37.0</td>\n",
       "      <td>25.0</td>\n",
       "      <td>12</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2015</td>\n",
       "      <td>41.902203</td>\n",
       "      <td>-87.760110</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>01/01/2015 12:01:00 AM</td>\n",
       "      <td>0</td>\n",
       "      <td>RESIDENCE</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>925</td>\n",
       "      <td>9.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>37.0</td>\n",
       "      <td>12</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2015</td>\n",
       "      <td>41.814283</td>\n",
       "      <td>-87.633512</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>01/01/2015 12:01:00 AM</td>\n",
       "      <td>0</td>\n",
       "      <td>RESIDENCE</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>1512</td>\n",
       "      <td>15.0</td>\n",
       "      <td>29.0</td>\n",
       "      <td>25.0</td>\n",
       "      <td>12</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2015</td>\n",
       "      <td>41.888997</td>\n",
       "      <td>-87.770622</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1584342</th>\n",
       "      <td>03/29/2021 11:35:00 PM</td>\n",
       "      <td>2</td>\n",
       "      <td>STREET</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>2533</td>\n",
       "      <td>25.0</td>\n",
       "      <td>37.0</td>\n",
       "      <td>25.0</td>\n",
       "      <td>23</td>\n",
       "      <td>29</td>\n",
       "      <td>3</td>\n",
       "      <td>2021</td>\n",
       "      <td>41.912013</td>\n",
       "      <td>-87.745929</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1584343</th>\n",
       "      <td>03/29/2021 11:45:00 PM</td>\n",
       "      <td>0</td>\n",
       "      <td>STREET</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>822</td>\n",
       "      <td>8.0</td>\n",
       "      <td>14.0</td>\n",
       "      <td>63.0</td>\n",
       "      <td>23</td>\n",
       "      <td>29</td>\n",
       "      <td>3</td>\n",
       "      <td>2021</td>\n",
       "      <td>41.795251</td>\n",
       "      <td>-87.709642</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1584344</th>\n",
       "      <td>03/29/2021 11:45:00 PM</td>\n",
       "      <td>0</td>\n",
       "      <td>SIDEWALK</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>1831</td>\n",
       "      <td>18.0</td>\n",
       "      <td>42.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>23</td>\n",
       "      <td>29</td>\n",
       "      <td>3</td>\n",
       "      <td>2021</td>\n",
       "      <td>41.889761</td>\n",
       "      <td>-87.627999</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1584345</th>\n",
       "      <td>03/29/2021 11:45:00 PM</td>\n",
       "      <td>0</td>\n",
       "      <td>APARTMENT</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>1931</td>\n",
       "      <td>19.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>23</td>\n",
       "      <td>29</td>\n",
       "      <td>3</td>\n",
       "      <td>2021</td>\n",
       "      <td>41.929029</td>\n",
       "      <td>-87.672011</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1584346</th>\n",
       "      <td>03/29/2021 11:55:00 PM</td>\n",
       "      <td>0</td>\n",
       "      <td>STREET</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>712</td>\n",
       "      <td>7.0</td>\n",
       "      <td>16.0</td>\n",
       "      <td>68.0</td>\n",
       "      <td>23</td>\n",
       "      <td>29</td>\n",
       "      <td>3</td>\n",
       "      <td>2021</td>\n",
       "      <td>41.784366</td>\n",
       "      <td>-87.650975</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1247344 rows × 15 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                           Date  Category Location Description  Arrest  \\\n",
       "13       01/01/2015 12:01:00 AM         2            RESIDENCE   False   \n",
       "15       01/01/2015 12:01:00 AM         1        ATHLETIC CLUB   False   \n",
       "16       01/01/2015 12:01:00 AM         0            RESIDENCE   False   \n",
       "17       01/01/2015 12:01:00 AM         0            RESIDENCE   False   \n",
       "18       01/01/2015 12:01:00 AM         0            RESIDENCE   False   \n",
       "...                         ...       ...                  ...     ...   \n",
       "1584342  03/29/2021 11:35:00 PM         2               STREET    True   \n",
       "1584343  03/29/2021 11:45:00 PM         0               STREET    True   \n",
       "1584344  03/29/2021 11:45:00 PM         0             SIDEWALK    True   \n",
       "1584345  03/29/2021 11:45:00 PM         0            APARTMENT   False   \n",
       "1584346  03/29/2021 11:55:00 PM         0               STREET   False   \n",
       "\n",
       "         Domestic  Beat  District  Ward  Community Area  Hour  Day  Month  \\\n",
       "13          False   221       2.0   3.0            38.0    12    1      1   \n",
       "15          False  1814      18.0   2.0             7.0    12    1      1   \n",
       "16          False  2532      25.0  37.0            25.0    12    1      1   \n",
       "17          False   925       9.0   3.0            37.0    12    1      1   \n",
       "18          False  1512      15.0  29.0            25.0    12    1      1   \n",
       "...           ...   ...       ...   ...             ...   ...  ...    ...   \n",
       "1584342     False  2533      25.0  37.0            25.0    23   29      3   \n",
       "1584343      True   822       8.0  14.0            63.0    23   29      3   \n",
       "1584344     False  1831      18.0  42.0             8.0    23   29      3   \n",
       "1584345      True  1931      19.0  32.0             7.0    23   29      3   \n",
       "1584346     False   712       7.0  16.0            68.0    23   29      3   \n",
       "\n",
       "         Year   Latitude  Longitude  \n",
       "13       2015  41.811212 -87.615663  \n",
       "15       2015  41.911151 -87.635982  \n",
       "16       2015  41.902203 -87.760110  \n",
       "17       2015  41.814283 -87.633512  \n",
       "18       2015  41.888997 -87.770622  \n",
       "...       ...        ...        ...  \n",
       "1584342  2021  41.912013 -87.745929  \n",
       "1584343  2021  41.795251 -87.709642  \n",
       "1584344  2021  41.889761 -87.627999  \n",
       "1584345  2021  41.929029 -87.672011  \n",
       "1584346  2021  41.784366 -87.650975  \n",
       "\n",
       "[1247344 rows x 15 columns]"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tmp = df['Date'].str.split(' ', 2, expand=True)\n",
    "tmp.insert(1, \"Hour\", pd.to_numeric(tmp[1].str.split(':', 1, expand=True)[0]))\n",
    "tmp.loc[tmp[2] == \"PM\", 'Hour'] += 12\n",
    "tmp.loc[tmp['Hour'] == 24, 'Hour'] = 0\n",
    "df.insert(9, \"Hour\", tmp['Hour'])\n",
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "74ceb699",
   "metadata": {},
   "source": [
    "On drop la colonne `Date`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "eb403477",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.drop('Date', axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c276696e",
   "metadata": {},
   "source": [
    "On transforme les colonne `Arrest` et `Domestic` booléennes en des entier (0 pour False et 1 pour True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "e74ce839",
   "metadata": {},
   "outputs": [],
   "source": [
    "df['Arrest'] = df['Arrest'].astype(int)\n",
    "df['Domestic'] = df['Domestic'].astype(int)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e6f73468",
   "metadata": {},
   "source": [
    "N'ayant pas cherché une meilleure manière pour exploiter l'attribut `Location Description`, on a décidé de simplement l'encoder avec un `OrdinalEncoder()`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "ce26eeb6",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Numeriser 'Location Description'\n",
    "from sklearn.preprocessing import OrdinalEncoder\n",
    "enc = OrdinalEncoder()\n",
    "df['Location Description'] = enc.fit_transform(df[['Location Description']])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7794dc62",
   "metadata": {},
   "source": [
    "On mélange notre jeu de données"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "603d2b6c",
   "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>Category</th>\n",
       "      <th>Location Description</th>\n",
       "      <th>Arrest</th>\n",
       "      <th>Domestic</th>\n",
       "      <th>Beat</th>\n",
       "      <th>District</th>\n",
       "      <th>Ward</th>\n",
       "      <th>Community Area</th>\n",
       "      <th>Hour</th>\n",
       "      <th>Day</th>\n",
       "      <th>Month</th>\n",
       "      <th>Year</th>\n",
       "      <th>Latitude</th>\n",
       "      <th>Longitude</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1057342</th>\n",
       "      <td>0</td>\n",
       "      <td>89.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>712</td>\n",
       "      <td>7.0</td>\n",
       "      <td>16.0</td>\n",
       "      <td>68.0</td>\n",
       "      <td>0</td>\n",
       "      <td>12</td>\n",
       "      <td>12</td>\n",
       "      <td>2018</td>\n",
       "      <td>41.787017</td>\n",
       "      <td>-87.645506</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1352020</th>\n",
       "      <td>0</td>\n",
       "      <td>24.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1925</td>\n",
       "      <td>19.0</td>\n",
       "      <td>46.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>18</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>2020</td>\n",
       "      <td>41.949954</td>\n",
       "      <td>-87.649572</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>922407</th>\n",
       "      <td>1</td>\n",
       "      <td>148.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>113</td>\n",
       "      <td>1.0</td>\n",
       "      <td>42.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>13</td>\n",
       "      <td>19</td>\n",
       "      <td>6</td>\n",
       "      <td>2018</td>\n",
       "      <td>41.879236</td>\n",
       "      <td>-87.624347</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3889</th>\n",
       "      <td>0</td>\n",
       "      <td>94.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>422</td>\n",
       "      <td>4.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>46.0</td>\n",
       "      <td>21</td>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "      <td>2015</td>\n",
       "      <td>41.751816</td>\n",
       "      <td>-87.560005</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>370424</th>\n",
       "      <td>0</td>\n",
       "      <td>147.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1115</td>\n",
       "      <td>11.0</td>\n",
       "      <td>28.0</td>\n",
       "      <td>26.0</td>\n",
       "      <td>20</td>\n",
       "      <td>31</td>\n",
       "      <td>5</td>\n",
       "      <td>2016</td>\n",
       "      <td>41.878856</td>\n",
       "      <td>-87.726672</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>450778</th>\n",
       "      <td>0</td>\n",
       "      <td>133.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2533</td>\n",
       "      <td>25.0</td>\n",
       "      <td>37.0</td>\n",
       "      <td>25.0</td>\n",
       "      <td>21</td>\n",
       "      <td>8</td>\n",
       "      <td>9</td>\n",
       "      <td>2016</td>\n",
       "      <td>41.909604</td>\n",
       "      <td>-87.747667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>76787</th>\n",
       "      <td>2</td>\n",
       "      <td>167.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>922</td>\n",
       "      <td>9.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>61.0</td>\n",
       "      <td>19</td>\n",
       "      <td>25</td>\n",
       "      <td>4</td>\n",
       "      <td>2015</td>\n",
       "      <td>41.808419</td>\n",
       "      <td>-87.684047</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1301102</th>\n",
       "      <td>1</td>\n",
       "      <td>74.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1924</td>\n",
       "      <td>19.0</td>\n",
       "      <td>44.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0</td>\n",
       "      <td>16</td>\n",
       "      <td>11</td>\n",
       "      <td>2019</td>\n",
       "      <td>41.940669</td>\n",
       "      <td>-87.655347</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>690010</th>\n",
       "      <td>1</td>\n",
       "      <td>125.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2532</td>\n",
       "      <td>25.0</td>\n",
       "      <td>37.0</td>\n",
       "      <td>25.0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>8</td>\n",
       "      <td>2017</td>\n",
       "      <td>41.903118</td>\n",
       "      <td>-87.759231</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>727690</th>\n",
       "      <td>2</td>\n",
       "      <td>147.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1121</td>\n",
       "      <td>11.0</td>\n",
       "      <td>27.0</td>\n",
       "      <td>23.0</td>\n",
       "      <td>19</td>\n",
       "      <td>18</td>\n",
       "      <td>9</td>\n",
       "      <td>2017</td>\n",
       "      <td>41.894314</td>\n",
       "      <td>-87.710181</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1247344 rows × 14 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         Category  Location Description  Arrest  Domestic  Beat  District  \\\n",
       "1057342         0                  89.0       0         0   712       7.0   \n",
       "1352020         0                  24.0       0         0  1925      19.0   \n",
       "922407          1                 148.0       0         0   113       1.0   \n",
       "3889            0                  94.0       0         0   422       4.0   \n",
       "370424          0                 147.0       1         0  1115      11.0   \n",
       "...           ...                   ...     ...       ...   ...       ...   \n",
       "450778          0                 133.0       0         0  2533      25.0   \n",
       "76787           2                 167.0       1         0   922       9.0   \n",
       "1301102         1                  74.0       0         0  1924      19.0   \n",
       "690010          1                 125.0       0         0  2532      25.0   \n",
       "727690          2                 147.0       1         0  1121      11.0   \n",
       "\n",
       "         Ward  Community Area  Hour  Day  Month  Year   Latitude  Longitude  \n",
       "1057342  16.0            68.0     0   12     12  2018  41.787017 -87.645506  \n",
       "1352020  46.0             6.0    18    1      2  2020  41.949954 -87.649572  \n",
       "922407   42.0            32.0    13   19      6  2018  41.879236 -87.624347  \n",
       "3889      7.0            46.0    21    6      1  2015  41.751816 -87.560005  \n",
       "370424   28.0            26.0    20   31      5  2016  41.878856 -87.726672  \n",
       "...       ...             ...   ...  ...    ...   ...        ...        ...  \n",
       "450778   37.0            25.0    21    8      9  2016  41.909604 -87.747667  \n",
       "76787    12.0            61.0    19   25      4  2015  41.808419 -87.684047  \n",
       "1301102  44.0             6.0     0   16     11  2019  41.940669 -87.655347  \n",
       "690010   37.0            25.0     0    1      8  2017  41.903118 -87.759231  \n",
       "727690   27.0            23.0    19   18      9  2017  41.894314 -87.710181  \n",
       "\n",
       "[1247344 rows x 14 columns]"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = df.sample(frac=1)\n",
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ede0b2ac",
   "metadata": {},
   "source": [
    "La matrice de corrélation ci-dessous nous montre que les attributs `Beat`, `District`, `Ward` et `Community Area` sont très corrélés (positivement ou négativement), corrélés aussi avec la `Latitude` et la `Longitude`.\n",
    "\n",
    "Ceci est normal comme ce sont tous des attributs qui représentent un emplacement sur la carte :\n",
    "\n",
    "**TL;DR**:\n",
    "* Beat : Indicates the beat where the incident occurred. A beat is the smallest police geographic area – each beat has a dedicated police beat car. Three to five beats make up a police sector, and three sectors make up a police district. The Chicago Police Department has 22 police districts. See the beats at https://data.cityofchicago.org/d/aerh-rz74.\n",
    "* Disctrict : Indicates the police district where the incident occurred. See the districts at https://data.cityofchicago.org/d/fthy-xz3r. (Chicago compte 22 districts de police).\n",
    "* Ward : The ward (City Council district) where the incident occurred. See the wards at https://data.cityofchicago.org/d/sp34-6z76.\n",
    "* Community Area : Indicates the community area where the incident occurred. Chicago has 77 community areas. See the community areas at https://data.cityofchicago.org/d/cauq-8yn6.\n",
    "\n",
    "(Source : https://www.kaggle.com/chicago/chicago-crime)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "8253358d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "corr_matrix = df.corr()\n",
    "plt.figure(figsize=(15, 5))\n",
    "sns.heatmap(data = corr_matrix, annot=True, linewidths=0.2)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cb40b2cd",
   "metadata": {},
   "source": [
    "On a choisi de garder l'attribut `District` qui possède 22 valeurs différentes, et de supprimer les autres attributs qui portent des informations similaires."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "d38839ae",
   "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>Category</th>\n",
       "      <th>Location Description</th>\n",
       "      <th>Arrest</th>\n",
       "      <th>Domestic</th>\n",
       "      <th>District</th>\n",
       "      <th>Hour</th>\n",
       "      <th>Day</th>\n",
       "      <th>Month</th>\n",
       "      <th>Year</th>\n",
       "      <th>Latitude</th>\n",
       "      <th>Longitude</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>89.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0</td>\n",
       "      <td>12</td>\n",
       "      <td>12</td>\n",
       "      <td>2018</td>\n",
       "      <td>41.787017</td>\n",
       "      <td>-87.645506</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>24.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>18</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>2020</td>\n",
       "      <td>41.949954</td>\n",
       "      <td>-87.649572</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>148.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>13</td>\n",
       "      <td>19</td>\n",
       "      <td>6</td>\n",
       "      <td>2018</td>\n",
       "      <td>41.879236</td>\n",
       "      <td>-87.624347</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>94.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>21</td>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "      <td>2015</td>\n",
       "      <td>41.751816</td>\n",
       "      <td>-87.560005</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>147.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>20</td>\n",
       "      <td>31</td>\n",
       "      <td>5</td>\n",
       "      <td>2016</td>\n",
       "      <td>41.878856</td>\n",
       "      <td>-87.726672</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1247339</th>\n",
       "      <td>0</td>\n",
       "      <td>133.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>25.0</td>\n",
       "      <td>21</td>\n",
       "      <td>8</td>\n",
       "      <td>9</td>\n",
       "      <td>2016</td>\n",
       "      <td>41.909604</td>\n",
       "      <td>-87.747667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1247340</th>\n",
       "      <td>2</td>\n",
       "      <td>167.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>19</td>\n",
       "      <td>25</td>\n",
       "      <td>4</td>\n",
       "      <td>2015</td>\n",
       "      <td>41.808419</td>\n",
       "      <td>-87.684047</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1247341</th>\n",
       "      <td>1</td>\n",
       "      <td>74.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>0</td>\n",
       "      <td>16</td>\n",
       "      <td>11</td>\n",
       "      <td>2019</td>\n",
       "      <td>41.940669</td>\n",
       "      <td>-87.655347</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1247342</th>\n",
       "      <td>1</td>\n",
       "      <td>125.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>25.0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>8</td>\n",
       "      <td>2017</td>\n",
       "      <td>41.903118</td>\n",
       "      <td>-87.759231</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1247343</th>\n",
       "      <td>2</td>\n",
       "      <td>147.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>19</td>\n",
       "      <td>18</td>\n",
       "      <td>9</td>\n",
       "      <td>2017</td>\n",
       "      <td>41.894314</td>\n",
       "      <td>-87.710181</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1247344 rows × 11 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         Category  Location Description  Arrest  Domestic  District  Hour  \\\n",
       "0               0                  89.0       0         0       7.0     0   \n",
       "1               0                  24.0       0         0      19.0    18   \n",
       "2               1                 148.0       0         0       1.0    13   \n",
       "3               0                  94.0       0         0       4.0    21   \n",
       "4               0                 147.0       1         0      11.0    20   \n",
       "...           ...                   ...     ...       ...       ...   ...   \n",
       "1247339         0                 133.0       0         0      25.0    21   \n",
       "1247340         2                 167.0       1         0       9.0    19   \n",
       "1247341         1                  74.0       0         0      19.0     0   \n",
       "1247342         1                 125.0       0         0      25.0     0   \n",
       "1247343         2                 147.0       1         0      11.0    19   \n",
       "\n",
       "         Day  Month  Year   Latitude  Longitude  \n",
       "0         12     12  2018  41.787017 -87.645506  \n",
       "1          1      2  2020  41.949954 -87.649572  \n",
       "2         19      6  2018  41.879236 -87.624347  \n",
       "3          6      1  2015  41.751816 -87.560005  \n",
       "4         31      5  2016  41.878856 -87.726672  \n",
       "...      ...    ...   ...        ...        ...  \n",
       "1247339    8      9  2016  41.909604 -87.747667  \n",
       "1247340   25      4  2015  41.808419 -87.684047  \n",
       "1247341   16     11  2019  41.940669 -87.655347  \n",
       "1247342    1      8  2017  41.903118 -87.759231  \n",
       "1247343   18      9  2017  41.894314 -87.710181  \n",
       "\n",
       "[1247344 rows x 11 columns]"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.drop(columns = ['Beat', 'Ward', 'Community Area'], inplace = True)\n",
    "df.reset_index(drop=True, inplace=True)\n",
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "87cee8df",
   "metadata": {},
   "source": [
    "On obtient alors la matrice de corrélation suivant :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "d9a2dafd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "corr_matrix = df.corr()\n",
    "plt.figure(figsize=(15, 5))\n",
    "sns.heatmap(data = corr_matrix, annot=True, linewidths=0.2)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "043e67fa",
   "metadata": {},
   "source": [
    "Avant de commencer la modélisation, et sachant que certains modèles nécessitent de garder le jeu de données, on sait qu'avec notre jeu de données avec plus de **1.5 million** d'exemples, certains modèles metteront une éternité pour s'entaîner avec nos simples ordinateurs limités par les ressources.\n",
    "\n",
    "On a donc décidé de travailler sur un échantillon (de 20 000 lignes) du dataset d'origine qui soit le plus représentatif que possible du dataset initial.\n",
    "\n",
    "On utilise la fonction `sample` pour échantillonner. Puis on fait appel au *test de Kolmogorov-Smirnov* pour vérifier que l'échantillonnage nous a pas fait perdre des données significatives.\n",
    "\n",
    "Pour vérifier cela, il faut que les `pvalues` soient toutes au moins supérieures à 5% (0.05).\n",
    "\n",
    "Pour plus de détails, lire la documentation de `ks_2samp` de `scipy`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "2621ab62",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.9991564221798757,\n",
       " 0.5142266183552056,\n",
       " 0.9966942362039569,\n",
       " 0.9999999998727649,\n",
       " 0.7585368527272978,\n",
       " 0.8509611576281594,\n",
       " 0.8275199887928899,\n",
       " 0.39384224332337325,\n",
       " 0.9999999999642781,\n",
       " 0.6304493377224243,\n",
       " 0.09910468690107133]"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from scipy.stats import ks_2samp\n",
    "\n",
    "sample_size = 20000\n",
    "subsample = df.sample(sample_size, axis = 0, random_state=1)\n",
    "pvalues = [None for _ in range(df.shape[1])]\n",
    "for col in range(df.shape[1]) :\n",
    "    # Using Kolmogorov-Smirnov test to check whether the sample truely represents the original dataset\n",
    "    _, pvalues[col] = ks_2samp(df.iloc[:,col], subsample.iloc[:,col])\n",
    "\n",
    "pvalues"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bbe22df0",
   "metadata": {},
   "source": [
    "On garde notre DataFrame originale dans `df_legacy` et l'échantillon dans `df`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "ae138086",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_legacy = df\n",
    "df = subsample\n",
    "df.reset_index(drop=True, inplace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5c716832",
   "metadata": {},
   "source": [
    "On vérifie que la matrice de corrélation pour l'échantillon est identique à celle du dataset d'origine."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "54605bfc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "corr_matrix = df.corr()\n",
    "plt.figure(figsize=(15, 5))\n",
    "sns.heatmap(data = corr_matrix, annot=True, linewidths=0.2)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0a808d5c",
   "metadata": {},
   "source": [
    "Et que la distribution des exemples par catégories est toujours respectée."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "be07f493",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(15, 5))\n",
    "plt.bar(df['Category'].value_counts().index.tolist(),df['Category'].value_counts())\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9e8f3d56",
   "metadata": {},
   "source": [
    "On sépare l'attribut qu'on aimerait prédire et qui contient les 3 classes du dataset et on le met dans la variable `y`, et le reste des attributs dans la variable `X`.\n",
    "\n",
    "On fait de la **classification multiclasses**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "0391c511",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = df.drop('Category', axis=1)\n",
    "y = df['Category']"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ad3a522a",
   "metadata": {},
   "source": [
    "On crée le jeu d'entraînement (train set : 70%) et le jeu de test (test set : 30%) :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "a41482d0",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Création de l'ensemble d'entrainement et de l'ensemble de test\n",
    "from sklearn.model_selection import train_test_split\n",
    "X_train, X_test, y_train, y_test = train_test_split(X.values, y.values, test_size = 0.3, random_state = 1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "78ed64f5",
   "metadata": {},
   "source": [
    "On standardize nos attributs grâce à un `StandardScaler` pour améliorer les performances lors de l'entraînement et éviter les problèmes numériques lors du calcul des gradients et de l'optimisation des fonctions objectif. (ne pas avoir les attributs sur des plages très différentes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "5227066e",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "scaler = StandardScaler()\n",
    "X_train = scaler.fit_transform(X_train)\n",
    "X_test = scaler.transform(X_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9783fea7",
   "metadata": {},
   "source": [
    "## 3. Modélisation :\n",
    "\n",
    "On passe maintenant à l'étape de modélisation. On crée différents *classifiers* (modèles pour classification multiclasses) et on compare leurs performances pour choisir le meilleur d'entre eux."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9fa658df",
   "metadata": {},
   "source": [
    "### 3.1 KNN :\n",
    "\n",
    "On commence par un modèle de type K Nearest Neighbor.\n",
    "\n",
    "On choisit l'hyperparamètre `n_neighbors` optimal avec `GridSearchCV` comme vu en cours.\n",
    "\n",
    "On a préféré utiliser le score `f1_weighted` (f1 score pondéré par les proportions des classes dans notre dataset) pour tous nos modèles puisque c'est un score qui représente mieux la performance lors d'une classification multiclasses avec un jeu de données non équilibré (c'est notre cas)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "9036e1e7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 10 candidates, totalling 50 fits\n"
     ]
    }
   ],
   "source": [
    "# Utiliser la méthode \"GridSearchCV\" pour faire une validation croisée du paramètre k d’un kNN (le nombre de plus proches voisins)\n",
    "# sur le jeu d’entraînement \n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.metrics import balanced_accuracy_score\n",
    "\n",
    "\n",
    "# Fixer les valeurs des hyperparamètres à tester\n",
    "param_grid = {'n_neighbors':[13, 15, 17, 19, 21, 23, 25, 27, 29, 31]}\n",
    "\n",
    "\n",
    "# Choisir un score à optimiser, ici f1-score pondéré (plus significatif que l'accuracy surtout en multiclass)\n",
    "score = 'f1_weighted'\n",
    "\n",
    "# Créer un classifieur kNN avec recherche d'hyperparamètre par validation croisée\n",
    "knn_grid = GridSearchCV(\n",
    "    KNeighborsClassifier(),\n",
    "    param_grid,\n",
    "    cv = 5,\n",
    "    verbose = 1,\n",
    "    scoring = score\n",
    ")\n",
    "\n",
    "# Optimiser ce classifieur sur le jeu d'entraînement\n",
    "clf_results = knn_grid.fit(X_train,y_train)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "99b8692c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best parameters : {'n_neighbors': 27}\n",
      "Best cross-validation score : 0.71\n"
     ]
    }
   ],
   "source": [
    "# Afficher le(s) hyperparamètre(s) optimaux\n",
    "print(\"Best parameters : {}\".format(knn_grid.best_params_))\n",
    "print(\"Best cross-validation score : {:.2f}\".format(knn_grid.best_score_))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "9650fdab",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Résultats de la validation croisée :\n",
      "f1_weighted = 0.707 (+/-0.013) for {'n_neighbors': 13}\n",
      "f1_weighted = 0.709 (+/-0.008) for {'n_neighbors': 15}\n",
      "f1_weighted = 0.709 (+/-0.010) for {'n_neighbors': 17}\n",
      "f1_weighted = 0.711 (+/-0.012) for {'n_neighbors': 19}\n",
      "f1_weighted = 0.711 (+/-0.012) for {'n_neighbors': 21}\n",
      "f1_weighted = 0.711 (+/-0.013) for {'n_neighbors': 23}\n",
      "f1_weighted = 0.711 (+/-0.015) for {'n_neighbors': 25}\n",
      "f1_weighted = 0.713 (+/-0.016) for {'n_neighbors': 27}\n",
      "f1_weighted = 0.712 (+/-0.016) for {'n_neighbors': 29}\n",
      "f1_weighted = 0.712 (+/-0.016) for {'n_neighbors': 31}\n"
     ]
    }
   ],
   "source": [
    "# Afficher les performances correspondantes\n",
    "print(\"Résultats de la validation croisée :\")\n",
    "for mean, std, params in zip(\n",
    "        knn_grid.cv_results_['mean_test_score'], # score moyen\n",
    "        knn_grid.cv_results_['std_test_score'],  # écart-type du score\n",
    "        knn_grid.cv_results_['params']           # valeur de l'hyperparamètre\n",
    "    ):\n",
    "\n",
    "    print(\"{} = {:.3f} (+/-{:.03f}) for {}\".format(\n",
    "        score,\n",
    "        mean,\n",
    "        std*2,\n",
    "        params\n",
    "    ) )"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "53fbe33b",
   "metadata": {},
   "source": [
    "**Efficacité moyenne du modèle par cross-validation (validation croisée)**\n",
    "\n",
    "On utilise la fonction `cross-validate` vue en cours avec des scores adaptés au cas multiclass :\n",
    "* précision pondérée\n",
    "* rappel pondéré\n",
    "* balanced accuracy\n",
    "* f1 score pondéré\n",
    "* roc auc avec du one vs rest et de la pondération\n",
    "\n",
    "(Détails dans **sklearn.metrics**)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "16cd8990",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 10 candidates, totalling 50 fits\n",
      "Fitting 5 folds for each of 10 candidates, totalling 50 fits\n",
      "Fitting 5 folds for each of 10 candidates, totalling 50 fits\n",
      "Fitting 5 folds for each of 10 candidates, totalling 50 fits\n",
      "Fitting 5 folds for each of 10 candidates, totalling 50 fits\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import cross_validate\n",
    "scores_knn = cross_validate(knn_grid, X_train, y_train, cv=5,\n",
    "            scoring=('precision_weighted','recall_weighted','balanced_accuracy','f1_weighted','roc_auc_ovr_weighted'))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7ce1c5f9",
   "metadata": {},
   "source": [
    "On affiche la performance de ce modèle : (71.2%)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "bec9ef0f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cross-validation performance : 0.712\n"
     ]
    }
   ],
   "source": [
    "print(\"Cross-validation performance : {:.3f}\".format(scores_knn['test_f1_weighted'].mean()))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "06c991f4",
   "metadata": {},
   "source": [
    "### 3.2 Random Forest :\n",
    "\n",
    "On passe à un modèle de type Random Forest.\n",
    "\n",
    "On choisit les hyperparamètre `n_estimators` et `max_depth` optimaux avec cette fois `RandomizedSearchCV`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "6726becf",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 50 candidates, totalling 250 fits\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "RandomizedSearchCV(cv=5, estimator=RandomForestClassifier(), n_iter=50,\n",
       "                   n_jobs=-1,\n",
       "                   param_distributions={'max_depth': [20, 40, 60, 80, 100],\n",
       "                                        'n_estimators': [20, 40, 60, 80, 100,\n",
       "                                                         120, 140, 160, 180,\n",
       "                                                         200]},\n",
       "                   random_state=1, scoring='f1_weighted', verbose=2)"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.model_selection import RandomizedSearchCV\n",
    "random_grid = {\n",
    " 'n_estimators': [int(x) for x in np.linspace(start = 20, stop = 200, num = 10)],\n",
    " 'max_depth': [int(x) for x in np.linspace(20, 100, num = 5)]\n",
    " }\n",
    "rf_random = RandomizedSearchCV(estimator = RandomForestClassifier(),\n",
    "                               param_distributions = random_grid,\n",
    "                               n_iter=50,\n",
    "                               cv=5,\n",
    "                               verbose=2,\n",
    "                               random_state=1,\n",
    "                               scoring='f1_weighted',\n",
    "                               n_jobs=-1)\n",
    "rf_random.fit(X_train,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "958a134c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best parameters : {'n_estimators': 120, 'max_depth': 80}\n",
      "Best cross-validation score : 0.74\n"
     ]
    }
   ],
   "source": [
    "# Afficher le(s) hyperparamètre(s) optimaux\n",
    "print(\"Best parameters : {}\".format(rf_random.best_params_))\n",
    "print(\"Best cross-validation score : {:.2f}\".format(rf_random.best_score_))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "437804f5",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Résultats de la validation croisée :\n",
      "f1_weighted = 0.729 (+/-0.009) for {'n_estimators': 20, 'max_depth': 20}\n",
      "f1_weighted = 0.731 (+/-0.013) for {'n_estimators': 40, 'max_depth': 20}\n",
      "f1_weighted = 0.735 (+/-0.017) for {'n_estimators': 60, 'max_depth': 20}\n",
      "f1_weighted = 0.736 (+/-0.014) for {'n_estimators': 80, 'max_depth': 20}\n",
      "f1_weighted = 0.734 (+/-0.014) for {'n_estimators': 100, 'max_depth': 20}\n",
      "f1_weighted = 0.734 (+/-0.012) for {'n_estimators': 120, 'max_depth': 20}\n",
      "f1_weighted = 0.734 (+/-0.015) for {'n_estimators': 140, 'max_depth': 20}\n",
      "f1_weighted = 0.734 (+/-0.012) for {'n_estimators': 160, 'max_depth': 20}\n",
      "f1_weighted = 0.736 (+/-0.011) for {'n_estimators': 180, 'max_depth': 20}\n",
      "f1_weighted = 0.734 (+/-0.017) for {'n_estimators': 200, 'max_depth': 20}\n",
      "f1_weighted = 0.726 (+/-0.020) for {'n_estimators': 20, 'max_depth': 40}\n",
      "f1_weighted = 0.730 (+/-0.014) for {'n_estimators': 40, 'max_depth': 40}\n",
      "f1_weighted = 0.733 (+/-0.009) for {'n_estimators': 60, 'max_depth': 40}\n",
      "f1_weighted = 0.735 (+/-0.011) for {'n_estimators': 80, 'max_depth': 40}\n",
      "f1_weighted = 0.734 (+/-0.015) for {'n_estimators': 100, 'max_depth': 40}\n",
      "f1_weighted = 0.735 (+/-0.016) for {'n_estimators': 120, 'max_depth': 40}\n",
      "f1_weighted = 0.735 (+/-0.012) for {'n_estimators': 140, 'max_depth': 40}\n",
      "f1_weighted = 0.735 (+/-0.013) for {'n_estimators': 160, 'max_depth': 40}\n",
      "f1_weighted = 0.734 (+/-0.010) for {'n_estimators': 180, 'max_depth': 40}\n",
      "f1_weighted = 0.734 (+/-0.015) for {'n_estimators': 200, 'max_depth': 40}\n",
      "f1_weighted = 0.724 (+/-0.018) for {'n_estimators': 20, 'max_depth': 60}\n",
      "f1_weighted = 0.729 (+/-0.017) for {'n_estimators': 40, 'max_depth': 60}\n",
      "f1_weighted = 0.733 (+/-0.010) for {'n_estimators': 60, 'max_depth': 60}\n",
      "f1_weighted = 0.734 (+/-0.015) for {'n_estimators': 80, 'max_depth': 60}\n",
      "f1_weighted = 0.732 (+/-0.016) for {'n_estimators': 100, 'max_depth': 60}\n",
      "f1_weighted = 0.735 (+/-0.012) for {'n_estimators': 120, 'max_depth': 60}\n",
      "f1_weighted = 0.735 (+/-0.012) for {'n_estimators': 140, 'max_depth': 60}\n",
      "f1_weighted = 0.733 (+/-0.014) for {'n_estimators': 160, 'max_depth': 60}\n",
      "f1_weighted = 0.735 (+/-0.010) for {'n_estimators': 180, 'max_depth': 60}\n",
      "f1_weighted = 0.733 (+/-0.013) for {'n_estimators': 200, 'max_depth': 60}\n",
      "f1_weighted = 0.725 (+/-0.017) for {'n_estimators': 20, 'max_depth': 80}\n",
      "f1_weighted = 0.730 (+/-0.010) for {'n_estimators': 40, 'max_depth': 80}\n",
      "f1_weighted = 0.732 (+/-0.012) for {'n_estimators': 60, 'max_depth': 80}\n",
      "f1_weighted = 0.735 (+/-0.013) for {'n_estimators': 80, 'max_depth': 80}\n",
      "f1_weighted = 0.733 (+/-0.011) for {'n_estimators': 100, 'max_depth': 80}\n",
      "f1_weighted = 0.736 (+/-0.012) for {'n_estimators': 120, 'max_depth': 80}\n",
      "f1_weighted = 0.735 (+/-0.011) for {'n_estimators': 140, 'max_depth': 80}\n",
      "f1_weighted = 0.734 (+/-0.015) for {'n_estimators': 160, 'max_depth': 80}\n",
      "f1_weighted = 0.735 (+/-0.018) for {'n_estimators': 180, 'max_depth': 80}\n",
      "f1_weighted = 0.736 (+/-0.014) for {'n_estimators': 200, 'max_depth': 80}\n",
      "f1_weighted = 0.720 (+/-0.013) for {'n_estimators': 20, 'max_depth': 100}\n",
      "f1_weighted = 0.729 (+/-0.012) for {'n_estimators': 40, 'max_depth': 100}\n",
      "f1_weighted = 0.732 (+/-0.010) for {'n_estimators': 60, 'max_depth': 100}\n",
      "f1_weighted = 0.734 (+/-0.016) for {'n_estimators': 80, 'max_depth': 100}\n",
      "f1_weighted = 0.731 (+/-0.017) for {'n_estimators': 100, 'max_depth': 100}\n",
      "f1_weighted = 0.734 (+/-0.012) for {'n_estimators': 120, 'max_depth': 100}\n",
      "f1_weighted = 0.732 (+/-0.009) for {'n_estimators': 140, 'max_depth': 100}\n",
      "f1_weighted = 0.733 (+/-0.009) for {'n_estimators': 160, 'max_depth': 100}\n",
      "f1_weighted = 0.734 (+/-0.014) for {'n_estimators': 180, 'max_depth': 100}\n",
      "f1_weighted = 0.735 (+/-0.011) for {'n_estimators': 200, 'max_depth': 100}\n"
     ]
    }
   ],
   "source": [
    "# Afficher les performances correspondantes\n",
    "print(\"Résultats de la validation croisée :\")\n",
    "for mean, std, params in zip(\n",
    "        rf_random.cv_results_['mean_test_score'], # score moyen\n",
    "        rf_random.cv_results_['std_test_score'],  # écart-type du score\n",
    "        rf_random.cv_results_['params']           # valeur de l'hyperparamètre\n",
    "    ):\n",
    "\n",
    "    print(\"{} = {:.3f} (+/-{:.03f}) for {}\".format(\n",
    "        score,\n",
    "        mean,\n",
    "        std*2,\n",
    "        params\n",
    "    ) )"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b2685bf5",
   "metadata": {},
   "source": [
    "**Calcul de l'efficacité moyenne du modèle par cross-validation (validation croisée)**\n",
    "\n",
    "Puis on l'affiche : (73.4%)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "ed34abd5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 50 candidates, totalling 250 fits\n",
      "Fitting 5 folds for each of 50 candidates, totalling 250 fits\n",
      "Fitting 5 folds for each of 50 candidates, totalling 250 fits\n",
      "Fitting 5 folds for each of 50 candidates, totalling 250 fits\n",
      "Fitting 5 folds for each of 50 candidates, totalling 250 fits\n",
      "Cross-validation performance : 0.734\n"
     ]
    }
   ],
   "source": [
    "scores_rf_random = cross_validate(rf_random, X_train, y_train, cv=5,\n",
    "                    scoring=('precision_weighted','recall_weighted','balanced_accuracy','f1_weighted','roc_auc_ovr_weighted'))\n",
    "print(\"Cross-validation performance : {:.3f}\".format(scores_rf_random['test_f1_weighted'].mean()))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3d448b0b",
   "metadata": {},
   "source": [
    "Comme attendu, le modèle Random Forest fait un peu mieux qu'un KNN."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "abb4b3db",
   "metadata": {},
   "source": [
    "### 3.3 XGBoost :\n",
    "\n",
    "On passe maintenant à un modèle très connu par son efficacité : **XGBoost**.\n",
    "\n",
    "Le modèle possède un bon nombre d'hyperparamètres, on choisit la meilleure combinaison entre certains d'entre eux pour les valeurs qu'on a donnée dans le dictionnaire `param_dist`, on utilise encore une fois `RandomizedSearchCV`.\n",
    "\n",
    "On utilise aussi cette fois un objet `KFold` pour choisir le nombre de plis sur nos données d'entraînement qu'on va utiliser pour la validation croisée dans `RandomizedSearchCV`.\n",
    "\n",
    "La **fonction objectif** qu'on spécifie pour le classifier dans notre cas multiclass est *'multi:softmax'*. Le **scoring** toujours avec du *'f1_weighted'*. Enfin, la **métrique d'évalution** (eval_metric) dans la fonction `fit` est le *'multi log loss'* (fonction de perte)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "755db42e",
   "metadata": {},
   "outputs": [],
   "source": [
    "import xgboost as xgb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "94a92af1",
   "metadata": {},
   "outputs": [],
   "source": [
    "from scipy import stats\n",
    "from sklearn.model_selection import KFold\n",
    "\n",
    "xgb_clf = xgb.XGBClassifier(use_label_encoder=False, objective='multi:softmax')\n",
    "param_dist = {'n_estimators': [int(x) for x in np.linspace(start = 20, stop = 200, num = 10)],\n",
    "              'learning_rate': [0.001, 0.003, 0.01, 0.03, 0.1, 0.3],\n",
    "              'subsample': (0.9 - 0.3) * random.random_sample(3)+ 0.3,\n",
    "              'max_depth': [int(x) for x in np.linspace(20, 100, num = 5)],\n",
    "              'colsample_bytree': (0.9 - 0.5) * random.random_sample(3) + 0.5\n",
    "             }\n",
    "\n",
    "numFolds = 5\n",
    "kfold_5 = KFold(random_state=1, shuffle = True, n_splits = numFolds)\n",
    "\n",
    "xgb_random = RandomizedSearchCV(xgb_clf, \n",
    "                         param_distributions = param_dist,\n",
    "                         cv = kfold_5,\n",
    "                         n_iter = 50,\n",
    "                         scoring = 'f1_weighted', \n",
    "                         error_score = 0, \n",
    "                         verbose = 3, \n",
    "                         n_jobs = -1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "96150c46",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 50 candidates, totalling 250 fits\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "RandomizedSearchCV(cv=KFold(n_splits=5, random_state=1, shuffle=True),\n",
       "                   error_score=0,\n",
       "                   estimator=XGBClassifier(base_score=None, booster=None,\n",
       "                                           colsample_bylevel=None,\n",
       "                                           colsample_bynode=None,\n",
       "                                           colsample_bytree=None, gamma=None,\n",
       "                                           gpu_id=None, importance_type='gain',\n",
       "                                           interaction_constraints=None,\n",
       "                                           learning_rate=None,\n",
       "                                           max_delta_step=None, max_depth=None,\n",
       "                                           min_child_weight=No...\n",
       "                                           validate_parameters=None,\n",
       "                                           verbosity=None),\n",
       "                   n_iter=50, n_jobs=-1,\n",
       "                   param_distributions={'colsample_bytree': array([0.86918892, 0.55963558, 0.75823576]),\n",
       "                                        'learning_rate': [0.001, 0.003, 0.01,\n",
       "                                                          0.03, 0.1, 0.3],\n",
       "                                        'max_depth': [20, 40, 60, 80, 100],\n",
       "                                        'n_estimators': [20, 40, 60, 80, 100,\n",
       "                                                         120, 140, 160, 180,\n",
       "                                                         200],\n",
       "                                        'subsample': array([0.7627155 , 0.49568248, 0.42770458])},\n",
       "                   scoring='f1_weighted', verbose=3)"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xgb_random.fit(X_train,y_train, eval_metric='mlogloss')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "74245ae8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best parameters : {'subsample': 0.49568248279310045, 'n_estimators': 100, 'max_depth': 80, 'learning_rate': 0.001, 'colsample_bytree': 0.8691889161717675}\n",
      "Best cross-validation score : 0.74\n"
     ]
    }
   ],
   "source": [
    "# Afficher le(s) hyperparamètre(s) optimaux\n",
    "print(\"Best parameters : {}\".format(xgb_random.best_params_))\n",
    "print(\"Best cross-validation score : {:.2f}\".format(xgb_random.best_score_))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "88ff8f74",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Résultats de la validation croisée :\n",
      "f1_weighted = 0.729 (+/-0.011) for {'subsample': 0.7627155014500968, 'n_estimators': 80, 'max_depth': 40, 'learning_rate': 0.01, 'colsample_bytree': 0.559635577743145}\n",
      "f1_weighted = 0.738 (+/-0.016) for {'subsample': 0.4277045805192762, 'n_estimators': 180, 'max_depth': 40, 'learning_rate': 0.001, 'colsample_bytree': 0.8691889161717675}\n",
      "f1_weighted = 0.735 (+/-0.016) for {'subsample': 0.7627155014500968, 'n_estimators': 20, 'max_depth': 80, 'learning_rate': 0.1, 'colsample_bytree': 0.8691889161717675}\n",
      "f1_weighted = 0.736 (+/-0.016) for {'subsample': 0.4277045805192762, 'n_estimators': 100, 'max_depth': 80, 'learning_rate': 0.003, 'colsample_bytree': 0.7582357629850743}\n",
      "f1_weighted = 0.734 (+/-0.021) for {'subsample': 0.7627155014500968, 'n_estimators': 40, 'max_depth': 80, 'learning_rate': 0.1, 'colsample_bytree': 0.7582357629850743}\n",
      "f1_weighted = 0.735 (+/-0.013) for {'subsample': 0.4277045805192762, 'n_estimators': 20, 'max_depth': 60, 'learning_rate': 0.03, 'colsample_bytree': 0.8691889161717675}\n",
      "f1_weighted = 0.738 (+/-0.012) for {'subsample': 0.4277045805192762, 'n_estimators': 40, 'max_depth': 60, 'learning_rate': 0.03, 'colsample_bytree': 0.8691889161717675}\n",
      "f1_weighted = 0.732 (+/-0.014) for {'subsample': 0.7627155014500968, 'n_estimators': 140, 'max_depth': 40, 'learning_rate': 0.03, 'colsample_bytree': 0.559635577743145}\n",
      "f1_weighted = 0.736 (+/-0.016) for {'subsample': 0.4277045805192762, 'n_estimators': 180, 'max_depth': 20, 'learning_rate': 0.001, 'colsample_bytree': 0.7582357629850743}\n",
      "f1_weighted = 0.733 (+/-0.013) for {'subsample': 0.4277045805192762, 'n_estimators': 20, 'max_depth': 60, 'learning_rate': 0.003, 'colsample_bytree': 0.7582357629850743}\n",
      "f1_weighted = 0.739 (+/-0.016) for {'subsample': 0.49568248279310045, 'n_estimators': 100, 'max_depth': 80, 'learning_rate': 0.001, 'colsample_bytree': 0.8691889161717675}\n",
      "f1_weighted = 0.733 (+/-0.015) for {'subsample': 0.7627155014500968, 'n_estimators': 140, 'max_depth': 40, 'learning_rate': 0.1, 'colsample_bytree': 0.7582357629850743}\n",
      "f1_weighted = 0.738 (+/-0.015) for {'subsample': 0.4277045805192762, 'n_estimators': 160, 'max_depth': 80, 'learning_rate': 0.01, 'colsample_bytree': 0.7582357629850743}\n",
      "f1_weighted = 0.734 (+/-0.023) for {'subsample': 0.7627155014500968, 'n_estimators': 20, 'max_depth': 40, 'learning_rate': 0.1, 'colsample_bytree': 0.7582357629850743}\n",
      "f1_weighted = 0.737 (+/-0.015) for {'subsample': 0.4277045805192762, 'n_estimators': 100, 'max_depth': 100, 'learning_rate': 0.001, 'colsample_bytree': 0.8691889161717675}\n",
      "f1_weighted = 0.726 (+/-0.019) for {'subsample': 0.49568248279310045, 'n_estimators': 100, 'max_depth': 80, 'learning_rate': 0.001, 'colsample_bytree': 0.559635577743145}\n",
      "f1_weighted = 0.722 (+/-0.018) for {'subsample': 0.49568248279310045, 'n_estimators': 120, 'max_depth': 20, 'learning_rate': 0.3, 'colsample_bytree': 0.559635577743145}\n",
      "f1_weighted = 0.726 (+/-0.015) for {'subsample': 0.7627155014500968, 'n_estimators': 40, 'max_depth': 80, 'learning_rate': 0.003, 'colsample_bytree': 0.559635577743145}\n",
      "f1_weighted = 0.722 (+/-0.009) for {'subsample': 0.4277045805192762, 'n_estimators': 20, 'max_depth': 60, 'learning_rate': 0.03, 'colsample_bytree': 0.559635577743145}\n",
      "f1_weighted = 0.737 (+/-0.015) for {'subsample': 0.49568248279310045, 'n_estimators': 100, 'max_depth': 20, 'learning_rate': 0.001, 'colsample_bytree': 0.7582357629850743}\n",
      "f1_weighted = 0.737 (+/-0.015) for {'subsample': 0.7627155014500968, 'n_estimators': 40, 'max_depth': 60, 'learning_rate': 0.03, 'colsample_bytree': 0.8691889161717675}\n",
      "f1_weighted = 0.716 (+/-0.014) for {'subsample': 0.4277045805192762, 'n_estimators': 180, 'max_depth': 40, 'learning_rate': 0.3, 'colsample_bytree': 0.7582357629850743}\n",
      "f1_weighted = 0.736 (+/-0.014) for {'subsample': 0.4277045805192762, 'n_estimators': 60, 'max_depth': 100, 'learning_rate': 0.01, 'colsample_bytree': 0.7582357629850743}\n",
      "f1_weighted = 0.723 (+/-0.011) for {'subsample': 0.4277045805192762, 'n_estimators': 20, 'max_depth': 80, 'learning_rate': 0.003, 'colsample_bytree': 0.559635577743145}\n",
      "f1_weighted = 0.729 (+/-0.011) for {'subsample': 0.7627155014500968, 'n_estimators': 60, 'max_depth': 80, 'learning_rate': 0.01, 'colsample_bytree': 0.559635577743145}\n",
      "f1_weighted = 0.737 (+/-0.015) for {'subsample': 0.49568248279310045, 'n_estimators': 160, 'max_depth': 60, 'learning_rate': 0.03, 'colsample_bytree': 0.8691889161717675}\n",
      "f1_weighted = 0.730 (+/-0.019) for {'subsample': 0.4277045805192762, 'n_estimators': 100, 'max_depth': 20, 'learning_rate': 0.01, 'colsample_bytree': 0.559635577743145}\n",
      "f1_weighted = 0.725 (+/-0.016) for {'subsample': 0.49568248279310045, 'n_estimators': 20, 'max_depth': 40, 'learning_rate': 0.3, 'colsample_bytree': 0.7582357629850743}\n",
      "f1_weighted = 0.729 (+/-0.013) for {'subsample': 0.7627155014500968, 'n_estimators': 60, 'max_depth': 20, 'learning_rate': 0.01, 'colsample_bytree': 0.559635577743145}\n",
      "f1_weighted = 0.738 (+/-0.017) for {'subsample': 0.49568248279310045, 'n_estimators': 140, 'max_depth': 100, 'learning_rate': 0.003, 'colsample_bytree': 0.7582357629850743}\n",
      "f1_weighted = 0.737 (+/-0.014) for {'subsample': 0.4277045805192762, 'n_estimators': 140, 'max_depth': 20, 'learning_rate': 0.01, 'colsample_bytree': 0.8691889161717675}\n",
      "f1_weighted = 0.727 (+/-0.013) for {'subsample': 0.7627155014500968, 'n_estimators': 120, 'max_depth': 80, 'learning_rate': 0.3, 'colsample_bytree': 0.7582357629850743}\n",
      "f1_weighted = 0.728 (+/-0.016) for {'subsample': 0.4277045805192762, 'n_estimators': 100, 'max_depth': 40, 'learning_rate': 0.003, 'colsample_bytree': 0.559635577743145}\n",
      "f1_weighted = 0.732 (+/-0.013) for {'subsample': 0.7627155014500968, 'n_estimators': 200, 'max_depth': 80, 'learning_rate': 0.1, 'colsample_bytree': 0.559635577743145}\n",
      "f1_weighted = 0.738 (+/-0.014) for {'subsample': 0.7627155014500968, 'n_estimators': 160, 'max_depth': 60, 'learning_rate': 0.03, 'colsample_bytree': 0.7582357629850743}\n",
      "f1_weighted = 0.737 (+/-0.020) for {'subsample': 0.7627155014500968, 'n_estimators': 120, 'max_depth': 20, 'learning_rate': 0.003, 'colsample_bytree': 0.7582357629850743}\n",
      "f1_weighted = 0.738 (+/-0.019) for {'subsample': 0.49568248279310045, 'n_estimators': 80, 'max_depth': 100, 'learning_rate': 0.01, 'colsample_bytree': 0.8691889161717675}\n",
      "f1_weighted = 0.719 (+/-0.016) for {'subsample': 0.49568248279310045, 'n_estimators': 120, 'max_depth': 80, 'learning_rate': 0.3, 'colsample_bytree': 0.8691889161717675}\n",
      "f1_weighted = 0.738 (+/-0.015) for {'subsample': 0.7627155014500968, 'n_estimators': 200, 'max_depth': 80, 'learning_rate': 0.001, 'colsample_bytree': 0.8691889161717675}\n",
      "f1_weighted = 0.729 (+/-0.015) for {'subsample': 0.49568248279310045, 'n_estimators': 180, 'max_depth': 20, 'learning_rate': 0.01, 'colsample_bytree': 0.559635577743145}\n",
      "f1_weighted = 0.738 (+/-0.019) for {'subsample': 0.4277045805192762, 'n_estimators': 140, 'max_depth': 40, 'learning_rate': 0.003, 'colsample_bytree': 0.8691889161717675}\n",
      "f1_weighted = 0.734 (+/-0.017) for {'subsample': 0.7627155014500968, 'n_estimators': 80, 'max_depth': 20, 'learning_rate': 0.1, 'colsample_bytree': 0.7582357629850743}\n",
      "f1_weighted = 0.736 (+/-0.019) for {'subsample': 0.4277045805192762, 'n_estimators': 140, 'max_depth': 100, 'learning_rate': 0.003, 'colsample_bytree': 0.7582357629850743}\n",
      "f1_weighted = 0.731 (+/-0.014) for {'subsample': 0.4277045805192762, 'n_estimators': 60, 'max_depth': 20, 'learning_rate': 0.03, 'colsample_bytree': 0.559635577743145}\n",
      "f1_weighted = 0.724 (+/-0.015) for {'subsample': 0.49568248279310045, 'n_estimators': 100, 'max_depth': 80, 'learning_rate': 0.3, 'colsample_bytree': 0.559635577743145}\n",
      "f1_weighted = 0.734 (+/-0.014) for {'subsample': 0.49568248279310045, 'n_estimators': 40, 'max_depth': 80, 'learning_rate': 0.1, 'colsample_bytree': 0.7582357629850743}\n",
      "f1_weighted = 0.737 (+/-0.017) for {'subsample': 0.7627155014500968, 'n_estimators': 100, 'max_depth': 100, 'learning_rate': 0.03, 'colsample_bytree': 0.7582357629850743}\n",
      "f1_weighted = 0.737 (+/-0.015) for {'subsample': 0.49568248279310045, 'n_estimators': 180, 'max_depth': 40, 'learning_rate': 0.001, 'colsample_bytree': 0.7582357629850743}\n",
      "f1_weighted = 0.738 (+/-0.015) for {'subsample': 0.7627155014500968, 'n_estimators': 200, 'max_depth': 20, 'learning_rate': 0.001, 'colsample_bytree': 0.8691889161717675}\n",
      "f1_weighted = 0.729 (+/-0.010) for {'subsample': 0.7627155014500968, 'n_estimators': 60, 'max_depth': 60, 'learning_rate': 0.03, 'colsample_bytree': 0.559635577743145}\n"
     ]
    }
   ],
   "source": [
    "# Afficher les performances correspondante\n",
    "print(\"Résultats de la validation croisée :\")\n",
    "for mean, std, params in zip(\n",
    "        xgb_random.cv_results_['mean_test_score'], # score moyen\n",
    "        xgb_random.cv_results_['std_test_score'],  # écart-type du score\n",
    "        xgb_random.cv_results_['params']           # valeur de l'hyperparamètre\n",
    "    ):\n",
    "\n",
    "    print(\"{} = {:.3f} (+/-{:.03f}) for {}\".format(\n",
    "        score,\n",
    "        mean,\n",
    "        std*2,\n",
    "        params\n",
    "    ) )"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e7e13ac1",
   "metadata": {},
   "source": [
    "**Calcul de l'efficacité moyenne du modèle par cross-validation (validation croisée)**\n",
    "\n",
    "Puis on l'affiche : (73.5%)\n",
    "\n",
    "Le warning s'affiche car `cross_validate` appelle la fonction `fit` du classifier `xgb_random` sans lui passer `eval_metric = 'mlogloss'` comme paramètre."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "812dc286",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 50 candidates, totalling 250 fits\n",
      "[03:23:35] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.3.0/src/learner.cc:1061: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'multi:softprob' was changed from 'merror' to 'mlogloss'. Explicitly set eval_metric if you'd like to restore the old behavior.\n",
      "Fitting 5 folds for each of 50 candidates, totalling 250 fits\n",
      "[03:32:37] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.3.0/src/learner.cc:1061: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'multi:softprob' was changed from 'merror' to 'mlogloss'. Explicitly set eval_metric if you'd like to restore the old behavior.\n",
      "Fitting 5 folds for each of 50 candidates, totalling 250 fits\n",
      "[03:42:19] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.3.0/src/learner.cc:1061: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'multi:softprob' was changed from 'merror' to 'mlogloss'. Explicitly set eval_metric if you'd like to restore the old behavior.\n",
      "Cross-validation performance : 0.735\n"
     ]
    }
   ],
   "source": [
    "scores_xgb_random = cross_validate(xgb_random, X_train, y_train, cv=3,\n",
    "                    scoring=('precision_weighted','recall_weighted','balanced_accuracy','f1_weighted','roc_auc_ovr_weighted'))\n",
    "print(\"Cross-validation performance : {:.3f}\".format(scores_xgb_random['test_f1_weighted'].mean()))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a21b17de",
   "metadata": {},
   "source": [
    "Le modèle XGBoost ne fait presque pas mieux qu'un modèle Random Forest. Il se peut que les hyperparamètres nécessitent plus de tuning, ou qu'il faut plus de données (mais pour entraîner ce modèle il nous faut déjà environ 15 minutes avec nos capacités computationnelles) ou que le modèle trouve du mal à approcher la fonction de décision car les données sont proches et qu'il risque l'**overfitting**."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c2348db3",
   "metadata": {},
   "source": [
    "## 3.4 Artificial Neural Network :\n",
    "\n",
    "On essaye maintenant un **réseau de neurones artificiel** en espérant avoir de meilleures performances (on avait implémenté un Logistic Regressor avant de passer au réseau de neurones mais il n'a pas fait mieux que le KNN, on a donc supprimé les cellules).\n",
    "\n",
    "Il faudrait avoir la bibliothèque `Keras` installée sur sa machine pour pouvoir exécuter les cellules suivantes.\n",
    "\n",
    "On utilise un réseau de neurones avec deux *hidden layers* avec des outputs de 16 et 32 dimensions respectivement avec comme fonction d'ativation **relu**, suivies de la dernière couche (output layer) avec du **softmax** comme c'est du multiclasses.\n",
    "\n",
    "On a dû implémenter une fonction `f1_weighted` pour la passer comme paramètre en tant que `metric` à la fonction `compile` du model.\n",
    "\n",
    "On crée ensuite un **wrapper** sklearn autour de notre modèle Keras. On l'entraîne sur notre jeu de données puis on lui applique de la validation croisée.\n",
    "\n",
    "Les hyperparamètres comme le *nombre d'epochs*, *batch_size*, le nombre de *hidden layers*, le *nombre de neurones dans chaque layer* etc. ne sont pas tunés comme il est fastidieux de retrouver les bons paramètres."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "5b2f69da",
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras import backend as K\n",
    "\n",
    "def f1_weighted(true, pred):\n",
    "\n",
    "    predLabels = K.argmax(pred, axis=-1)\n",
    "    pred = K.one_hot(predLabels, 1) \n",
    "\n",
    "\n",
    "    ground_positives = K.sum(true, axis=0) + K.epsilon()       # = TP + FN\n",
    "    pred_positives = K.sum(pred, axis=0) + K.epsilon()         # = TP + FP\n",
    "    true_positives = K.sum(true * pred, axis=0) + K.epsilon()  # = TP\n",
    "    \n",
    "    precision = true_positives / pred_positives \n",
    "    recall = true_positives / ground_positives\n",
    "        #both = 1 if ground_positives == 0 or pred_positives == 0\n",
    "\n",
    "    f1 = 2 * (precision * recall) / (precision + recall + K.epsilon())\n",
    "\n",
    "    weighted_f1 = f1 * ground_positives / K.sum(ground_positives) \n",
    "    weighted_f1 = K.sum(weighted_f1)\n",
    "\n",
    "    \n",
    "    return 1 - weighted_f1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "id": "81203046",
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.layers import InputLayer\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "from keras.wrappers.scikit_learn import KerasClassifier\n",
    "from keras.utils import np_utils\n",
    "\n",
    "def baseline_model():\n",
    "    model = Sequential()\n",
    "    model.add(Dense(16, input_dim=10, activation='relu'))\n",
    "    model.add(Dense(32, activation='relu'))\n",
    "    model.add(Dense(3, activation='softmax'))\n",
    "    model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=[f1_weighted, 'accuracy'])\n",
    "    return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "fd9bb579",
   "metadata": {},
   "outputs": [],
   "source": [
    "ann = KerasClassifier(build_fn=baseline_model, epochs=200, batch_size=200)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "id": "38b46549",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "70/70 [==============================] - 1s 1ms/step - loss: 1.0474 - f1_weighted: 0.7905 - accuracy: 0.4789\n",
      "Epoch 2/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.7486 - f1_weighted: 0.7537 - accuracy: 0.6937\n",
      "Epoch 3/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6761 - f1_weighted: 0.7657 - accuracy: 0.7179\n",
      "Epoch 4/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6567 - f1_weighted: 0.7657 - accuracy: 0.7210\n",
      "Epoch 5/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6525 - f1_weighted: 0.7645 - accuracy: 0.7233\n",
      "Epoch 6/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6550 - f1_weighted: 0.7672 - accuracy: 0.7252\n",
      "Epoch 7/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6315 - f1_weighted: 0.7615 - accuracy: 0.7331\n",
      "Epoch 8/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6416 - f1_weighted: 0.7628 - accuracy: 0.7269\n",
      "Epoch 9/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6328 - f1_weighted: 0.7689 - accuracy: 0.7345\n",
      "Epoch 10/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6464 - f1_weighted: 0.7659 - accuracy: 0.7272\n",
      "Epoch 11/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6383 - f1_weighted: 0.7720 - accuracy: 0.7277\n",
      "Epoch 12/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6318 - f1_weighted: 0.7610 - accuracy: 0.7322\n",
      "Epoch 13/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6344 - f1_weighted: 0.7586 - accuracy: 0.7281\n",
      "Epoch 14/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6379 - f1_weighted: 0.7651 - accuracy: 0.7282\n",
      "Epoch 15/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6398 - f1_weighted: 0.7701 - accuracy: 0.7258\n",
      "Epoch 16/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6321 - f1_weighted: 0.7682 - accuracy: 0.7332\n",
      "Epoch 17/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6369 - f1_weighted: 0.7661 - accuracy: 0.7283\n",
      "Epoch 18/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6296 - f1_weighted: 0.7673 - accuracy: 0.7336\n",
      "Epoch 19/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6239 - f1_weighted: 0.7652 - accuracy: 0.7350\n",
      "Epoch 20/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6254 - f1_weighted: 0.7676 - accuracy: 0.7351\n",
      "Epoch 21/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6306 - f1_weighted: 0.7693 - accuracy: 0.7323\n",
      "Epoch 22/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6231 - f1_weighted: 0.7625 - accuracy: 0.7325\n",
      "Epoch 23/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6309 - f1_weighted: 0.7663 - accuracy: 0.7296\n",
      "Epoch 24/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6309 - f1_weighted: 0.7707 - accuracy: 0.7301\n",
      "Epoch 25/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6314 - f1_weighted: 0.7649 - accuracy: 0.7315\n",
      "Epoch 26/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6154 - f1_weighted: 0.7632 - accuracy: 0.7394\n",
      "Epoch 27/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6263 - f1_weighted: 0.7634 - accuracy: 0.7295\n",
      "Epoch 28/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6184 - f1_weighted: 0.7733 - accuracy: 0.7418\n",
      "Epoch 29/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6161 - f1_weighted: 0.7642 - accuracy: 0.7413\n",
      "Epoch 30/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6222 - f1_weighted: 0.7617 - accuracy: 0.7331\n",
      "Epoch 31/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6183 - f1_weighted: 0.7655 - accuracy: 0.7399\n",
      "Epoch 32/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6247 - f1_weighted: 0.7674 - accuracy: 0.7321\n",
      "Epoch 33/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6247 - f1_weighted: 0.7662 - accuracy: 0.7352\n",
      "Epoch 34/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6205 - f1_weighted: 0.7677 - accuracy: 0.7356\n",
      "Epoch 35/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6166 - f1_weighted: 0.7732 - accuracy: 0.7399\n",
      "Epoch 36/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6187 - f1_weighted: 0.7674 - accuracy: 0.7340\n",
      "Epoch 37/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6124 - f1_weighted: 0.7623 - accuracy: 0.7373\n",
      "Epoch 38/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6160 - f1_weighted: 0.7630 - accuracy: 0.7391\n",
      "Epoch 39/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6201 - f1_weighted: 0.7635 - accuracy: 0.7354\n",
      "Epoch 40/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6147 - f1_weighted: 0.7673 - accuracy: 0.7426\n",
      "Epoch 41/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6145 - f1_weighted: 0.7589 - accuracy: 0.7368\n",
      "Epoch 42/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6005 - f1_weighted: 0.7707 - accuracy: 0.7464\n",
      "Epoch 43/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6155 - f1_weighted: 0.7623 - accuracy: 0.7358\n",
      "Epoch 44/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6149 - f1_weighted: 0.7642 - accuracy: 0.7370\n",
      "Epoch 45/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6125 - f1_weighted: 0.7671 - accuracy: 0.7420\n",
      "Epoch 46/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6062 - f1_weighted: 0.7613 - accuracy: 0.7387\n",
      "Epoch 47/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6133 - f1_weighted: 0.7680 - accuracy: 0.7387\n",
      "Epoch 48/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6151 - f1_weighted: 0.7582 - accuracy: 0.7374\n",
      "Epoch 49/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6140 - f1_weighted: 0.7617 - accuracy: 0.7348\n",
      "Epoch 50/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6005 - f1_weighted: 0.7689 - accuracy: 0.7459\n",
      "Epoch 51/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6091 - f1_weighted: 0.7624 - accuracy: 0.7395\n",
      "Epoch 52/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6090 - f1_weighted: 0.7586 - accuracy: 0.7425\n",
      "Epoch 53/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6073 - f1_weighted: 0.7669 - accuracy: 0.7423\n",
      "Epoch 54/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6059 - f1_weighted: 0.7583 - accuracy: 0.7447\n",
      "Epoch 55/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6049 - f1_weighted: 0.7609 - accuracy: 0.7429\n",
      "Epoch 56/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6062 - f1_weighted: 0.7639 - accuracy: 0.7425\n",
      "Epoch 57/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6109 - f1_weighted: 0.7597 - accuracy: 0.7410\n",
      "Epoch 58/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6122 - f1_weighted: 0.7592 - accuracy: 0.7409\n",
      "Epoch 59/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6074 - f1_weighted: 0.7641 - accuracy: 0.7408\n",
      "Epoch 60/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6066 - f1_weighted: 0.7630 - accuracy: 0.7445\n",
      "Epoch 61/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6095 - f1_weighted: 0.7557 - accuracy: 0.7407\n",
      "Epoch 62/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6112 - f1_weighted: 0.7558 - accuracy: 0.7373\n",
      "Epoch 63/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6101 - f1_weighted: 0.7660 - accuracy: 0.7465\n",
      "Epoch 64/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6109 - f1_weighted: 0.7656 - accuracy: 0.7406\n",
      "Epoch 65/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6162 - f1_weighted: 0.7607 - accuracy: 0.7348\n",
      "Epoch 66/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6128 - f1_weighted: 0.7643 - accuracy: 0.7374\n",
      "Epoch 67/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6109 - f1_weighted: 0.7624 - accuracy: 0.7414\n",
      "Epoch 68/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5979 - f1_weighted: 0.7658 - accuracy: 0.7481\n",
      "Epoch 69/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6057 - f1_weighted: 0.7570 - accuracy: 0.7439\n",
      "Epoch 70/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5966 - f1_weighted: 0.7618 - accuracy: 0.7462\n",
      "Epoch 71/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6040 - f1_weighted: 0.7659 - accuracy: 0.7481\n",
      "Epoch 72/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5982 - f1_weighted: 0.7666 - accuracy: 0.7494\n",
      "Epoch 73/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6093 - f1_weighted: 0.7626 - accuracy: 0.7430\n",
      "Epoch 74/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6037 - f1_weighted: 0.7596 - accuracy: 0.7425\n",
      "Epoch 75/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5954 - f1_weighted: 0.7639 - accuracy: 0.7476\n",
      "Epoch 76/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6013 - f1_weighted: 0.7631 - accuracy: 0.7439\n",
      "Epoch 77/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6115 - f1_weighted: 0.7628 - accuracy: 0.7406\n",
      "Epoch 78/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6087 - f1_weighted: 0.7668 - accuracy: 0.7417\n",
      "Epoch 79/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6039 - f1_weighted: 0.7637 - accuracy: 0.7454\n",
      "Epoch 80/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6068 - f1_weighted: 0.7588 - accuracy: 0.7423\n",
      "Epoch 81/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5967 - f1_weighted: 0.7622 - accuracy: 0.7433\n",
      "Epoch 82/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5979 - f1_weighted: 0.7659 - accuracy: 0.7473\n",
      "Epoch 83/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6017 - f1_weighted: 0.7582 - accuracy: 0.7440\n",
      "Epoch 84/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6130 - f1_weighted: 0.7592 - accuracy: 0.7340\n",
      "Epoch 85/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6025 - f1_weighted: 0.7513 - accuracy: 0.7433\n",
      "Epoch 86/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6021 - f1_weighted: 0.7598 - accuracy: 0.7423\n",
      "Epoch 87/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6075 - f1_weighted: 0.7571 - accuracy: 0.7417\n",
      "Epoch 88/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6016 - f1_weighted: 0.7570 - accuracy: 0.7451\n",
      "Epoch 89/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5935 - f1_weighted: 0.7613 - accuracy: 0.7490\n",
      "Epoch 90/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5998 - f1_weighted: 0.7603 - accuracy: 0.7418\n",
      "Epoch 91/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6068 - f1_weighted: 0.7563 - accuracy: 0.7400\n",
      "Epoch 92/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6006 - f1_weighted: 0.7501 - accuracy: 0.7423\n",
      "Epoch 93/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6058 - f1_weighted: 0.7627 - accuracy: 0.7399\n",
      "Epoch 94/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5916 - f1_weighted: 0.7538 - accuracy: 0.7487\n",
      "Epoch 95/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6006 - f1_weighted: 0.7608 - accuracy: 0.7461\n",
      "Epoch 96/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5955 - f1_weighted: 0.7595 - accuracy: 0.7488\n",
      "Epoch 97/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6013 - f1_weighted: 0.7552 - accuracy: 0.7474\n",
      "Epoch 98/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6008 - f1_weighted: 0.7616 - accuracy: 0.7445\n",
      "Epoch 99/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5996 - f1_weighted: 0.7557 - accuracy: 0.7441\n",
      "Epoch 100/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5938 - f1_weighted: 0.7568 - accuracy: 0.7477\n",
      "Epoch 101/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5964 - f1_weighted: 0.7520 - accuracy: 0.7434\n",
      "Epoch 102/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5990 - f1_weighted: 0.7584 - accuracy: 0.7430\n",
      "Epoch 103/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6053 - f1_weighted: 0.7572 - accuracy: 0.7414\n",
      "Epoch 104/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5971 - f1_weighted: 0.7593 - accuracy: 0.7437\n",
      "Epoch 105/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5959 - f1_weighted: 0.7667 - accuracy: 0.7508\n",
      "Epoch 106/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5946 - f1_weighted: 0.7664 - accuracy: 0.7498\n",
      "Epoch 107/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6009 - f1_weighted: 0.7641 - accuracy: 0.7439\n",
      "Epoch 108/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6022 - f1_weighted: 0.7523 - accuracy: 0.7432\n",
      "Epoch 109/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5961 - f1_weighted: 0.7529 - accuracy: 0.7448\n",
      "Epoch 110/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5965 - f1_weighted: 0.7561 - accuracy: 0.7458\n",
      "Epoch 111/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6039 - f1_weighted: 0.7547 - accuracy: 0.7405\n",
      "Epoch 112/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5955 - f1_weighted: 0.7483 - accuracy: 0.7480\n",
      "Epoch 113/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6001 - f1_weighted: 0.7597 - accuracy: 0.7426\n",
      "Epoch 114/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5902 - f1_weighted: 0.7548 - accuracy: 0.7483\n",
      "Epoch 115/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5891 - f1_weighted: 0.7655 - accuracy: 0.7487\n",
      "Epoch 116/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5877 - f1_weighted: 0.7537 - accuracy: 0.7504\n",
      "Epoch 117/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6021 - f1_weighted: 0.7557 - accuracy: 0.7424\n",
      "Epoch 118/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5980 - f1_weighted: 0.7536 - accuracy: 0.7441\n",
      "Epoch 119/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6003 - f1_weighted: 0.7563 - accuracy: 0.7434\n",
      "Epoch 120/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5929 - f1_weighted: 0.7595 - accuracy: 0.7506\n",
      "Epoch 121/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5949 - f1_weighted: 0.7571 - accuracy: 0.7450\n",
      "Epoch 122/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5898 - f1_weighted: 0.7589 - accuracy: 0.7475\n",
      "Epoch 123/200\n",
      "70/70 [==============================] - 0s 906us/step - loss: 0.6004 - f1_weighted: 0.7569 - accuracy: 0.7454\n",
      "Epoch 124/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5988 - f1_weighted: 0.7534 - accuracy: 0.7428\n",
      "Epoch 125/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5912 - f1_weighted: 0.7522 - accuracy: 0.7453\n",
      "Epoch 126/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5988 - f1_weighted: 0.7552 - accuracy: 0.7436\n",
      "Epoch 127/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6004 - f1_weighted: 0.7561 - accuracy: 0.7416\n",
      "Epoch 128/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5947 - f1_weighted: 0.7547 - accuracy: 0.7458\n",
      "Epoch 129/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6037 - f1_weighted: 0.7592 - accuracy: 0.7420\n",
      "Epoch 130/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5990 - f1_weighted: 0.7531 - accuracy: 0.7440\n",
      "Epoch 131/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6143 - f1_weighted: 0.7547 - accuracy: 0.7331\n",
      "Epoch 132/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5981 - f1_weighted: 0.7614 - accuracy: 0.7444\n",
      "Epoch 133/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5905 - f1_weighted: 0.7492 - accuracy: 0.7453\n",
      "Epoch 134/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6047 - f1_weighted: 0.7619 - accuracy: 0.7375\n",
      "Epoch 135/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5948 - f1_weighted: 0.7596 - accuracy: 0.7460\n",
      "Epoch 136/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6012 - f1_weighted: 0.7552 - accuracy: 0.7421\n",
      "Epoch 137/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5951 - f1_weighted: 0.7540 - accuracy: 0.7446\n",
      "Epoch 138/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6049 - f1_weighted: 0.7522 - accuracy: 0.7399\n",
      "Epoch 139/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6029 - f1_weighted: 0.7569 - accuracy: 0.7419\n",
      "Epoch 140/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6017 - f1_weighted: 0.7564 - accuracy: 0.7402\n",
      "Epoch 141/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5962 - f1_weighted: 0.7572 - accuracy: 0.7453\n",
      "Epoch 142/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5976 - f1_weighted: 0.7608 - accuracy: 0.7447\n",
      "Epoch 143/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6040 - f1_weighted: 0.7570 - accuracy: 0.7377\n",
      "Epoch 144/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5896 - f1_weighted: 0.7501 - accuracy: 0.7486\n",
      "Epoch 145/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5954 - f1_weighted: 0.7620 - accuracy: 0.7435\n",
      "Epoch 146/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5894 - f1_weighted: 0.7538 - accuracy: 0.7474\n",
      "Epoch 147/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6004 - f1_weighted: 0.7542 - accuracy: 0.7431\n",
      "Epoch 148/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5906 - f1_weighted: 0.7564 - accuracy: 0.7485\n",
      "Epoch 149/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5925 - f1_weighted: 0.7540 - accuracy: 0.7413\n",
      "Epoch 150/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5978 - f1_weighted: 0.7579 - accuracy: 0.7448\n",
      "Epoch 151/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6040 - f1_weighted: 0.7466 - accuracy: 0.7389\n",
      "Epoch 152/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5906 - f1_weighted: 0.7569 - accuracy: 0.7455\n",
      "Epoch 153/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5920 - f1_weighted: 0.7542 - accuracy: 0.7473\n",
      "Epoch 154/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5950 - f1_weighted: 0.7561 - accuracy: 0.7455\n",
      "Epoch 155/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5838 - f1_weighted: 0.7542 - accuracy: 0.7493\n",
      "Epoch 156/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6087 - f1_weighted: 0.7548 - accuracy: 0.7339\n",
      "Epoch 157/200\n",
      "70/70 [==============================] - 0s 906us/step - loss: 0.5970 - f1_weighted: 0.7546 - accuracy: 0.7419\n",
      "Epoch 158/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5974 - f1_weighted: 0.7486 - accuracy: 0.7397\n",
      "Epoch 159/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5920 - f1_weighted: 0.7592 - accuracy: 0.7467\n",
      "Epoch 160/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5963 - f1_weighted: 0.7525 - accuracy: 0.7420\n",
      "Epoch 161/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5903 - f1_weighted: 0.7528 - accuracy: 0.7451\n",
      "Epoch 162/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5957 - f1_weighted: 0.7508 - accuracy: 0.7400\n",
      "Epoch 163/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6016 - f1_weighted: 0.7473 - accuracy: 0.7396\n",
      "Epoch 164/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5888 - f1_weighted: 0.7554 - accuracy: 0.7451\n",
      "Epoch 165/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5885 - f1_weighted: 0.7545 - accuracy: 0.7492\n",
      "Epoch 166/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5906 - f1_weighted: 0.7649 - accuracy: 0.7437\n",
      "Epoch 167/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5867 - f1_weighted: 0.7518 - accuracy: 0.7466\n",
      "Epoch 168/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5943 - f1_weighted: 0.7498 - accuracy: 0.7443\n",
      "Epoch 169/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5952 - f1_weighted: 0.7546 - accuracy: 0.7417\n",
      "Epoch 170/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5900 - f1_weighted: 0.7518 - accuracy: 0.7447\n",
      "Epoch 171/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5904 - f1_weighted: 0.7512 - accuracy: 0.7440\n",
      "Epoch 172/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5880 - f1_weighted: 0.7551 - accuracy: 0.7484\n",
      "Epoch 173/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5907 - f1_weighted: 0.7491 - accuracy: 0.7453\n",
      "Epoch 174/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5907 - f1_weighted: 0.7589 - accuracy: 0.7479\n",
      "Epoch 175/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5911 - f1_weighted: 0.7560 - accuracy: 0.7431\n",
      "Epoch 176/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5944 - f1_weighted: 0.7583 - accuracy: 0.7497\n",
      "Epoch 177/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5910 - f1_weighted: 0.7560 - accuracy: 0.7488\n",
      "Epoch 178/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5954 - f1_weighted: 0.7538 - accuracy: 0.7415\n",
      "Epoch 179/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5893 - f1_weighted: 0.7458 - accuracy: 0.7469\n",
      "Epoch 180/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5879 - f1_weighted: 0.7540 - accuracy: 0.7452\n",
      "Epoch 181/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6006 - f1_weighted: 0.7454 - accuracy: 0.7377\n",
      "Epoch 182/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6003 - f1_weighted: 0.7557 - accuracy: 0.7400\n",
      "Epoch 183/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5834 - f1_weighted: 0.7495 - accuracy: 0.7516\n",
      "Epoch 184/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5866 - f1_weighted: 0.7530 - accuracy: 0.7472\n",
      "Epoch 185/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5972 - f1_weighted: 0.7540 - accuracy: 0.7397\n",
      "Epoch 186/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5924 - f1_weighted: 0.7539 - accuracy: 0.7465\n",
      "Epoch 187/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5875 - f1_weighted: 0.7474 - accuracy: 0.7490\n",
      "Epoch 188/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5871 - f1_weighted: 0.7482 - accuracy: 0.7476\n",
      "Epoch 189/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5880 - f1_weighted: 0.7547 - accuracy: 0.7447\n",
      "Epoch 190/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5946 - f1_weighted: 0.7519 - accuracy: 0.7429\n",
      "Epoch 191/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5876 - f1_weighted: 0.7579 - accuracy: 0.7471\n",
      "Epoch 192/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5849 - f1_weighted: 0.7459 - accuracy: 0.7468\n",
      "Epoch 193/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5902 - f1_weighted: 0.7563 - accuracy: 0.7498\n",
      "Epoch 194/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5913 - f1_weighted: 0.7484 - accuracy: 0.7428\n",
      "Epoch 195/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5917 - f1_weighted: 0.7543 - accuracy: 0.7422\n",
      "Epoch 196/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5854 - f1_weighted: 0.7510 - accuracy: 0.7491\n",
      "Epoch 197/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5910 - f1_weighted: 0.7476 - accuracy: 0.7439\n",
      "Epoch 198/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5878 - f1_weighted: 0.7512 - accuracy: 0.7423\n",
      "Epoch 199/200\n",
      "70/70 [==============================] - 0s 1ms/step - loss: 0.6008 - f1_weighted: 0.7535 - accuracy: 0.7402\n",
      "Epoch 200/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "70/70 [==============================] - 0s 1ms/step - loss: 0.5783 - f1_weighted: 0.7519 - accuracy: 0.7555\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x1f3537d15b0>"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Execution du modèle\n",
    "ann.fit(X_train, y_train, epochs=200, batch_size=200)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "id": "909d0361",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "56/56 [==============================] - 1s 1ms/step - loss: 1.1104 - f1_weighted: 0.4679 - accuracy: 0.3907\n",
      "Epoch 2/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.8004 - f1_weighted: 0.7455 - accuracy: 0.6773\n",
      "Epoch 3/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6966 - f1_weighted: 0.7634 - accuracy: 0.7120\n",
      "Epoch 4/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6750 - f1_weighted: 0.7631 - accuracy: 0.7147\n",
      "Epoch 5/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6549 - f1_weighted: 0.7635 - accuracy: 0.7257\n",
      "Epoch 6/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6381 - f1_weighted: 0.7615 - accuracy: 0.7337\n",
      "Epoch 7/200\n",
      "56/56 [==============================] - 0s 853us/step - loss: 0.6480 - f1_weighted: 0.7683 - accuracy: 0.7220\n",
      "Epoch 8/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6414 - f1_weighted: 0.7613 - accuracy: 0.7280\n",
      "Epoch 9/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6354 - f1_weighted: 0.7590 - accuracy: 0.7328\n",
      "Epoch 10/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6531 - f1_weighted: 0.7669 - accuracy: 0.7185\n",
      "Epoch 11/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6352 - f1_weighted: 0.7670 - accuracy: 0.7292\n",
      "Epoch 12/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6261 - f1_weighted: 0.7678 - accuracy: 0.7351\n",
      "Epoch 13/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6322 - f1_weighted: 0.7711 - accuracy: 0.7317\n",
      "Epoch 14/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6445 - f1_weighted: 0.7647 - accuracy: 0.7280\n",
      "Epoch 15/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6315 - f1_weighted: 0.7619 - accuracy: 0.7268\n",
      "Epoch 16/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6379 - f1_weighted: 0.7648 - accuracy: 0.7307\n",
      "Epoch 17/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6343 - f1_weighted: 0.7661 - accuracy: 0.7319\n",
      "Epoch 18/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6397 - f1_weighted: 0.7655 - accuracy: 0.7289\n",
      "Epoch 19/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6300 - f1_weighted: 0.7616 - accuracy: 0.7342\n",
      "Epoch 20/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6322 - f1_weighted: 0.7667 - accuracy: 0.7313\n",
      "Epoch 21/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6310 - f1_weighted: 0.7698 - accuracy: 0.7334\n",
      "Epoch 22/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6417 - f1_weighted: 0.7641 - accuracy: 0.7232\n",
      "Epoch 23/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6221 - f1_weighted: 0.7682 - accuracy: 0.7383\n",
      "Epoch 24/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6279 - f1_weighted: 0.7721 - accuracy: 0.7347\n",
      "Epoch 25/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6359 - f1_weighted: 0.7672 - accuracy: 0.7311\n",
      "Epoch 26/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6201 - f1_weighted: 0.7677 - accuracy: 0.7397\n",
      "Epoch 27/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6274 - f1_weighted: 0.7654 - accuracy: 0.7329\n",
      "Epoch 28/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6188 - f1_weighted: 0.7638 - accuracy: 0.7373\n",
      "Epoch 29/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6160 - f1_weighted: 0.7683 - accuracy: 0.7386\n",
      "Epoch 30/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6229 - f1_weighted: 0.7723 - accuracy: 0.7309\n",
      "Epoch 31/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6288 - f1_weighted: 0.7656 - accuracy: 0.7338\n",
      "Epoch 32/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6094 - f1_weighted: 0.7711 - accuracy: 0.7442\n",
      "Epoch 33/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6191 - f1_weighted: 0.7585 - accuracy: 0.7388\n",
      "Epoch 34/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6208 - f1_weighted: 0.7624 - accuracy: 0.7367\n",
      "Epoch 35/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6243 - f1_weighted: 0.7655 - accuracy: 0.7316\n",
      "Epoch 36/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6194 - f1_weighted: 0.7625 - accuracy: 0.7420\n",
      "Epoch 37/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6240 - f1_weighted: 0.7624 - accuracy: 0.7337\n",
      "Epoch 38/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6280 - f1_weighted: 0.7666 - accuracy: 0.7343\n",
      "Epoch 39/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6107 - f1_weighted: 0.7618 - accuracy: 0.7360\n",
      "Epoch 40/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6197 - f1_weighted: 0.7626 - accuracy: 0.7376\n",
      "Epoch 41/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6225 - f1_weighted: 0.7719 - accuracy: 0.7357\n",
      "Epoch 42/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6286 - f1_weighted: 0.7685 - accuracy: 0.7313\n",
      "Epoch 43/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6145 - f1_weighted: 0.7736 - accuracy: 0.7423\n",
      "Epoch 44/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6271 - f1_weighted: 0.7682 - accuracy: 0.7301\n",
      "Epoch 45/200\n",
      "56/56 [==============================] - 0s 944us/step - loss: 0.6145 - f1_weighted: 0.7609 - accuracy: 0.7371\n",
      "Epoch 46/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6205 - f1_weighted: 0.7617 - accuracy: 0.7320\n",
      "Epoch 47/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6135 - f1_weighted: 0.7716 - accuracy: 0.7390\n",
      "Epoch 48/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6163 - f1_weighted: 0.7733 - accuracy: 0.7385\n",
      "Epoch 49/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6284 - f1_weighted: 0.7628 - accuracy: 0.7310\n",
      "Epoch 50/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6193 - f1_weighted: 0.7654 - accuracy: 0.7387\n",
      "Epoch 51/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6211 - f1_weighted: 0.7631 - accuracy: 0.7390\n",
      "Epoch 52/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6221 - f1_weighted: 0.7696 - accuracy: 0.7340\n",
      "Epoch 53/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6288 - f1_weighted: 0.7699 - accuracy: 0.7302\n",
      "Epoch 54/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6151 - f1_weighted: 0.7723 - accuracy: 0.7445\n",
      "Epoch 55/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6006 - f1_weighted: 0.7606 - accuracy: 0.7444\n",
      "Epoch 56/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6207 - f1_weighted: 0.7607 - accuracy: 0.7333\n",
      "Epoch 57/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6156 - f1_weighted: 0.7661 - accuracy: 0.7351\n",
      "Epoch 58/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6140 - f1_weighted: 0.7639 - accuracy: 0.7350\n",
      "Epoch 59/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6045 - f1_weighted: 0.7623 - accuracy: 0.7454\n",
      "Epoch 60/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6049 - f1_weighted: 0.7569 - accuracy: 0.7434\n",
      "Epoch 61/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6092 - f1_weighted: 0.7629 - accuracy: 0.7397\n",
      "Epoch 62/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6180 - f1_weighted: 0.7623 - accuracy: 0.7332\n",
      "Epoch 63/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6184 - f1_weighted: 0.7638 - accuracy: 0.7336\n",
      "Epoch 64/200\n",
      "56/56 [==============================] - 0s 962us/step - loss: 0.6030 - f1_weighted: 0.7598 - accuracy: 0.7398\n",
      "Epoch 65/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6158 - f1_weighted: 0.7671 - accuracy: 0.7382\n",
      "Epoch 66/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6045 - f1_weighted: 0.7685 - accuracy: 0.7456\n",
      "Epoch 67/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6119 - f1_weighted: 0.7567 - accuracy: 0.7394\n",
      "Epoch 68/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6074 - f1_weighted: 0.7693 - accuracy: 0.7425\n",
      "Epoch 69/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6034 - f1_weighted: 0.7648 - accuracy: 0.7394\n",
      "Epoch 70/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6014 - f1_weighted: 0.7603 - accuracy: 0.7420\n",
      "Epoch 71/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6091 - f1_weighted: 0.7551 - accuracy: 0.7404\n",
      "Epoch 72/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6070 - f1_weighted: 0.7609 - accuracy: 0.7415\n",
      "Epoch 73/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6051 - f1_weighted: 0.7626 - accuracy: 0.7414\n",
      "Epoch 74/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6113 - f1_weighted: 0.7618 - accuracy: 0.7389\n",
      "Epoch 75/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6037 - f1_weighted: 0.7558 - accuracy: 0.7414\n",
      "Epoch 76/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6039 - f1_weighted: 0.7671 - accuracy: 0.7442\n",
      "Epoch 77/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6175 - f1_weighted: 0.7592 - accuracy: 0.7348\n",
      "Epoch 78/200\n",
      "56/56 [==============================] - 0s 944us/step - loss: 0.6056 - f1_weighted: 0.7555 - accuracy: 0.7410\n",
      "Epoch 79/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6057 - f1_weighted: 0.7506 - accuracy: 0.7396\n",
      "Epoch 80/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6082 - f1_weighted: 0.7617 - accuracy: 0.7428\n",
      "Epoch 81/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6037 - f1_weighted: 0.7577 - accuracy: 0.7420\n",
      "Epoch 82/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6049 - f1_weighted: 0.7679 - accuracy: 0.7382\n",
      "Epoch 83/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5970 - f1_weighted: 0.7596 - accuracy: 0.7453\n",
      "Epoch 84/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6077 - f1_weighted: 0.7564 - accuracy: 0.7376\n",
      "Epoch 85/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6118 - f1_weighted: 0.7654 - accuracy: 0.7351\n",
      "Epoch 86/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6037 - f1_weighted: 0.7586 - accuracy: 0.7420\n",
      "Epoch 87/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6061 - f1_weighted: 0.7532 - accuracy: 0.7368\n",
      "Epoch 88/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5994 - f1_weighted: 0.7665 - accuracy: 0.7461\n",
      "Epoch 89/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6025 - f1_weighted: 0.7571 - accuracy: 0.7404\n",
      "Epoch 90/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6083 - f1_weighted: 0.7547 - accuracy: 0.7409\n",
      "Epoch 91/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5916 - f1_weighted: 0.7535 - accuracy: 0.7464\n",
      "Epoch 92/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5997 - f1_weighted: 0.7615 - accuracy: 0.7424\n",
      "Epoch 93/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6060 - f1_weighted: 0.7542 - accuracy: 0.7377\n",
      "Epoch 94/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5986 - f1_weighted: 0.7667 - accuracy: 0.7457\n",
      "Epoch 95/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5939 - f1_weighted: 0.7547 - accuracy: 0.7454\n",
      "Epoch 96/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6135 - f1_weighted: 0.7571 - accuracy: 0.7397\n",
      "Epoch 97/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5952 - f1_weighted: 0.7553 - accuracy: 0.7437\n",
      "Epoch 98/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6000 - f1_weighted: 0.7613 - accuracy: 0.7451\n",
      "Epoch 99/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5942 - f1_weighted: 0.7587 - accuracy: 0.7453\n",
      "Epoch 100/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5922 - f1_weighted: 0.7597 - accuracy: 0.7468\n",
      "Epoch 101/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5975 - f1_weighted: 0.7550 - accuracy: 0.7473\n",
      "Epoch 102/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6012 - f1_weighted: 0.7569 - accuracy: 0.7408\n",
      "Epoch 103/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6084 - f1_weighted: 0.7563 - accuracy: 0.7430\n",
      "Epoch 104/200\n",
      "56/56 [==============================] - 0s 935us/step - loss: 0.6049 - f1_weighted: 0.7512 - accuracy: 0.7451\n",
      "Epoch 105/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6010 - f1_weighted: 0.7618 - accuracy: 0.7400\n",
      "Epoch 106/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5926 - f1_weighted: 0.7620 - accuracy: 0.7481\n",
      "Epoch 107/200\n",
      "56/56 [==============================] - 0s 964us/step - loss: 0.6035 - f1_weighted: 0.7559 - accuracy: 0.7391\n",
      "Epoch 108/200\n",
      "56/56 [==============================] - 0s 853us/step - loss: 0.5924 - f1_weighted: 0.7592 - accuracy: 0.7515\n",
      "Epoch 109/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6005 - f1_weighted: 0.7538 - accuracy: 0.7453\n",
      "Epoch 110/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5985 - f1_weighted: 0.7509 - accuracy: 0.7469\n",
      "Epoch 111/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6037 - f1_weighted: 0.7564 - accuracy: 0.7407\n",
      "Epoch 112/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6015 - f1_weighted: 0.7438 - accuracy: 0.7419\n",
      "Epoch 113/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5966 - f1_weighted: 0.7573 - accuracy: 0.7457\n",
      "Epoch 114/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6018 - f1_weighted: 0.7553 - accuracy: 0.7395\n",
      "Epoch 115/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5965 - f1_weighted: 0.7528 - accuracy: 0.7455\n",
      "Epoch 116/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5964 - f1_weighted: 0.7523 - accuracy: 0.7444\n",
      "Epoch 117/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5951 - f1_weighted: 0.7462 - accuracy: 0.7470\n",
      "Epoch 118/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5957 - f1_weighted: 0.7566 - accuracy: 0.7473\n",
      "Epoch 119/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5992 - f1_weighted: 0.7540 - accuracy: 0.7459\n",
      "Epoch 120/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5926 - f1_weighted: 0.7547 - accuracy: 0.7482\n",
      "Epoch 121/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6019 - f1_weighted: 0.7476 - accuracy: 0.7426\n",
      "Epoch 122/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5896 - f1_weighted: 0.7543 - accuracy: 0.7462\n",
      "Epoch 123/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5997 - f1_weighted: 0.7535 - accuracy: 0.7448\n",
      "Epoch 124/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6031 - f1_weighted: 0.7578 - accuracy: 0.7434\n",
      "Epoch 125/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6070 - f1_weighted: 0.7514 - accuracy: 0.7404\n",
      "Epoch 126/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6036 - f1_weighted: 0.7492 - accuracy: 0.7412\n",
      "Epoch 127/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5893 - f1_weighted: 0.7500 - accuracy: 0.7441\n",
      "Epoch 128/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5982 - f1_weighted: 0.7512 - accuracy: 0.7441\n",
      "Epoch 129/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5957 - f1_weighted: 0.7506 - accuracy: 0.7424\n",
      "Epoch 130/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5928 - f1_weighted: 0.7551 - accuracy: 0.7449\n",
      "Epoch 131/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6023 - f1_weighted: 0.7522 - accuracy: 0.7389\n",
      "Epoch 132/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5951 - f1_weighted: 0.7585 - accuracy: 0.7469\n",
      "Epoch 133/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5958 - f1_weighted: 0.7425 - accuracy: 0.7431\n",
      "Epoch 134/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5960 - f1_weighted: 0.7444 - accuracy: 0.7445\n",
      "Epoch 135/200\n",
      "56/56 [==============================] - 0s 853us/step - loss: 0.5943 - f1_weighted: 0.7548 - accuracy: 0.7475\n",
      "Epoch 136/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5938 - f1_weighted: 0.7520 - accuracy: 0.7438\n",
      "Epoch 137/200\n",
      "56/56 [==============================] - 0s 853us/step - loss: 0.5898 - f1_weighted: 0.7535 - accuracy: 0.7485\n",
      "Epoch 138/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5947 - f1_weighted: 0.7542 - accuracy: 0.7491\n",
      "Epoch 139/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5896 - f1_weighted: 0.7399 - accuracy: 0.7471\n",
      "Epoch 140/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5933 - f1_weighted: 0.7499 - accuracy: 0.7425\n",
      "Epoch 141/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5907 - f1_weighted: 0.7519 - accuracy: 0.7477\n",
      "Epoch 142/200\n",
      "56/56 [==============================] - 0s 796us/step - loss: 0.6020 - f1_weighted: 0.7472 - accuracy: 0.7360\n",
      "Epoch 143/200\n",
      "56/56 [==============================] - 0s 852us/step - loss: 0.5963 - f1_weighted: 0.7505 - accuracy: 0.7496\n",
      "Epoch 144/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5930 - f1_weighted: 0.7515 - accuracy: 0.7481\n",
      "Epoch 145/200\n",
      "56/56 [==============================] - 0s 964us/step - loss: 0.5890 - f1_weighted: 0.7495 - accuracy: 0.7492\n",
      "Epoch 146/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5887 - f1_weighted: 0.7509 - accuracy: 0.7452\n",
      "Epoch 147/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5894 - f1_weighted: 0.7480 - accuracy: 0.7468\n",
      "Epoch 148/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5903 - f1_weighted: 0.7495 - accuracy: 0.7482\n",
      "Epoch 149/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5907 - f1_weighted: 0.7469 - accuracy: 0.7483\n",
      "Epoch 150/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6028 - f1_weighted: 0.7548 - accuracy: 0.7423\n",
      "Epoch 151/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5866 - f1_weighted: 0.7431 - accuracy: 0.7459\n",
      "Epoch 152/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5997 - f1_weighted: 0.7510 - accuracy: 0.7423\n",
      "Epoch 153/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5970 - f1_weighted: 0.7477 - accuracy: 0.7431\n",
      "Epoch 154/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5924 - f1_weighted: 0.7512 - accuracy: 0.7474\n",
      "Epoch 155/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5971 - f1_weighted: 0.7508 - accuracy: 0.7478\n",
      "Epoch 156/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5860 - f1_weighted: 0.7481 - accuracy: 0.7486\n",
      "Epoch 157/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5932 - f1_weighted: 0.7426 - accuracy: 0.7447\n",
      "Epoch 158/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5897 - f1_weighted: 0.7476 - accuracy: 0.7504\n",
      "Epoch 159/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5836 - f1_weighted: 0.7425 - accuracy: 0.7475\n",
      "Epoch 160/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5900 - f1_weighted: 0.7444 - accuracy: 0.7488\n",
      "Epoch 161/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5938 - f1_weighted: 0.7473 - accuracy: 0.7440\n",
      "Epoch 162/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5973 - f1_weighted: 0.7377 - accuracy: 0.7401\n",
      "Epoch 163/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6013 - f1_weighted: 0.7532 - accuracy: 0.7407\n",
      "Epoch 164/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5960 - f1_weighted: 0.7454 - accuracy: 0.7389\n",
      "Epoch 165/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5961 - f1_weighted: 0.7450 - accuracy: 0.7455\n",
      "Epoch 166/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5931 - f1_weighted: 0.7431 - accuracy: 0.7463\n",
      "Epoch 167/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6042 - f1_weighted: 0.7478 - accuracy: 0.7376\n",
      "Epoch 168/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6014 - f1_weighted: 0.7415 - accuracy: 0.7427\n",
      "Epoch 169/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5881 - f1_weighted: 0.7445 - accuracy: 0.7472\n",
      "Epoch 170/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5908 - f1_weighted: 0.7440 - accuracy: 0.7454\n",
      "Epoch 171/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5970 - f1_weighted: 0.7448 - accuracy: 0.7431\n",
      "Epoch 172/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5929 - f1_weighted: 0.7512 - accuracy: 0.7477\n",
      "Epoch 173/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5827 - f1_weighted: 0.7477 - accuracy: 0.7498\n",
      "Epoch 174/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5890 - f1_weighted: 0.7405 - accuracy: 0.7448\n",
      "Epoch 175/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5927 - f1_weighted: 0.7441 - accuracy: 0.7458\n",
      "Epoch 176/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5980 - f1_weighted: 0.7428 - accuracy: 0.7425\n",
      "Epoch 177/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5895 - f1_weighted: 0.7488 - accuracy: 0.7452\n",
      "Epoch 178/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5828 - f1_weighted: 0.7475 - accuracy: 0.7509\n",
      "Epoch 179/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5978 - f1_weighted: 0.7478 - accuracy: 0.7454\n",
      "Epoch 180/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5884 - f1_weighted: 0.7498 - accuracy: 0.7453\n",
      "Epoch 181/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5922 - f1_weighted: 0.7448 - accuracy: 0.7452\n",
      "Epoch 182/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5916 - f1_weighted: 0.7399 - accuracy: 0.7497\n",
      "Epoch 183/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5941 - f1_weighted: 0.7410 - accuracy: 0.7423\n",
      "Epoch 184/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5922 - f1_weighted: 0.7403 - accuracy: 0.7489\n",
      "Epoch 185/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5953 - f1_weighted: 0.7414 - accuracy: 0.7403\n",
      "Epoch 186/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5796 - f1_weighted: 0.7436 - accuracy: 0.7505\n",
      "Epoch 187/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5942 - f1_weighted: 0.7407 - accuracy: 0.7486\n",
      "Epoch 188/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5952 - f1_weighted: 0.7421 - accuracy: 0.7464\n",
      "Epoch 189/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5909 - f1_weighted: 0.7373 - accuracy: 0.7481\n",
      "Epoch 190/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5809 - f1_weighted: 0.7374 - accuracy: 0.7553\n",
      "Epoch 191/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5917 - f1_weighted: 0.7441 - accuracy: 0.7483\n",
      "Epoch 192/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5870 - f1_weighted: 0.7374 - accuracy: 0.7535\n",
      "Epoch 193/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5845 - f1_weighted: 0.7499 - accuracy: 0.7517\n",
      "Epoch 194/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5983 - f1_weighted: 0.7461 - accuracy: 0.7420\n",
      "Epoch 195/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5975 - f1_weighted: 0.7408 - accuracy: 0.7465\n",
      "Epoch 196/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5928 - f1_weighted: 0.7367 - accuracy: 0.7415\n",
      "Epoch 197/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5890 - f1_weighted: 0.7529 - accuracy: 0.7482\n",
      "Epoch 198/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5893 - f1_weighted: 0.7400 - accuracy: 0.7519\n",
      "Epoch 199/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "56/56 [==============================] - 0s 971us/step - loss: 0.5927 - f1_weighted: 0.7481 - accuracy: 0.7471\n",
      "Epoch 200/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5867 - f1_weighted: 0.7446 - accuracy: 0.7477\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\lenovo\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\keras\\engine\\sequential.py:450: UserWarning: `model.predict_classes()` is deprecated and will be removed after 2021-01-01. Please use instead:* `np.argmax(model.predict(x), axis=-1)`,   if your model does multi-class classification   (e.g. if it uses a `softmax` last-layer activation).* `(model.predict(x) > 0.5).astype(\"int32\")`,   if your model does binary classification   (e.g. if it uses a `sigmoid` last-layer activation).\n",
      "  warnings.warn('`model.predict_classes()` is deprecated and '\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "56/56 [==============================] - 1s 1ms/step - loss: 1.1234 - f1_weighted: 0.8051 - accuracy: 0.3654\n",
      "Epoch 2/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.8607 - f1_weighted: 0.7267 - accuracy: 0.6213\n",
      "Epoch 3/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.7268 - f1_weighted: 0.7530 - accuracy: 0.7044\n",
      "Epoch 4/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6637 - f1_weighted: 0.7501 - accuracy: 0.7243\n",
      "Epoch 5/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6519 - f1_weighted: 0.7517 - accuracy: 0.7245\n",
      "Epoch 6/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6515 - f1_weighted: 0.7585 - accuracy: 0.7237\n",
      "Epoch 7/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6463 - f1_weighted: 0.7618 - accuracy: 0.7249\n",
      "Epoch 8/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6391 - f1_weighted: 0.7610 - accuracy: 0.7273\n",
      "Epoch 9/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6430 - f1_weighted: 0.7613 - accuracy: 0.7241\n",
      "Epoch 10/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6454 - f1_weighted: 0.7625 - accuracy: 0.7217\n",
      "Epoch 11/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6379 - f1_weighted: 0.7598 - accuracy: 0.7274\n",
      "Epoch 12/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6462 - f1_weighted: 0.7553 - accuracy: 0.7256\n",
      "Epoch 13/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6316 - f1_weighted: 0.7681 - accuracy: 0.7292\n",
      "Epoch 14/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6301 - f1_weighted: 0.7673 - accuracy: 0.7320\n",
      "Epoch 15/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6340 - f1_weighted: 0.7660 - accuracy: 0.7317\n",
      "Epoch 16/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6441 - f1_weighted: 0.7689 - accuracy: 0.7224\n",
      "Epoch 17/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6308 - f1_weighted: 0.7647 - accuracy: 0.7324\n",
      "Epoch 18/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6277 - f1_weighted: 0.7687 - accuracy: 0.7310\n",
      "Epoch 19/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6216 - f1_weighted: 0.7639 - accuracy: 0.7333\n",
      "Epoch 20/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6242 - f1_weighted: 0.7620 - accuracy: 0.7347\n",
      "Epoch 21/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6220 - f1_weighted: 0.7663 - accuracy: 0.7322\n",
      "Epoch 22/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6282 - f1_weighted: 0.7595 - accuracy: 0.7316\n",
      "Epoch 23/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6122 - f1_weighted: 0.7638 - accuracy: 0.7401\n",
      "Epoch 24/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6238 - f1_weighted: 0.7670 - accuracy: 0.7349\n",
      "Epoch 25/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6233 - f1_weighted: 0.7622 - accuracy: 0.7264\n",
      "Epoch 26/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6185 - f1_weighted: 0.7612 - accuracy: 0.7371\n",
      "Epoch 27/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6236 - f1_weighted: 0.7637 - accuracy: 0.7294\n",
      "Epoch 28/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6191 - f1_weighted: 0.7682 - accuracy: 0.7329\n",
      "Epoch 29/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6123 - f1_weighted: 0.7600 - accuracy: 0.7363\n",
      "Epoch 30/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6124 - f1_weighted: 0.7704 - accuracy: 0.7403\n",
      "Epoch 31/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6162 - f1_weighted: 0.7733 - accuracy: 0.7342\n",
      "Epoch 32/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6186 - f1_weighted: 0.7601 - accuracy: 0.7351\n",
      "Epoch 33/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6111 - f1_weighted: 0.7592 - accuracy: 0.7369\n",
      "Epoch 34/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6099 - f1_weighted: 0.7631 - accuracy: 0.7394\n",
      "Epoch 35/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6192 - f1_weighted: 0.7627 - accuracy: 0.7293\n",
      "Epoch 36/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6128 - f1_weighted: 0.7661 - accuracy: 0.7360\n",
      "Epoch 37/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6161 - f1_weighted: 0.7649 - accuracy: 0.7368\n",
      "Epoch 38/200\n",
      "56/56 [==============================] - 0s 852us/step - loss: 0.6101 - f1_weighted: 0.7609 - accuracy: 0.7401\n",
      "Epoch 39/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6162 - f1_weighted: 0.7562 - accuracy: 0.7368\n",
      "Epoch 40/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5993 - f1_weighted: 0.7631 - accuracy: 0.7413\n",
      "Epoch 41/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6178 - f1_weighted: 0.7630 - accuracy: 0.7319\n",
      "Epoch 42/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6065 - f1_weighted: 0.7660 - accuracy: 0.7417\n",
      "Epoch 43/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6104 - f1_weighted: 0.7609 - accuracy: 0.7353\n",
      "Epoch 44/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6063 - f1_weighted: 0.7604 - accuracy: 0.7381\n",
      "Epoch 45/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6015 - f1_weighted: 0.7571 - accuracy: 0.7445\n",
      "Epoch 46/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6047 - f1_weighted: 0.7542 - accuracy: 0.7404\n",
      "Epoch 47/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6003 - f1_weighted: 0.7573 - accuracy: 0.7452\n",
      "Epoch 48/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5996 - f1_weighted: 0.7605 - accuracy: 0.7427\n",
      "Epoch 49/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6061 - f1_weighted: 0.7628 - accuracy: 0.7411\n",
      "Epoch 50/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6133 - f1_weighted: 0.7606 - accuracy: 0.7375\n",
      "Epoch 51/200\n",
      "56/56 [==============================] - 0s 873us/step - loss: 0.6083 - f1_weighted: 0.7555 - accuracy: 0.7422\n",
      "Epoch 52/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6081 - f1_weighted: 0.7618 - accuracy: 0.7381\n",
      "Epoch 53/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6096 - f1_weighted: 0.7586 - accuracy: 0.7379\n",
      "Epoch 54/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6068 - f1_weighted: 0.7671 - accuracy: 0.7422\n",
      "Epoch 55/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6010 - f1_weighted: 0.7626 - accuracy: 0.7449\n",
      "Epoch 56/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5978 - f1_weighted: 0.7496 - accuracy: 0.7452\n",
      "Epoch 57/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6063 - f1_weighted: 0.7541 - accuracy: 0.7377\n",
      "Epoch 58/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6001 - f1_weighted: 0.7561 - accuracy: 0.7454\n",
      "Epoch 59/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6021 - f1_weighted: 0.7614 - accuracy: 0.7399\n",
      "Epoch 60/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6052 - f1_weighted: 0.7612 - accuracy: 0.7409\n",
      "Epoch 61/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5950 - f1_weighted: 0.7526 - accuracy: 0.7435\n",
      "Epoch 62/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6043 - f1_weighted: 0.7600 - accuracy: 0.7407\n",
      "Epoch 63/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6094 - f1_weighted: 0.7543 - accuracy: 0.7417\n",
      "Epoch 64/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5949 - f1_weighted: 0.7577 - accuracy: 0.7482\n",
      "Epoch 65/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6223 - f1_weighted: 0.7612 - accuracy: 0.7314\n",
      "Epoch 66/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5971 - f1_weighted: 0.7494 - accuracy: 0.7422\n",
      "Epoch 67/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6070 - f1_weighted: 0.7507 - accuracy: 0.7374\n",
      "Epoch 68/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6067 - f1_weighted: 0.7540 - accuracy: 0.7387\n",
      "Epoch 69/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5993 - f1_weighted: 0.7539 - accuracy: 0.7420\n",
      "Epoch 70/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5988 - f1_weighted: 0.7563 - accuracy: 0.7419\n",
      "Epoch 71/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5917 - f1_weighted: 0.7599 - accuracy: 0.7462\n",
      "Epoch 72/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5986 - f1_weighted: 0.7577 - accuracy: 0.7469\n",
      "Epoch 73/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5998 - f1_weighted: 0.7597 - accuracy: 0.7409\n",
      "Epoch 74/200\n",
      "56/56 [==============================] - 0s 991us/step - loss: 0.5931 - f1_weighted: 0.7527 - accuracy: 0.7464\n",
      "Epoch 75/200\n",
      "56/56 [==============================] - 0s 944us/step - loss: 0.5963 - f1_weighted: 0.7552 - accuracy: 0.7480\n",
      "Epoch 76/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5990 - f1_weighted: 0.7555 - accuracy: 0.7433\n",
      "Epoch 77/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5990 - f1_weighted: 0.7559 - accuracy: 0.7472\n",
      "Epoch 78/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6007 - f1_weighted: 0.7569 - accuracy: 0.7438\n",
      "Epoch 79/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6004 - f1_weighted: 0.7532 - accuracy: 0.7435\n",
      "Epoch 80/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5931 - f1_weighted: 0.7569 - accuracy: 0.7448\n",
      "Epoch 81/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6102 - f1_weighted: 0.7581 - accuracy: 0.7358\n",
      "Epoch 82/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6051 - f1_weighted: 0.7470 - accuracy: 0.7408\n",
      "Epoch 83/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5980 - f1_weighted: 0.7456 - accuracy: 0.7445\n",
      "Epoch 84/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5907 - f1_weighted: 0.7492 - accuracy: 0.7504\n",
      "Epoch 85/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5937 - f1_weighted: 0.7606 - accuracy: 0.7486\n",
      "Epoch 86/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6050 - f1_weighted: 0.7474 - accuracy: 0.7388\n",
      "Epoch 87/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5948 - f1_weighted: 0.7513 - accuracy: 0.7458\n",
      "Epoch 88/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5987 - f1_weighted: 0.7504 - accuracy: 0.7398\n",
      "Epoch 89/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5989 - f1_weighted: 0.7500 - accuracy: 0.7447\n",
      "Epoch 90/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5871 - f1_weighted: 0.7545 - accuracy: 0.7534\n",
      "Epoch 91/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5917 - f1_weighted: 0.7489 - accuracy: 0.7511\n",
      "Epoch 92/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5878 - f1_weighted: 0.7514 - accuracy: 0.7510\n",
      "Epoch 93/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5934 - f1_weighted: 0.7541 - accuracy: 0.7471\n",
      "Epoch 94/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5979 - f1_weighted: 0.7599 - accuracy: 0.7457\n",
      "Epoch 95/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5842 - f1_weighted: 0.7503 - accuracy: 0.7510\n",
      "Epoch 96/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6040 - f1_weighted: 0.7449 - accuracy: 0.7450\n",
      "Epoch 97/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6027 - f1_weighted: 0.7513 - accuracy: 0.7403\n",
      "Epoch 98/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5952 - f1_weighted: 0.7521 - accuracy: 0.7474\n",
      "Epoch 99/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5884 - f1_weighted: 0.7505 - accuracy: 0.7481\n",
      "Epoch 100/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5963 - f1_weighted: 0.7513 - accuracy: 0.7463\n",
      "Epoch 101/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5892 - f1_weighted: 0.7532 - accuracy: 0.7507\n",
      "Epoch 102/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5938 - f1_weighted: 0.7575 - accuracy: 0.7530\n",
      "Epoch 103/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6050 - f1_weighted: 0.7558 - accuracy: 0.7412\n",
      "Epoch 104/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5951 - f1_weighted: 0.7492 - accuracy: 0.7481\n",
      "Epoch 105/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6000 - f1_weighted: 0.7448 - accuracy: 0.7422\n",
      "Epoch 106/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5920 - f1_weighted: 0.7510 - accuracy: 0.7456\n",
      "Epoch 107/200\n",
      "56/56 [==============================] - 0s 853us/step - loss: 0.5822 - f1_weighted: 0.7491 - accuracy: 0.7530\n",
      "Epoch 108/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5923 - f1_weighted: 0.7524 - accuracy: 0.7508\n",
      "Epoch 109/200\n",
      "56/56 [==============================] - 0s 853us/step - loss: 0.5926 - f1_weighted: 0.7504 - accuracy: 0.7476\n",
      "Epoch 110/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5950 - f1_weighted: 0.7556 - accuracy: 0.7506\n",
      "Epoch 111/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5917 - f1_weighted: 0.7453 - accuracy: 0.7479\n",
      "Epoch 112/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5882 - f1_weighted: 0.7478 - accuracy: 0.7503\n",
      "Epoch 113/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5875 - f1_weighted: 0.7478 - accuracy: 0.7490\n",
      "Epoch 114/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5859 - f1_weighted: 0.7444 - accuracy: 0.7495\n",
      "Epoch 115/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6023 - f1_weighted: 0.7510 - accuracy: 0.7392\n",
      "Epoch 116/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5988 - f1_weighted: 0.7546 - accuracy: 0.7428\n",
      "Epoch 117/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6072 - f1_weighted: 0.7550 - accuracy: 0.7422\n",
      "Epoch 118/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5934 - f1_weighted: 0.7506 - accuracy: 0.7478\n",
      "Epoch 119/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5853 - f1_weighted: 0.7487 - accuracy: 0.7532\n",
      "Epoch 120/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5941 - f1_weighted: 0.7461 - accuracy: 0.7479\n",
      "Epoch 121/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5920 - f1_weighted: 0.7603 - accuracy: 0.7478\n",
      "Epoch 122/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5953 - f1_weighted: 0.7416 - accuracy: 0.7450\n",
      "Epoch 123/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5900 - f1_weighted: 0.7504 - accuracy: 0.7461\n",
      "Epoch 124/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5892 - f1_weighted: 0.7379 - accuracy: 0.7457\n",
      "Epoch 125/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5940 - f1_weighted: 0.7471 - accuracy: 0.7473\n",
      "Epoch 126/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5816 - f1_weighted: 0.7475 - accuracy: 0.7511\n",
      "Epoch 127/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5838 - f1_weighted: 0.7471 - accuracy: 0.7523\n",
      "Epoch 128/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5833 - f1_weighted: 0.7507 - accuracy: 0.7484\n",
      "Epoch 129/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5853 - f1_weighted: 0.7419 - accuracy: 0.7518\n",
      "Epoch 130/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5762 - f1_weighted: 0.7542 - accuracy: 0.7560\n",
      "Epoch 131/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5939 - f1_weighted: 0.7449 - accuracy: 0.7462\n",
      "Epoch 132/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5949 - f1_weighted: 0.7446 - accuracy: 0.7460\n",
      "Epoch 133/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5994 - f1_weighted: 0.7456 - accuracy: 0.7442\n",
      "Epoch 134/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6017 - f1_weighted: 0.7508 - accuracy: 0.7450\n",
      "Epoch 135/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5955 - f1_weighted: 0.7514 - accuracy: 0.7474\n",
      "Epoch 136/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5702 - f1_weighted: 0.7477 - accuracy: 0.7566\n",
      "Epoch 137/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5894 - f1_weighted: 0.7529 - accuracy: 0.7501\n",
      "Epoch 138/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5977 - f1_weighted: 0.7508 - accuracy: 0.7445\n",
      "Epoch 139/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5975 - f1_weighted: 0.7498 - accuracy: 0.7459\n",
      "Epoch 140/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5874 - f1_weighted: 0.7446 - accuracy: 0.7497\n",
      "Epoch 141/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5847 - f1_weighted: 0.7406 - accuracy: 0.7504\n",
      "Epoch 142/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5926 - f1_weighted: 0.7483 - accuracy: 0.7479\n",
      "Epoch 143/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5847 - f1_weighted: 0.7507 - accuracy: 0.7514\n",
      "Epoch 144/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5859 - f1_weighted: 0.7552 - accuracy: 0.7529\n",
      "Epoch 145/200\n",
      "56/56 [==============================] - 0s 991us/step - loss: 0.5918 - f1_weighted: 0.7437 - accuracy: 0.7446\n",
      "Epoch 146/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5903 - f1_weighted: 0.7394 - accuracy: 0.7467\n",
      "Epoch 147/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5920 - f1_weighted: 0.7471 - accuracy: 0.7444\n",
      "Epoch 148/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6012 - f1_weighted: 0.7459 - accuracy: 0.7424\n",
      "Epoch 149/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5810 - f1_weighted: 0.7455 - accuracy: 0.7526\n",
      "Epoch 150/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5825 - f1_weighted: 0.7414 - accuracy: 0.7541\n",
      "Epoch 151/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5848 - f1_weighted: 0.7419 - accuracy: 0.7538\n",
      "Epoch 152/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5854 - f1_weighted: 0.7493 - accuracy: 0.7488\n",
      "Epoch 153/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5969 - f1_weighted: 0.7413 - accuracy: 0.7476\n",
      "Epoch 154/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5798 - f1_weighted: 0.7410 - accuracy: 0.7565\n",
      "Epoch 155/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5901 - f1_weighted: 0.7378 - accuracy: 0.7492\n",
      "Epoch 156/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5898 - f1_weighted: 0.7425 - accuracy: 0.7503\n",
      "Epoch 157/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5885 - f1_weighted: 0.7465 - accuracy: 0.7453\n",
      "Epoch 158/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5859 - f1_weighted: 0.7402 - accuracy: 0.7470\n",
      "Epoch 159/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5967 - f1_weighted: 0.7430 - accuracy: 0.7385\n",
      "Epoch 160/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5766 - f1_weighted: 0.7428 - accuracy: 0.7555\n",
      "Epoch 161/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5927 - f1_weighted: 0.7446 - accuracy: 0.7462\n",
      "Epoch 162/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5833 - f1_weighted: 0.7409 - accuracy: 0.7485\n",
      "Epoch 163/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5955 - f1_weighted: 0.7402 - accuracy: 0.7413\n",
      "Epoch 164/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5981 - f1_weighted: 0.7377 - accuracy: 0.7425\n",
      "Epoch 165/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5966 - f1_weighted: 0.7372 - accuracy: 0.7418\n",
      "Epoch 166/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5936 - f1_weighted: 0.7464 - accuracy: 0.7482\n",
      "Epoch 167/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5859 - f1_weighted: 0.7439 - accuracy: 0.7539\n",
      "Epoch 168/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5832 - f1_weighted: 0.7458 - accuracy: 0.7526\n",
      "Epoch 169/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5919 - f1_weighted: 0.7446 - accuracy: 0.7506\n",
      "Epoch 170/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5830 - f1_weighted: 0.7401 - accuracy: 0.7516\n",
      "Epoch 171/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5779 - f1_weighted: 0.7415 - accuracy: 0.7571\n",
      "Epoch 172/200\n",
      "56/56 [==============================] - 0s 853us/step - loss: 0.5884 - f1_weighted: 0.7329 - accuracy: 0.7489\n",
      "Epoch 173/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5849 - f1_weighted: 0.7406 - accuracy: 0.7467\n",
      "Epoch 174/200\n",
      "56/56 [==============================] - 0s 853us/step - loss: 0.5707 - f1_weighted: 0.7511 - accuracy: 0.7594\n",
      "Epoch 175/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5898 - f1_weighted: 0.7350 - accuracy: 0.7473\n",
      "Epoch 176/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5747 - f1_weighted: 0.7462 - accuracy: 0.7540\n",
      "Epoch 177/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5876 - f1_weighted: 0.7408 - accuracy: 0.7495\n",
      "Epoch 178/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5817 - f1_weighted: 0.7358 - accuracy: 0.7554\n",
      "Epoch 179/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5960 - f1_weighted: 0.7325 - accuracy: 0.7445\n",
      "Epoch 180/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5908 - f1_weighted: 0.7363 - accuracy: 0.7500\n",
      "Epoch 181/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5916 - f1_weighted: 0.7391 - accuracy: 0.7452\n",
      "Epoch 182/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5878 - f1_weighted: 0.7430 - accuracy: 0.7474\n",
      "Epoch 183/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5857 - f1_weighted: 0.7436 - accuracy: 0.7481\n",
      "Epoch 184/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5931 - f1_weighted: 0.7468 - accuracy: 0.7486\n",
      "Epoch 185/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5838 - f1_weighted: 0.7466 - accuracy: 0.7506\n",
      "Epoch 186/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5716 - f1_weighted: 0.7468 - accuracy: 0.7563\n",
      "Epoch 187/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5899 - f1_weighted: 0.7408 - accuracy: 0.7493\n",
      "Epoch 188/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5828 - f1_weighted: 0.7422 - accuracy: 0.7458\n",
      "Epoch 189/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5937 - f1_weighted: 0.7498 - accuracy: 0.7423\n",
      "Epoch 190/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5933 - f1_weighted: 0.7398 - accuracy: 0.7481\n",
      "Epoch 191/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5969 - f1_weighted: 0.7410 - accuracy: 0.7427\n",
      "Epoch 192/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5757 - f1_weighted: 0.7458 - accuracy: 0.7557\n",
      "Epoch 193/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5806 - f1_weighted: 0.7465 - accuracy: 0.7574\n",
      "Epoch 194/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5798 - f1_weighted: 0.7434 - accuracy: 0.7477\n",
      "Epoch 195/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5851 - f1_weighted: 0.7447 - accuracy: 0.7488\n",
      "Epoch 196/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5842 - f1_weighted: 0.7372 - accuracy: 0.7496\n",
      "Epoch 197/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5848 - f1_weighted: 0.7390 - accuracy: 0.7494\n",
      "Epoch 198/200\n",
      "56/56 [==============================] - 0s 853us/step - loss: 0.5867 - f1_weighted: 0.7420 - accuracy: 0.7482\n",
      "Epoch 199/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5859 - f1_weighted: 0.7444 - accuracy: 0.7496\n",
      "Epoch 200/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5883 - f1_weighted: 0.7369 - accuracy: 0.7446\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\lenovo\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\keras\\engine\\sequential.py:450: UserWarning: `model.predict_classes()` is deprecated and will be removed after 2021-01-01. Please use instead:* `np.argmax(model.predict(x), axis=-1)`,   if your model does multi-class classification   (e.g. if it uses a `softmax` last-layer activation).* `(model.predict(x) > 0.5).astype(\"int32\")`,   if your model does binary classification   (e.g. if it uses a `sigmoid` last-layer activation).\n",
      "  warnings.warn('`model.predict_classes()` is deprecated and '\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "56/56 [==============================] - 1s 1ms/step - loss: 1.1270 - f1_weighted: 0.9806 - accuracy: 0.3625\n",
      "Epoch 2/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.8672 - f1_weighted: 0.8166 - accuracy: 0.6037\n",
      "Epoch 3/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.7777 - f1_weighted: 0.7586 - accuracy: 0.6627\n",
      "Epoch 4/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.7193 - f1_weighted: 0.7325 - accuracy: 0.6836\n",
      "Epoch 5/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6864 - f1_weighted: 0.7448 - accuracy: 0.7012\n",
      "Epoch 6/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6668 - f1_weighted: 0.7549 - accuracy: 0.7128\n",
      "Epoch 7/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6538 - f1_weighted: 0.7488 - accuracy: 0.7197\n",
      "Epoch 8/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6543 - f1_weighted: 0.7646 - accuracy: 0.7197\n",
      "Epoch 9/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6404 - f1_weighted: 0.7580 - accuracy: 0.7232\n",
      "Epoch 10/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6467 - f1_weighted: 0.7639 - accuracy: 0.7250\n",
      "Epoch 11/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6378 - f1_weighted: 0.7600 - accuracy: 0.7291\n",
      "Epoch 12/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6463 - f1_weighted: 0.7607 - accuracy: 0.7243\n",
      "Epoch 13/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6386 - f1_weighted: 0.7614 - accuracy: 0.7271\n",
      "Epoch 14/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6365 - f1_weighted: 0.7587 - accuracy: 0.7305\n",
      "Epoch 15/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6316 - f1_weighted: 0.7675 - accuracy: 0.7320\n",
      "Epoch 16/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6352 - f1_weighted: 0.7611 - accuracy: 0.7264\n",
      "Epoch 17/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6392 - f1_weighted: 0.7620 - accuracy: 0.7261\n",
      "Epoch 18/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6328 - f1_weighted: 0.7704 - accuracy: 0.7321\n",
      "Epoch 19/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6329 - f1_weighted: 0.7639 - accuracy: 0.7326\n",
      "Epoch 20/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6205 - f1_weighted: 0.7609 - accuracy: 0.7385\n",
      "Epoch 21/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6290 - f1_weighted: 0.7714 - accuracy: 0.7313\n",
      "Epoch 22/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6337 - f1_weighted: 0.7629 - accuracy: 0.7297\n",
      "Epoch 23/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6235 - f1_weighted: 0.7634 - accuracy: 0.7327\n",
      "Epoch 24/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6277 - f1_weighted: 0.7670 - accuracy: 0.7308\n",
      "Epoch 25/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6248 - f1_weighted: 0.7711 - accuracy: 0.7336\n",
      "Epoch 26/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6354 - f1_weighted: 0.7674 - accuracy: 0.7242\n",
      "Epoch 27/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6231 - f1_weighted: 0.7658 - accuracy: 0.7376\n",
      "Epoch 28/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6230 - f1_weighted: 0.7636 - accuracy: 0.7326\n",
      "Epoch 29/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6268 - f1_weighted: 0.7707 - accuracy: 0.7296\n",
      "Epoch 30/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6352 - f1_weighted: 0.7628 - accuracy: 0.7249\n",
      "Epoch 31/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6273 - f1_weighted: 0.7651 - accuracy: 0.7314\n",
      "Epoch 32/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6257 - f1_weighted: 0.7671 - accuracy: 0.7318\n",
      "Epoch 33/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6276 - f1_weighted: 0.7681 - accuracy: 0.7312\n",
      "Epoch 34/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6117 - f1_weighted: 0.7653 - accuracy: 0.7353\n",
      "Epoch 35/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6259 - f1_weighted: 0.7647 - accuracy: 0.7297\n",
      "Epoch 36/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6113 - f1_weighted: 0.7606 - accuracy: 0.7347\n",
      "Epoch 37/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6226 - f1_weighted: 0.7647 - accuracy: 0.7266\n",
      "Epoch 38/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6121 - f1_weighted: 0.7691 - accuracy: 0.7362\n",
      "Epoch 39/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6119 - f1_weighted: 0.7627 - accuracy: 0.7380\n",
      "Epoch 40/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6123 - f1_weighted: 0.7632 - accuracy: 0.7347\n",
      "Epoch 41/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6139 - f1_weighted: 0.7676 - accuracy: 0.7394\n",
      "Epoch 42/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6137 - f1_weighted: 0.7611 - accuracy: 0.7336\n",
      "Epoch 43/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6217 - f1_weighted: 0.7705 - accuracy: 0.7304\n",
      "Epoch 44/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6123 - f1_weighted: 0.7690 - accuracy: 0.7330\n",
      "Epoch 45/200\n",
      "56/56 [==============================] - 0s 853us/step - loss: 0.6092 - f1_weighted: 0.7649 - accuracy: 0.7367\n",
      "Epoch 46/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6153 - f1_weighted: 0.7634 - accuracy: 0.7313\n",
      "Epoch 47/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6126 - f1_weighted: 0.7618 - accuracy: 0.7349\n",
      "Epoch 48/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6105 - f1_weighted: 0.7558 - accuracy: 0.7374\n",
      "Epoch 49/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6242 - f1_weighted: 0.7650 - accuracy: 0.7298\n",
      "Epoch 50/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6123 - f1_weighted: 0.7544 - accuracy: 0.7379\n",
      "Epoch 51/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6117 - f1_weighted: 0.7538 - accuracy: 0.7321\n",
      "Epoch 52/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6120 - f1_weighted: 0.7593 - accuracy: 0.7336\n",
      "Epoch 53/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6181 - f1_weighted: 0.7667 - accuracy: 0.7284\n",
      "Epoch 54/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6080 - f1_weighted: 0.7624 - accuracy: 0.7352\n",
      "Epoch 55/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6188 - f1_weighted: 0.7540 - accuracy: 0.7307\n",
      "Epoch 56/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6062 - f1_weighted: 0.7631 - accuracy: 0.7370\n",
      "Epoch 57/200\n",
      "56/56 [==============================] - 0s 853us/step - loss: 0.6115 - f1_weighted: 0.7611 - accuracy: 0.7371\n",
      "Epoch 58/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6042 - f1_weighted: 0.7536 - accuracy: 0.7400\n",
      "Epoch 59/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6058 - f1_weighted: 0.7588 - accuracy: 0.7391\n",
      "Epoch 60/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5991 - f1_weighted: 0.7527 - accuracy: 0.7419\n",
      "Epoch 61/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6148 - f1_weighted: 0.7584 - accuracy: 0.7327\n",
      "Epoch 62/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6121 - f1_weighted: 0.7559 - accuracy: 0.7343\n",
      "Epoch 63/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6137 - f1_weighted: 0.7623 - accuracy: 0.7334\n",
      "Epoch 64/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6103 - f1_weighted: 0.7532 - accuracy: 0.7352\n",
      "Epoch 65/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6082 - f1_weighted: 0.7549 - accuracy: 0.7349\n",
      "Epoch 66/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6176 - f1_weighted: 0.7513 - accuracy: 0.7289\n",
      "Epoch 67/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6053 - f1_weighted: 0.7597 - accuracy: 0.7395\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 68/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6043 - f1_weighted: 0.7457 - accuracy: 0.7361\n",
      "Epoch 69/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5994 - f1_weighted: 0.7540 - accuracy: 0.7401\n",
      "Epoch 70/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6143 - f1_weighted: 0.7602 - accuracy: 0.7337\n",
      "Epoch 71/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5996 - f1_weighted: 0.7599 - accuracy: 0.7374\n",
      "Epoch 72/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6087 - f1_weighted: 0.7536 - accuracy: 0.7355\n",
      "Epoch 73/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6032 - f1_weighted: 0.7518 - accuracy: 0.7375\n",
      "Epoch 74/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5967 - f1_weighted: 0.7531 - accuracy: 0.7437\n",
      "Epoch 75/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6106 - f1_weighted: 0.7566 - accuracy: 0.7353\n",
      "Epoch 76/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5990 - f1_weighted: 0.7529 - accuracy: 0.7395\n",
      "Epoch 77/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6158 - f1_weighted: 0.7438 - accuracy: 0.7310\n",
      "Epoch 78/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6173 - f1_weighted: 0.7520 - accuracy: 0.7312\n",
      "Epoch 79/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6152 - f1_weighted: 0.7608 - accuracy: 0.7332\n",
      "Epoch 80/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6088 - f1_weighted: 0.7562 - accuracy: 0.7346\n",
      "Epoch 81/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6044 - f1_weighted: 0.7502 - accuracy: 0.7394\n",
      "Epoch 82/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5980 - f1_weighted: 0.7553 - accuracy: 0.7401\n",
      "Epoch 83/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6137 - f1_weighted: 0.7426 - accuracy: 0.7344\n",
      "Epoch 84/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6085 - f1_weighted: 0.7481 - accuracy: 0.7321\n",
      "Epoch 85/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6046 - f1_weighted: 0.7569 - accuracy: 0.7372\n",
      "Epoch 86/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5906 - f1_weighted: 0.7441 - accuracy: 0.7440\n",
      "Epoch 87/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6008 - f1_weighted: 0.7445 - accuracy: 0.7380\n",
      "Epoch 88/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6009 - f1_weighted: 0.7557 - accuracy: 0.7435\n",
      "Epoch 89/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5959 - f1_weighted: 0.7479 - accuracy: 0.7422\n",
      "Epoch 90/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6022 - f1_weighted: 0.7434 - accuracy: 0.7345\n",
      "Epoch 91/200\n",
      "56/56 [==============================] - 0s 853us/step - loss: 0.5943 - f1_weighted: 0.7343 - accuracy: 0.7423\n",
      "Epoch 92/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6012 - f1_weighted: 0.7520 - accuracy: 0.7434\n",
      "Epoch 93/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6090 - f1_weighted: 0.7526 - accuracy: 0.7357\n",
      "Epoch 94/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6073 - f1_weighted: 0.7569 - accuracy: 0.7346\n",
      "Epoch 95/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6039 - f1_weighted: 0.7533 - accuracy: 0.7363\n",
      "Epoch 96/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6044 - f1_weighted: 0.7503 - accuracy: 0.7328\n",
      "Epoch 97/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5984 - f1_weighted: 0.7468 - accuracy: 0.7369\n",
      "Epoch 98/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6093 - f1_weighted: 0.7486 - accuracy: 0.7321\n",
      "Epoch 99/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6000 - f1_weighted: 0.7451 - accuracy: 0.7387\n",
      "Epoch 100/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6034 - f1_weighted: 0.7524 - accuracy: 0.7372\n",
      "Epoch 101/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6035 - f1_weighted: 0.7525 - accuracy: 0.7348\n",
      "Epoch 102/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6043 - f1_weighted: 0.7491 - accuracy: 0.7368\n",
      "Epoch 103/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6039 - f1_weighted: 0.7496 - accuracy: 0.7357\n",
      "Epoch 104/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5950 - f1_weighted: 0.7532 - accuracy: 0.7425\n",
      "Epoch 105/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5953 - f1_weighted: 0.7429 - accuracy: 0.7416\n",
      "Epoch 106/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6040 - f1_weighted: 0.7480 - accuracy: 0.7357\n",
      "Epoch 107/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6038 - f1_weighted: 0.7483 - accuracy: 0.7349\n",
      "Epoch 108/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6004 - f1_weighted: 0.7484 - accuracy: 0.7354\n",
      "Epoch 109/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5943 - f1_weighted: 0.7522 - accuracy: 0.7421\n",
      "Epoch 110/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5920 - f1_weighted: 0.7494 - accuracy: 0.7451\n",
      "Epoch 111/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6047 - f1_weighted: 0.7500 - accuracy: 0.7356\n",
      "Epoch 112/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6020 - f1_weighted: 0.7477 - accuracy: 0.7370\n",
      "Epoch 113/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5965 - f1_weighted: 0.7425 - accuracy: 0.7377\n",
      "Epoch 114/200\n",
      "56/56 [==============================] - 0s 843us/step - loss: 0.6015 - f1_weighted: 0.7505 - accuracy: 0.7390\n",
      "Epoch 115/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6083 - f1_weighted: 0.7391 - accuracy: 0.7335\n",
      "Epoch 116/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5964 - f1_weighted: 0.7547 - accuracy: 0.7411\n",
      "Epoch 117/200\n",
      "56/56 [==============================] - 0s 853us/step - loss: 0.5944 - f1_weighted: 0.7408 - accuracy: 0.7407\n",
      "Epoch 118/200\n",
      "56/56 [==============================] - 0s 964us/step - loss: 0.6013 - f1_weighted: 0.7551 - accuracy: 0.7400\n",
      "Epoch 119/200\n",
      "56/56 [==============================] - 0s 853us/step - loss: 0.6061 - f1_weighted: 0.7401 - accuracy: 0.7334\n",
      "Epoch 120/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5917 - f1_weighted: 0.7547 - accuracy: 0.7436\n",
      "Epoch 121/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5956 - f1_weighted: 0.7481 - accuracy: 0.7431\n",
      "Epoch 122/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5966 - f1_weighted: 0.7383 - accuracy: 0.7385\n",
      "Epoch 123/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5945 - f1_weighted: 0.7485 - accuracy: 0.7418\n",
      "Epoch 124/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5944 - f1_weighted: 0.7456 - accuracy: 0.7415\n",
      "Epoch 125/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5862 - f1_weighted: 0.7477 - accuracy: 0.7486\n",
      "Epoch 126/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5996 - f1_weighted: 0.7379 - accuracy: 0.7365\n",
      "Epoch 127/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5963 - f1_weighted: 0.7417 - accuracy: 0.7422\n",
      "Epoch 128/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5986 - f1_weighted: 0.7423 - accuracy: 0.7457\n",
      "Epoch 129/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6056 - f1_weighted: 0.7398 - accuracy: 0.7332\n",
      "Epoch 130/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6120 - f1_weighted: 0.7356 - accuracy: 0.7276\n",
      "Epoch 131/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6086 - f1_weighted: 0.7513 - accuracy: 0.7314\n",
      "Epoch 132/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5995 - f1_weighted: 0.7514 - accuracy: 0.7411\n",
      "Epoch 133/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6056 - f1_weighted: 0.7503 - accuracy: 0.7353\n",
      "Epoch 134/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6008 - f1_weighted: 0.7433 - accuracy: 0.7348\n",
      "Epoch 135/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6113 - f1_weighted: 0.7401 - accuracy: 0.7310\n",
      "Epoch 136/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5971 - f1_weighted: 0.7487 - accuracy: 0.7390\n",
      "Epoch 137/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5890 - f1_weighted: 0.7551 - accuracy: 0.7477\n",
      "Epoch 138/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6063 - f1_weighted: 0.7389 - accuracy: 0.7298\n",
      "Epoch 139/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5934 - f1_weighted: 0.7454 - accuracy: 0.7445\n",
      "Epoch 140/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5900 - f1_weighted: 0.7478 - accuracy: 0.7378\n",
      "Epoch 141/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5927 - f1_weighted: 0.7444 - accuracy: 0.7420\n",
      "Epoch 142/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5956 - f1_weighted: 0.7456 - accuracy: 0.7384\n",
      "Epoch 143/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5972 - f1_weighted: 0.7421 - accuracy: 0.7406\n",
      "Epoch 144/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5894 - f1_weighted: 0.7416 - accuracy: 0.7418\n",
      "Epoch 145/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6036 - f1_weighted: 0.7437 - accuracy: 0.7358\n",
      "Epoch 146/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5863 - f1_weighted: 0.7349 - accuracy: 0.7455\n",
      "Epoch 147/200\n",
      "56/56 [==============================] - 0s 853us/step - loss: 0.5972 - f1_weighted: 0.7443 - accuracy: 0.7407\n",
      "Epoch 148/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6053 - f1_weighted: 0.7390 - accuracy: 0.7384\n",
      "Epoch 149/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5991 - f1_weighted: 0.7533 - accuracy: 0.7397\n",
      "Epoch 150/200\n",
      "56/56 [==============================] - 0s 853us/step - loss: 0.5968 - f1_weighted: 0.7424 - accuracy: 0.7416\n",
      "Epoch 151/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6000 - f1_weighted: 0.7319 - accuracy: 0.7385\n",
      "Epoch 152/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5888 - f1_weighted: 0.7436 - accuracy: 0.7432\n",
      "Epoch 153/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5924 - f1_weighted: 0.7394 - accuracy: 0.7413\n",
      "Epoch 154/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5982 - f1_weighted: 0.7458 - accuracy: 0.7400\n",
      "Epoch 155/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6035 - f1_weighted: 0.7471 - accuracy: 0.7368\n",
      "Epoch 156/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5882 - f1_weighted: 0.7435 - accuracy: 0.7419\n",
      "Epoch 157/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6020 - f1_weighted: 0.7394 - accuracy: 0.7368\n",
      "Epoch 158/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5983 - f1_weighted: 0.7500 - accuracy: 0.7399\n",
      "Epoch 159/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5895 - f1_weighted: 0.7414 - accuracy: 0.7442\n",
      "Epoch 160/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5870 - f1_weighted: 0.7372 - accuracy: 0.7475\n",
      "Epoch 161/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5926 - f1_weighted: 0.7350 - accuracy: 0.7416\n",
      "Epoch 162/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5959 - f1_weighted: 0.7394 - accuracy: 0.7397\n",
      "Epoch 163/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5947 - f1_weighted: 0.7495 - accuracy: 0.7404\n",
      "Epoch 164/200\n",
      "56/56 [==============================] - 0s 984us/step - loss: 0.5877 - f1_weighted: 0.7413 - accuracy: 0.7447\n",
      "Epoch 165/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5923 - f1_weighted: 0.7399 - accuracy: 0.7410\n",
      "Epoch 166/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5907 - f1_weighted: 0.7433 - accuracy: 0.7448\n",
      "Epoch 167/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5998 - f1_weighted: 0.7397 - accuracy: 0.7382\n",
      "Epoch 168/200\n",
      "56/56 [==============================] - 0s 853us/step - loss: 0.5953 - f1_weighted: 0.7466 - accuracy: 0.7390\n",
      "Epoch 169/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5907 - f1_weighted: 0.7398 - accuracy: 0.7455\n",
      "Epoch 170/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5981 - f1_weighted: 0.7414 - accuracy: 0.7370\n",
      "Epoch 171/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5994 - f1_weighted: 0.7433 - accuracy: 0.7384\n",
      "Epoch 172/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5940 - f1_weighted: 0.7352 - accuracy: 0.7447\n",
      "Epoch 173/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5853 - f1_weighted: 0.7462 - accuracy: 0.7501\n",
      "Epoch 174/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5884 - f1_weighted: 0.7498 - accuracy: 0.7488\n",
      "Epoch 175/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6045 - f1_weighted: 0.7391 - accuracy: 0.7344\n",
      "Epoch 176/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5933 - f1_weighted: 0.7386 - accuracy: 0.7415\n",
      "Epoch 177/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6002 - f1_weighted: 0.7417 - accuracy: 0.7388\n",
      "Epoch 178/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6002 - f1_weighted: 0.7464 - accuracy: 0.7354\n",
      "Epoch 179/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5976 - f1_weighted: 0.7393 - accuracy: 0.7378\n",
      "Epoch 180/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5918 - f1_weighted: 0.7481 - accuracy: 0.7438\n",
      "Epoch 181/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5890 - f1_weighted: 0.7396 - accuracy: 0.7503\n",
      "Epoch 182/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5970 - f1_weighted: 0.7450 - accuracy: 0.7406\n",
      "Epoch 183/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5834 - f1_weighted: 0.7441 - accuracy: 0.7451\n",
      "Epoch 184/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5822 - f1_weighted: 0.7454 - accuracy: 0.7524\n",
      "Epoch 185/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5954 - f1_weighted: 0.7323 - accuracy: 0.7386\n",
      "Epoch 186/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5898 - f1_weighted: 0.7478 - accuracy: 0.7425\n",
      "Epoch 187/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6085 - f1_weighted: 0.7403 - accuracy: 0.7311\n",
      "Epoch 188/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5856 - f1_weighted: 0.7426 - accuracy: 0.7434\n",
      "Epoch 189/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5921 - f1_weighted: 0.7394 - accuracy: 0.7402\n",
      "Epoch 190/200\n",
      "56/56 [==============================] - ETA: 0s - loss: 0.6096 - f1_weighted: 0.7532 - accuracy: 0.73 - 0s 1ms/step - loss: 0.6063 - f1_weighted: 0.7515 - accuracy: 0.7349\n",
      "Epoch 191/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5775 - f1_weighted: 0.7422 - accuracy: 0.7502\n",
      "Epoch 192/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5944 - f1_weighted: 0.7340 - accuracy: 0.7388\n",
      "Epoch 193/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5802 - f1_weighted: 0.7423 - accuracy: 0.7556\n",
      "Epoch 194/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5972 - f1_weighted: 0.7415 - accuracy: 0.7406\n",
      "Epoch 195/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6023 - f1_weighted: 0.7363 - accuracy: 0.7376\n",
      "Epoch 196/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5807 - f1_weighted: 0.7346 - accuracy: 0.7445\n",
      "Epoch 197/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5974 - f1_weighted: 0.7378 - accuracy: 0.7376\n",
      "Epoch 198/200\n",
      "56/56 [==============================] - ETA: 0s - loss: 0.5900 - f1_weighted: 0.7403 - accuracy: 0.74 - 0s 1ms/step - loss: 0.5902 - f1_weighted: 0.7395 - accuracy: 0.7447\n",
      "Epoch 199/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5838 - f1_weighted: 0.7335 - accuracy: 0.7502\n",
      "Epoch 200/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5890 - f1_weighted: 0.7448 - accuracy: 0.7443\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\lenovo\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\keras\\engine\\sequential.py:450: UserWarning: `model.predict_classes()` is deprecated and will be removed after 2021-01-01. Please use instead:* `np.argmax(model.predict(x), axis=-1)`,   if your model does multi-class classification   (e.g. if it uses a `softmax` last-layer activation).* `(model.predict(x) > 0.5).astype(\"int32\")`,   if your model does binary classification   (e.g. if it uses a `sigmoid` last-layer activation).\n",
      "  warnings.warn('`model.predict_classes()` is deprecated and '\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "56/56 [==============================] - 1s 1ms/step - loss: 1.3193 - f1_weighted: 0.5857 - accuracy: 0.2709\n",
      "Epoch 2/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.8757 - f1_weighted: 0.6880 - accuracy: 0.6179\n",
      "Epoch 3/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.7390 - f1_weighted: 0.7258 - accuracy: 0.6921\n",
      "Epoch 4/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6965 - f1_weighted: 0.7398 - accuracy: 0.7123\n",
      "Epoch 5/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6624 - f1_weighted: 0.7512 - accuracy: 0.7246\n",
      "Epoch 6/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6640 - f1_weighted: 0.7660 - accuracy: 0.7224\n",
      "Epoch 7/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6591 - f1_weighted: 0.7636 - accuracy: 0.7214\n",
      "Epoch 8/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6473 - f1_weighted: 0.7639 - accuracy: 0.7279\n",
      "Epoch 9/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6406 - f1_weighted: 0.7674 - accuracy: 0.7306\n",
      "Epoch 10/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6361 - f1_weighted: 0.7658 - accuracy: 0.7307\n",
      "Epoch 11/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6397 - f1_weighted: 0.7665 - accuracy: 0.7298\n",
      "Epoch 12/200\n",
      "56/56 [==============================] - 0s 853us/step - loss: 0.6341 - f1_weighted: 0.7678 - accuracy: 0.7282\n",
      "Epoch 13/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6336 - f1_weighted: 0.7619 - accuracy: 0.7268\n",
      "Epoch 14/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6398 - f1_weighted: 0.7706 - accuracy: 0.7270\n",
      "Epoch 15/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6285 - f1_weighted: 0.7655 - accuracy: 0.7329\n",
      "Epoch 16/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6225 - f1_weighted: 0.7622 - accuracy: 0.7358\n",
      "Epoch 17/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6321 - f1_weighted: 0.7765 - accuracy: 0.7296\n",
      "Epoch 18/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6297 - f1_weighted: 0.7649 - accuracy: 0.7315\n",
      "Epoch 19/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6179 - f1_weighted: 0.7636 - accuracy: 0.7369\n",
      "Epoch 20/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6229 - f1_weighted: 0.7631 - accuracy: 0.7352\n",
      "Epoch 21/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6278 - f1_weighted: 0.7669 - accuracy: 0.7362\n",
      "Epoch 22/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6252 - f1_weighted: 0.7641 - accuracy: 0.7348\n",
      "Epoch 23/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6199 - f1_weighted: 0.7640 - accuracy: 0.7358\n",
      "Epoch 24/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6209 - f1_weighted: 0.7644 - accuracy: 0.7312\n",
      "Epoch 25/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6264 - f1_weighted: 0.7688 - accuracy: 0.7330\n",
      "Epoch 26/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6260 - f1_weighted: 0.7684 - accuracy: 0.7368\n",
      "Epoch 27/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6179 - f1_weighted: 0.7673 - accuracy: 0.7410\n",
      "Epoch 28/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6305 - f1_weighted: 0.7673 - accuracy: 0.7315\n",
      "Epoch 29/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6199 - f1_weighted: 0.7646 - accuracy: 0.7374\n",
      "Epoch 30/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6073 - f1_weighted: 0.7633 - accuracy: 0.7463\n",
      "Epoch 31/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6272 - f1_weighted: 0.7742 - accuracy: 0.7330\n",
      "Epoch 32/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6163 - f1_weighted: 0.7653 - accuracy: 0.7373\n",
      "Epoch 33/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6190 - f1_weighted: 0.7603 - accuracy: 0.7388\n",
      "Epoch 34/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6164 - f1_weighted: 0.7708 - accuracy: 0.7400\n",
      "Epoch 35/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6303 - f1_weighted: 0.7626 - accuracy: 0.7266\n",
      "Epoch 36/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6115 - f1_weighted: 0.7628 - accuracy: 0.7394\n",
      "Epoch 37/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6158 - f1_weighted: 0.7657 - accuracy: 0.7385\n",
      "Epoch 38/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6173 - f1_weighted: 0.7683 - accuracy: 0.7425\n",
      "Epoch 39/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6118 - f1_weighted: 0.7639 - accuracy: 0.7402\n",
      "Epoch 40/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6216 - f1_weighted: 0.7591 - accuracy: 0.7344\n",
      "Epoch 41/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6041 - f1_weighted: 0.7651 - accuracy: 0.7459\n",
      "Epoch 42/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6107 - f1_weighted: 0.7627 - accuracy: 0.7376\n",
      "Epoch 43/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6124 - f1_weighted: 0.7650 - accuracy: 0.7397\n",
      "Epoch 44/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6106 - f1_weighted: 0.7670 - accuracy: 0.7400\n",
      "Epoch 45/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6134 - f1_weighted: 0.7656 - accuracy: 0.7404\n",
      "Epoch 46/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6130 - f1_weighted: 0.7635 - accuracy: 0.7372\n",
      "Epoch 47/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5984 - f1_weighted: 0.7657 - accuracy: 0.7478\n",
      "Epoch 48/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6124 - f1_weighted: 0.7583 - accuracy: 0.7378\n",
      "Epoch 49/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6103 - f1_weighted: 0.7710 - accuracy: 0.7408\n",
      "Epoch 50/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6100 - f1_weighted: 0.7649 - accuracy: 0.7431\n",
      "Epoch 51/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6152 - f1_weighted: 0.7639 - accuracy: 0.7393\n",
      "Epoch 52/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6119 - f1_weighted: 0.7602 - accuracy: 0.7399\n",
      "Epoch 53/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6070 - f1_weighted: 0.7618 - accuracy: 0.7405\n",
      "Epoch 54/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6159 - f1_weighted: 0.7665 - accuracy: 0.7417\n",
      "Epoch 55/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6118 - f1_weighted: 0.7674 - accuracy: 0.7416\n",
      "Epoch 56/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6050 - f1_weighted: 0.7604 - accuracy: 0.7450\n",
      "Epoch 57/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5921 - f1_weighted: 0.7626 - accuracy: 0.7516\n",
      "Epoch 58/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6114 - f1_weighted: 0.7554 - accuracy: 0.7390\n",
      "Epoch 59/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6088 - f1_weighted: 0.7604 - accuracy: 0.7406\n",
      "Epoch 60/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6020 - f1_weighted: 0.7576 - accuracy: 0.7437\n",
      "Epoch 61/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6064 - f1_weighted: 0.7673 - accuracy: 0.7429\n",
      "Epoch 62/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5998 - f1_weighted: 0.7648 - accuracy: 0.7490\n",
      "Epoch 63/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6109 - f1_weighted: 0.7623 - accuracy: 0.7401\n",
      "Epoch 64/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6082 - f1_weighted: 0.7524 - accuracy: 0.7397\n",
      "Epoch 65/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6082 - f1_weighted: 0.7604 - accuracy: 0.7448\n",
      "Epoch 66/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5946 - f1_weighted: 0.7614 - accuracy: 0.7509\n",
      "Epoch 67/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6009 - f1_weighted: 0.7529 - accuracy: 0.7425\n",
      "Epoch 68/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6108 - f1_weighted: 0.7574 - accuracy: 0.7429\n",
      "Epoch 69/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5955 - f1_weighted: 0.7678 - accuracy: 0.7473\n",
      "Epoch 70/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5971 - f1_weighted: 0.7647 - accuracy: 0.7476\n",
      "Epoch 71/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5936 - f1_weighted: 0.7596 - accuracy: 0.7498\n",
      "Epoch 72/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5988 - f1_weighted: 0.7619 - accuracy: 0.7440\n",
      "Epoch 73/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6011 - f1_weighted: 0.7544 - accuracy: 0.7443\n",
      "Epoch 74/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5957 - f1_weighted: 0.7559 - accuracy: 0.7461\n",
      "Epoch 75/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6077 - f1_weighted: 0.7590 - accuracy: 0.7426\n",
      "Epoch 76/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6044 - f1_weighted: 0.7504 - accuracy: 0.7430\n",
      "Epoch 77/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6073 - f1_weighted: 0.7520 - accuracy: 0.7379\n",
      "Epoch 78/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6002 - f1_weighted: 0.7593 - accuracy: 0.7448\n",
      "Epoch 79/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5977 - f1_weighted: 0.7659 - accuracy: 0.7451\n",
      "Epoch 80/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6024 - f1_weighted: 0.7528 - accuracy: 0.7432\n",
      "Epoch 81/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6112 - f1_weighted: 0.7547 - accuracy: 0.7392\n",
      "Epoch 82/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6045 - f1_weighted: 0.7643 - accuracy: 0.7407\n",
      "Epoch 83/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6033 - f1_weighted: 0.7598 - accuracy: 0.7448\n",
      "Epoch 84/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5942 - f1_weighted: 0.7539 - accuracy: 0.7458\n",
      "Epoch 85/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6067 - f1_weighted: 0.7540 - accuracy: 0.7422\n",
      "Epoch 86/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5928 - f1_weighted: 0.7583 - accuracy: 0.7454\n",
      "Epoch 87/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6023 - f1_weighted: 0.7555 - accuracy: 0.7417\n",
      "Epoch 88/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5951 - f1_weighted: 0.7537 - accuracy: 0.7421\n",
      "Epoch 89/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5957 - f1_weighted: 0.7556 - accuracy: 0.7473\n",
      "Epoch 90/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6003 - f1_weighted: 0.7561 - accuracy: 0.7411\n",
      "Epoch 91/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6025 - f1_weighted: 0.7487 - accuracy: 0.7413\n",
      "Epoch 92/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5976 - f1_weighted: 0.7572 - accuracy: 0.7503\n",
      "Epoch 93/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6012 - f1_weighted: 0.7568 - accuracy: 0.7431\n",
      "Epoch 94/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6066 - f1_weighted: 0.7566 - accuracy: 0.7401\n",
      "Epoch 95/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6085 - f1_weighted: 0.7501 - accuracy: 0.7362\n",
      "Epoch 96/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6011 - f1_weighted: 0.7491 - accuracy: 0.7457\n",
      "Epoch 97/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6076 - f1_weighted: 0.7578 - accuracy: 0.7387\n",
      "Epoch 98/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6058 - f1_weighted: 0.7578 - accuracy: 0.7383\n",
      "Epoch 99/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5979 - f1_weighted: 0.7602 - accuracy: 0.7462\n",
      "Epoch 100/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5905 - f1_weighted: 0.7480 - accuracy: 0.7464\n",
      "Epoch 101/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5979 - f1_weighted: 0.7551 - accuracy: 0.7421\n",
      "Epoch 102/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5884 - f1_weighted: 0.7494 - accuracy: 0.7494\n",
      "Epoch 103/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5914 - f1_weighted: 0.7504 - accuracy: 0.7436\n",
      "Epoch 104/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5947 - f1_weighted: 0.7523 - accuracy: 0.7455\n",
      "Epoch 105/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5969 - f1_weighted: 0.7561 - accuracy: 0.7452\n",
      "Epoch 106/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5979 - f1_weighted: 0.7578 - accuracy: 0.7449\n",
      "Epoch 107/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5932 - f1_weighted: 0.7458 - accuracy: 0.7429\n",
      "Epoch 108/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5937 - f1_weighted: 0.7522 - accuracy: 0.7463\n",
      "Epoch 109/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6012 - f1_weighted: 0.7631 - accuracy: 0.7405\n",
      "Epoch 110/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5870 - f1_weighted: 0.7496 - accuracy: 0.7503\n",
      "Epoch 111/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5879 - f1_weighted: 0.7527 - accuracy: 0.7485\n",
      "Epoch 112/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5860 - f1_weighted: 0.7479 - accuracy: 0.7486\n",
      "Epoch 113/200\n",
      "56/56 [==============================] - 0s 853us/step - loss: 0.5879 - f1_weighted: 0.7576 - accuracy: 0.7509\n",
      "Epoch 114/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6017 - f1_weighted: 0.7530 - accuracy: 0.7421\n",
      "Epoch 115/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5933 - f1_weighted: 0.7459 - accuracy: 0.7420\n",
      "Epoch 116/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5930 - f1_weighted: 0.7491 - accuracy: 0.7455\n",
      "Epoch 117/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5963 - f1_weighted: 0.7567 - accuracy: 0.7444\n",
      "Epoch 118/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5947 - f1_weighted: 0.7529 - accuracy: 0.7498\n",
      "Epoch 119/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6056 - f1_weighted: 0.7543 - accuracy: 0.7389\n",
      "Epoch 120/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5895 - f1_weighted: 0.7440 - accuracy: 0.7478\n",
      "Epoch 121/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5980 - f1_weighted: 0.7516 - accuracy: 0.7389\n",
      "Epoch 122/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5859 - f1_weighted: 0.7528 - accuracy: 0.7528\n",
      "Epoch 123/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6035 - f1_weighted: 0.7560 - accuracy: 0.7442\n",
      "Epoch 124/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5943 - f1_weighted: 0.7574 - accuracy: 0.7462\n",
      "Epoch 125/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5908 - f1_weighted: 0.7535 - accuracy: 0.7436\n",
      "Epoch 126/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5830 - f1_weighted: 0.7497 - accuracy: 0.7527\n",
      "Epoch 127/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5931 - f1_weighted: 0.7449 - accuracy: 0.7476\n",
      "Epoch 128/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6000 - f1_weighted: 0.7521 - accuracy: 0.7385\n",
      "Epoch 129/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6008 - f1_weighted: 0.7600 - accuracy: 0.7419\n",
      "Epoch 130/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5901 - f1_weighted: 0.7525 - accuracy: 0.7493\n",
      "Epoch 131/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5989 - f1_weighted: 0.7531 - accuracy: 0.7423\n",
      "Epoch 132/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5958 - f1_weighted: 0.7439 - accuracy: 0.7434\n",
      "Epoch 133/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5969 - f1_weighted: 0.7402 - accuracy: 0.7424\n",
      "Epoch 134/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5975 - f1_weighted: 0.7508 - accuracy: 0.7443\n",
      "Epoch 135/200\n",
      "56/56 [==============================] - ETA: 0s - loss: 0.5885 - f1_weighted: 0.7465 - accuracy: 0.74 - 0s 1ms/step - loss: 0.5899 - f1_weighted: 0.7484 - accuracy: 0.7457\n",
      "Epoch 136/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5930 - f1_weighted: 0.7521 - accuracy: 0.7470\n",
      "Epoch 137/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5874 - f1_weighted: 0.7452 - accuracy: 0.7441\n",
      "Epoch 138/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5977 - f1_weighted: 0.7518 - accuracy: 0.7439\n",
      "Epoch 139/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5959 - f1_weighted: 0.7504 - accuracy: 0.7431\n",
      "Epoch 140/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5982 - f1_weighted: 0.7498 - accuracy: 0.7451\n",
      "Epoch 141/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5821 - f1_weighted: 0.7576 - accuracy: 0.7519\n",
      "Epoch 142/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5987 - f1_weighted: 0.7458 - accuracy: 0.7428\n",
      "Epoch 143/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5930 - f1_weighted: 0.7537 - accuracy: 0.7488\n",
      "Epoch 144/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5936 - f1_weighted: 0.7495 - accuracy: 0.7463\n",
      "Epoch 145/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5932 - f1_weighted: 0.7464 - accuracy: 0.7455\n",
      "Epoch 146/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5886 - f1_weighted: 0.7526 - accuracy: 0.7486\n",
      "Epoch 147/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5860 - f1_weighted: 0.7568 - accuracy: 0.7486\n",
      "Epoch 148/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5978 - f1_weighted: 0.7448 - accuracy: 0.7410\n",
      "Epoch 149/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5825 - f1_weighted: 0.7469 - accuracy: 0.7550\n",
      "Epoch 150/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5891 - f1_weighted: 0.7473 - accuracy: 0.7436\n",
      "Epoch 151/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5930 - f1_weighted: 0.7486 - accuracy: 0.7429\n",
      "Epoch 152/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5835 - f1_weighted: 0.7535 - accuracy: 0.7527\n",
      "Epoch 153/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5908 - f1_weighted: 0.7479 - accuracy: 0.7511\n",
      "Epoch 154/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5939 - f1_weighted: 0.7436 - accuracy: 0.7457\n",
      "Epoch 155/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5849 - f1_weighted: 0.7514 - accuracy: 0.7500\n",
      "Epoch 156/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5915 - f1_weighted: 0.7472 - accuracy: 0.7459\n",
      "Epoch 157/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5951 - f1_weighted: 0.7571 - accuracy: 0.7445\n",
      "Epoch 158/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5819 - f1_weighted: 0.7471 - accuracy: 0.7534\n",
      "Epoch 159/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5894 - f1_weighted: 0.7555 - accuracy: 0.7489\n",
      "Epoch 160/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5820 - f1_weighted: 0.7485 - accuracy: 0.7534\n",
      "Epoch 161/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5987 - f1_weighted: 0.7541 - accuracy: 0.7441\n",
      "Epoch 162/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5930 - f1_weighted: 0.7361 - accuracy: 0.7431\n",
      "Epoch 163/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5920 - f1_weighted: 0.7478 - accuracy: 0.7485\n",
      "Epoch 164/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5785 - f1_weighted: 0.7490 - accuracy: 0.7559\n",
      "Epoch 165/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5916 - f1_weighted: 0.7522 - accuracy: 0.7443\n",
      "Epoch 166/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5847 - f1_weighted: 0.7502 - accuracy: 0.7509\n",
      "Epoch 167/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5966 - f1_weighted: 0.7416 - accuracy: 0.7427\n",
      "Epoch 168/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5783 - f1_weighted: 0.7481 - accuracy: 0.7490\n",
      "Epoch 169/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5859 - f1_weighted: 0.7454 - accuracy: 0.7473\n",
      "Epoch 170/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5843 - f1_weighted: 0.7487 - accuracy: 0.7559\n",
      "Epoch 171/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5844 - f1_weighted: 0.7434 - accuracy: 0.7507\n",
      "Epoch 172/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5839 - f1_weighted: 0.7465 - accuracy: 0.7459\n",
      "Epoch 173/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5892 - f1_weighted: 0.7440 - accuracy: 0.7463\n",
      "Epoch 174/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5806 - f1_weighted: 0.7457 - accuracy: 0.7524\n",
      "Epoch 175/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5828 - f1_weighted: 0.7536 - accuracy: 0.7503\n",
      "Epoch 176/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5851 - f1_weighted: 0.7375 - accuracy: 0.7536\n",
      "Epoch 177/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5785 - f1_weighted: 0.7458 - accuracy: 0.7552\n",
      "Epoch 178/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5895 - f1_weighted: 0.7477 - accuracy: 0.7484\n",
      "Epoch 179/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5872 - f1_weighted: 0.7504 - accuracy: 0.7477\n",
      "Epoch 180/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5907 - f1_weighted: 0.7492 - accuracy: 0.7432\n",
      "Epoch 181/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5872 - f1_weighted: 0.7493 - accuracy: 0.7497\n",
      "Epoch 182/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5935 - f1_weighted: 0.7405 - accuracy: 0.7422\n",
      "Epoch 183/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5836 - f1_weighted: 0.7493 - accuracy: 0.7504\n",
      "Epoch 184/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5951 - f1_weighted: 0.7470 - accuracy: 0.7461\n",
      "Epoch 185/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5927 - f1_weighted: 0.7563 - accuracy: 0.7469\n",
      "Epoch 186/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5922 - f1_weighted: 0.7481 - accuracy: 0.7404\n",
      "Epoch 187/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5831 - f1_weighted: 0.7459 - accuracy: 0.7506\n",
      "Epoch 188/200\n",
      "56/56 [==============================] - 0s 944us/step - loss: 0.5914 - f1_weighted: 0.7358 - accuracy: 0.7435\n",
      "Epoch 189/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5793 - f1_weighted: 0.7449 - accuracy: 0.7499\n",
      "Epoch 190/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5863 - f1_weighted: 0.7509 - accuracy: 0.7489\n",
      "Epoch 191/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5969 - f1_weighted: 0.7418 - accuracy: 0.7449\n",
      "Epoch 192/200\n",
      "56/56 [==============================] - 0s 962us/step - loss: 0.5856 - f1_weighted: 0.7443 - accuracy: 0.7501\n",
      "Epoch 193/200\n",
      "56/56 [==============================] - 0s 853us/step - loss: 0.5824 - f1_weighted: 0.7431 - accuracy: 0.7520\n",
      "Epoch 194/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5862 - f1_weighted: 0.7402 - accuracy: 0.7484\n",
      "Epoch 195/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5861 - f1_weighted: 0.7491 - accuracy: 0.7474\n",
      "Epoch 196/200\n",
      "56/56 [==============================] - 0s 853us/step - loss: 0.5903 - f1_weighted: 0.7407 - accuracy: 0.7467\n",
      "Epoch 197/200\n",
      "56/56 [==============================] - 0s 992us/step - loss: 0.5925 - f1_weighted: 0.7410 - accuracy: 0.7391\n",
      "Epoch 198/200\n",
      "56/56 [==============================] - 0s 944us/step - loss: 0.5800 - f1_weighted: 0.7546 - accuracy: 0.7559\n",
      "Epoch 199/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5819 - f1_weighted: 0.7432 - accuracy: 0.7517\n",
      "Epoch 200/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5876 - f1_weighted: 0.7441 - accuracy: 0.7495\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\lenovo\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\keras\\engine\\sequential.py:450: UserWarning: `model.predict_classes()` is deprecated and will be removed after 2021-01-01. Please use instead:* `np.argmax(model.predict(x), axis=-1)`,   if your model does multi-class classification   (e.g. if it uses a `softmax` last-layer activation).* `(model.predict(x) > 0.5).astype(\"int32\")`,   if your model does binary classification   (e.g. if it uses a `sigmoid` last-layer activation).\n",
      "  warnings.warn('`model.predict_classes()` is deprecated and '\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "56/56 [==============================] - 1s 1ms/step - loss: 1.0414 - f1_weighted: 0.4840 - accuracy: 0.4255\n",
      "Epoch 2/200\n",
      "56/56 [==============================] - 0s 853us/step - loss: 0.8238 - f1_weighted: 0.7159 - accuracy: 0.6526\n",
      "Epoch 3/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.7281 - f1_weighted: 0.7603 - accuracy: 0.6908\n",
      "Epoch 4/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6767 - f1_weighted: 0.7615 - accuracy: 0.7156\n",
      "Epoch 5/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6645 - f1_weighted: 0.7574 - accuracy: 0.7216\n",
      "Epoch 6/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6443 - f1_weighted: 0.7655 - accuracy: 0.7310\n",
      "Epoch 7/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6477 - f1_weighted: 0.7665 - accuracy: 0.7255\n",
      "Epoch 8/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6525 - f1_weighted: 0.7663 - accuracy: 0.7229\n",
      "Epoch 9/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6439 - f1_weighted: 0.7624 - accuracy: 0.7290\n",
      "Epoch 10/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6429 - f1_weighted: 0.7629 - accuracy: 0.7224\n",
      "Epoch 11/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6449 - f1_weighted: 0.7580 - accuracy: 0.7224\n",
      "Epoch 12/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6362 - f1_weighted: 0.7630 - accuracy: 0.7285\n",
      "Epoch 13/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6393 - f1_weighted: 0.7677 - accuracy: 0.7271\n",
      "Epoch 14/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6406 - f1_weighted: 0.7658 - accuracy: 0.7271\n",
      "Epoch 15/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6251 - f1_weighted: 0.7637 - accuracy: 0.7346\n",
      "Epoch 16/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6367 - f1_weighted: 0.7681 - accuracy: 0.7293\n",
      "Epoch 17/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6367 - f1_weighted: 0.7656 - accuracy: 0.7239\n",
      "Epoch 18/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6339 - f1_weighted: 0.7681 - accuracy: 0.7247\n",
      "Epoch 19/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6223 - f1_weighted: 0.7661 - accuracy: 0.7327\n",
      "Epoch 20/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6281 - f1_weighted: 0.7625 - accuracy: 0.7312\n",
      "Epoch 21/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6422 - f1_weighted: 0.7685 - accuracy: 0.7219\n",
      "Epoch 22/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6380 - f1_weighted: 0.7627 - accuracy: 0.7245\n",
      "Epoch 23/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6258 - f1_weighted: 0.7688 - accuracy: 0.7378\n",
      "Epoch 24/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6180 - f1_weighted: 0.7627 - accuracy: 0.7335\n",
      "Epoch 25/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6276 - f1_weighted: 0.7601 - accuracy: 0.7310\n",
      "Epoch 26/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6283 - f1_weighted: 0.7666 - accuracy: 0.7314\n",
      "Epoch 27/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6189 - f1_weighted: 0.7615 - accuracy: 0.7370\n",
      "Epoch 28/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6176 - f1_weighted: 0.7668 - accuracy: 0.7360\n",
      "Epoch 29/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6290 - f1_weighted: 0.7631 - accuracy: 0.7265\n",
      "Epoch 30/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6154 - f1_weighted: 0.7651 - accuracy: 0.7341\n",
      "Epoch 31/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6191 - f1_weighted: 0.7649 - accuracy: 0.7343\n",
      "Epoch 32/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6192 - f1_weighted: 0.7675 - accuracy: 0.7341\n",
      "Epoch 33/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6170 - f1_weighted: 0.7607 - accuracy: 0.7337\n",
      "Epoch 34/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6282 - f1_weighted: 0.7684 - accuracy: 0.7238\n",
      "Epoch 35/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6190 - f1_weighted: 0.7647 - accuracy: 0.7354\n",
      "Epoch 36/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6195 - f1_weighted: 0.7685 - accuracy: 0.7369\n",
      "Epoch 37/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6230 - f1_weighted: 0.7666 - accuracy: 0.7309\n",
      "Epoch 38/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6173 - f1_weighted: 0.7659 - accuracy: 0.7319\n",
      "Epoch 39/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6241 - f1_weighted: 0.7606 - accuracy: 0.7259\n",
      "Epoch 40/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6192 - f1_weighted: 0.7687 - accuracy: 0.7283\n",
      "Epoch 41/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6216 - f1_weighted: 0.7700 - accuracy: 0.7305\n",
      "Epoch 42/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6134 - f1_weighted: 0.7686 - accuracy: 0.7359\n",
      "Epoch 43/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6242 - f1_weighted: 0.7647 - accuracy: 0.7287\n",
      "Epoch 44/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6081 - f1_weighted: 0.7646 - accuracy: 0.7416\n",
      "Epoch 45/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6168 - f1_weighted: 0.7699 - accuracy: 0.7361\n",
      "Epoch 46/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6165 - f1_weighted: 0.7595 - accuracy: 0.7355\n",
      "Epoch 47/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6153 - f1_weighted: 0.7658 - accuracy: 0.7348\n",
      "Epoch 48/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6063 - f1_weighted: 0.7575 - accuracy: 0.7389\n",
      "Epoch 49/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6046 - f1_weighted: 0.7607 - accuracy: 0.7426\n",
      "Epoch 50/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6161 - f1_weighted: 0.7643 - accuracy: 0.7381\n",
      "Epoch 51/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6234 - f1_weighted: 0.7644 - accuracy: 0.7323\n",
      "Epoch 52/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6112 - f1_weighted: 0.7592 - accuracy: 0.7362\n",
      "Epoch 53/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6152 - f1_weighted: 0.7716 - accuracy: 0.7327\n",
      "Epoch 54/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6132 - f1_weighted: 0.7645 - accuracy: 0.7374\n",
      "Epoch 55/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6108 - f1_weighted: 0.7636 - accuracy: 0.7342\n",
      "Epoch 56/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6156 - f1_weighted: 0.7670 - accuracy: 0.7317\n",
      "Epoch 57/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6157 - f1_weighted: 0.7525 - accuracy: 0.7318\n",
      "Epoch 58/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6126 - f1_weighted: 0.7711 - accuracy: 0.7382\n",
      "Epoch 59/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6135 - f1_weighted: 0.7582 - accuracy: 0.7362\n",
      "Epoch 60/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6124 - f1_weighted: 0.7577 - accuracy: 0.7388\n",
      "Epoch 61/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6139 - f1_weighted: 0.7738 - accuracy: 0.7371\n",
      "Epoch 62/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6120 - f1_weighted: 0.7576 - accuracy: 0.7325\n",
      "Epoch 63/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.6165 - f1_weighted: 0.7618 - accuracy: 0.7349\n",
      "Epoch 64/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6054 - f1_weighted: 0.7629 - accuracy: 0.7357\n",
      "Epoch 65/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6171 - f1_weighted: 0.7624 - accuracy: 0.7337\n",
      "Epoch 66/200\n",
      "56/56 [==============================] - 0s 853us/step - loss: 0.6134 - f1_weighted: 0.7609 - accuracy: 0.7295\n",
      "Epoch 67/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6025 - f1_weighted: 0.7579 - accuracy: 0.7381\n",
      "Epoch 68/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6080 - f1_weighted: 0.7573 - accuracy: 0.7379\n",
      "Epoch 69/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6200 - f1_weighted: 0.7630 - accuracy: 0.7287\n",
      "Epoch 70/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6040 - f1_weighted: 0.7606 - accuracy: 0.7385\n",
      "Epoch 71/200\n",
      "56/56 [==============================] - 0s 964us/step - loss: 0.6064 - f1_weighted: 0.7666 - accuracy: 0.7374\n",
      "Epoch 72/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6117 - f1_weighted: 0.7622 - accuracy: 0.7393\n",
      "Epoch 73/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6033 - f1_weighted: 0.7618 - accuracy: 0.7380\n",
      "Epoch 74/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6082 - f1_weighted: 0.7608 - accuracy: 0.7386\n",
      "Epoch 75/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6078 - f1_weighted: 0.7546 - accuracy: 0.7311\n",
      "Epoch 76/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6034 - f1_weighted: 0.7582 - accuracy: 0.7408\n",
      "Epoch 77/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6009 - f1_weighted: 0.7649 - accuracy: 0.7432\n",
      "Epoch 78/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6100 - f1_weighted: 0.7669 - accuracy: 0.7365\n",
      "Epoch 79/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6054 - f1_weighted: 0.7717 - accuracy: 0.7385\n",
      "Epoch 80/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6106 - f1_weighted: 0.7634 - accuracy: 0.7350\n",
      "Epoch 81/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6114 - f1_weighted: 0.7532 - accuracy: 0.7376\n",
      "Epoch 82/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6104 - f1_weighted: 0.7544 - accuracy: 0.7355\n",
      "Epoch 83/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6090 - f1_weighted: 0.7588 - accuracy: 0.7394\n",
      "Epoch 84/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6020 - f1_weighted: 0.7545 - accuracy: 0.7454\n",
      "Epoch 85/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5969 - f1_weighted: 0.7560 - accuracy: 0.7429\n",
      "Epoch 86/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6119 - f1_weighted: 0.7563 - accuracy: 0.7393\n",
      "Epoch 87/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6198 - f1_weighted: 0.7622 - accuracy: 0.7318\n",
      "Epoch 88/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6004 - f1_weighted: 0.7545 - accuracy: 0.7384\n",
      "Epoch 89/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6107 - f1_weighted: 0.7593 - accuracy: 0.7379\n",
      "Epoch 90/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6026 - f1_weighted: 0.7592 - accuracy: 0.7412\n",
      "Epoch 91/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6028 - f1_weighted: 0.7576 - accuracy: 0.7425\n",
      "Epoch 92/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5987 - f1_weighted: 0.7542 - accuracy: 0.7402\n",
      "Epoch 93/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6037 - f1_weighted: 0.7508 - accuracy: 0.7396\n",
      "Epoch 94/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6099 - f1_weighted: 0.7581 - accuracy: 0.7359\n",
      "Epoch 95/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6014 - f1_weighted: 0.7502 - accuracy: 0.7422\n",
      "Epoch 96/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6050 - f1_weighted: 0.7586 - accuracy: 0.7386\n",
      "Epoch 97/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6088 - f1_weighted: 0.7553 - accuracy: 0.7346\n",
      "Epoch 98/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6003 - f1_weighted: 0.7508 - accuracy: 0.7482\n",
      "Epoch 99/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5965 - f1_weighted: 0.7655 - accuracy: 0.7489\n",
      "Epoch 100/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5986 - f1_weighted: 0.7566 - accuracy: 0.7423\n",
      "Epoch 101/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5999 - f1_weighted: 0.7517 - accuracy: 0.7424\n",
      "Epoch 102/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5984 - f1_weighted: 0.7568 - accuracy: 0.7413\n",
      "Epoch 103/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6092 - f1_weighted: 0.7524 - accuracy: 0.7386\n",
      "Epoch 104/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5961 - f1_weighted: 0.7638 - accuracy: 0.7476\n",
      "Epoch 105/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6088 - f1_weighted: 0.7502 - accuracy: 0.7388\n",
      "Epoch 106/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6086 - f1_weighted: 0.7576 - accuracy: 0.7395\n",
      "Epoch 107/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5975 - f1_weighted: 0.7593 - accuracy: 0.7417\n",
      "Epoch 108/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6064 - f1_weighted: 0.7595 - accuracy: 0.7374\n",
      "Epoch 109/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6016 - f1_weighted: 0.7512 - accuracy: 0.7448\n",
      "Epoch 110/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5980 - f1_weighted: 0.7538 - accuracy: 0.7422\n",
      "Epoch 111/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6039 - f1_weighted: 0.7523 - accuracy: 0.7405\n",
      "Epoch 112/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5988 - f1_weighted: 0.7544 - accuracy: 0.7429\n",
      "Epoch 113/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6108 - f1_weighted: 0.7521 - accuracy: 0.7377\n",
      "Epoch 114/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6047 - f1_weighted: 0.7578 - accuracy: 0.7385\n",
      "Epoch 115/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5923 - f1_weighted: 0.7512 - accuracy: 0.7476\n",
      "Epoch 116/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6027 - f1_weighted: 0.7508 - accuracy: 0.7389\n",
      "Epoch 117/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6001 - f1_weighted: 0.7567 - accuracy: 0.7454\n",
      "Epoch 118/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6046 - f1_weighted: 0.7500 - accuracy: 0.7402\n",
      "Epoch 119/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6006 - f1_weighted: 0.7588 - accuracy: 0.7412\n",
      "Epoch 120/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5962 - f1_weighted: 0.7502 - accuracy: 0.7409\n",
      "Epoch 121/200\n",
      "56/56 [==============================] - 0s 2ms/step - loss: 0.5951 - f1_weighted: 0.7476 - accuracy: 0.7441\n",
      "Epoch 122/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5922 - f1_weighted: 0.7518 - accuracy: 0.7478\n",
      "Epoch 123/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5997 - f1_weighted: 0.7472 - accuracy: 0.7468\n",
      "Epoch 124/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6090 - f1_weighted: 0.7555 - accuracy: 0.7388\n",
      "Epoch 125/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5973 - f1_weighted: 0.7470 - accuracy: 0.7444\n",
      "Epoch 126/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5961 - f1_weighted: 0.7586 - accuracy: 0.7425\n",
      "Epoch 127/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6046 - f1_weighted: 0.7532 - accuracy: 0.7408\n",
      "Epoch 128/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5960 - f1_weighted: 0.7539 - accuracy: 0.7447\n",
      "Epoch 129/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5942 - f1_weighted: 0.7548 - accuracy: 0.7499\n",
      "Epoch 130/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5985 - f1_weighted: 0.7496 - accuracy: 0.7440\n",
      "Epoch 131/200\n",
      "56/56 [==============================] - 0s 944us/step - loss: 0.6036 - f1_weighted: 0.7441 - accuracy: 0.7397\n",
      "Epoch 132/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5920 - f1_weighted: 0.7500 - accuracy: 0.7471\n",
      "Epoch 133/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5952 - f1_weighted: 0.7499 - accuracy: 0.7452\n",
      "Epoch 134/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5950 - f1_weighted: 0.7470 - accuracy: 0.7418\n",
      "Epoch 135/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5957 - f1_weighted: 0.7534 - accuracy: 0.7465\n",
      "Epoch 136/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5921 - f1_weighted: 0.7528 - accuracy: 0.7461\n",
      "Epoch 137/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5951 - f1_weighted: 0.7552 - accuracy: 0.7483\n",
      "Epoch 138/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5906 - f1_weighted: 0.7528 - accuracy: 0.7451\n",
      "Epoch 139/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5984 - f1_weighted: 0.7523 - accuracy: 0.7423\n",
      "Epoch 140/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5923 - f1_weighted: 0.7548 - accuracy: 0.7479\n",
      "Epoch 141/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5941 - f1_weighted: 0.7488 - accuracy: 0.7424\n",
      "Epoch 142/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5994 - f1_weighted: 0.7567 - accuracy: 0.7424\n",
      "Epoch 143/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5864 - f1_weighted: 0.7647 - accuracy: 0.7563\n",
      "Epoch 144/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5966 - f1_weighted: 0.7527 - accuracy: 0.7419\n",
      "Epoch 145/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5931 - f1_weighted: 0.7394 - accuracy: 0.7473\n",
      "Epoch 146/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5898 - f1_weighted: 0.7538 - accuracy: 0.7471\n",
      "Epoch 147/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5999 - f1_weighted: 0.7513 - accuracy: 0.7376\n",
      "Epoch 148/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5882 - f1_weighted: 0.7530 - accuracy: 0.7462\n",
      "Epoch 149/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5970 - f1_weighted: 0.7524 - accuracy: 0.7459\n",
      "Epoch 150/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5831 - f1_weighted: 0.7481 - accuracy: 0.7478\n",
      "Epoch 151/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5885 - f1_weighted: 0.7514 - accuracy: 0.7459\n",
      "Epoch 152/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5969 - f1_weighted: 0.7434 - accuracy: 0.7404\n",
      "Epoch 153/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5855 - f1_weighted: 0.7467 - accuracy: 0.7486\n",
      "Epoch 154/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5945 - f1_weighted: 0.7457 - accuracy: 0.7428\n",
      "Epoch 155/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6058 - f1_weighted: 0.7398 - accuracy: 0.7328\n",
      "Epoch 156/200\n",
      "56/56 [==============================] - 0s 907us/step - loss: 0.5968 - f1_weighted: 0.7511 - accuracy: 0.7474\n",
      "Epoch 157/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6051 - f1_weighted: 0.7516 - accuracy: 0.7417\n",
      "Epoch 158/200\n",
      "56/56 [==============================] - 0s 871us/step - loss: 0.5978 - f1_weighted: 0.7444 - accuracy: 0.7407\n",
      "Epoch 159/200\n",
      "56/56 [==============================] - 0s 944us/step - loss: 0.5893 - f1_weighted: 0.7427 - accuracy: 0.7462\n",
      "Epoch 160/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6062 - f1_weighted: 0.7415 - accuracy: 0.7386\n",
      "Epoch 161/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5959 - f1_weighted: 0.7455 - accuracy: 0.7443\n",
      "Epoch 162/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5931 - f1_weighted: 0.7442 - accuracy: 0.7455\n",
      "Epoch 163/200\n",
      "56/56 [==============================] - 0s 853us/step - loss: 0.5931 - f1_weighted: 0.7461 - accuracy: 0.7477\n",
      "Epoch 164/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5885 - f1_weighted: 0.7456 - accuracy: 0.7458\n",
      "Epoch 165/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5902 - f1_weighted: 0.7456 - accuracy: 0.7456\n",
      "Epoch 166/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5991 - f1_weighted: 0.7469 - accuracy: 0.7432\n",
      "Epoch 167/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5912 - f1_weighted: 0.7437 - accuracy: 0.7469\n",
      "Epoch 168/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.6019 - f1_weighted: 0.7469 - accuracy: 0.7360\n",
      "Epoch 169/200\n",
      "56/56 [==============================] - 0s 873us/step - loss: 0.5867 - f1_weighted: 0.7393 - accuracy: 0.7506\n",
      "Epoch 170/200\n",
      "56/56 [==============================] - 0s 944us/step - loss: 0.5973 - f1_weighted: 0.7426 - accuracy: 0.7394\n",
      "Epoch 171/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5889 - f1_weighted: 0.7535 - accuracy: 0.7481\n",
      "Epoch 172/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5947 - f1_weighted: 0.7397 - accuracy: 0.7438\n",
      "Epoch 173/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5990 - f1_weighted: 0.7451 - accuracy: 0.7405\n",
      "Epoch 174/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5880 - f1_weighted: 0.7443 - accuracy: 0.7482\n",
      "Epoch 175/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5906 - f1_weighted: 0.7387 - accuracy: 0.7480\n",
      "Epoch 176/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5876 - f1_weighted: 0.7447 - accuracy: 0.7480\n",
      "Epoch 177/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5872 - f1_weighted: 0.7431 - accuracy: 0.7527\n",
      "Epoch 178/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5885 - f1_weighted: 0.7438 - accuracy: 0.7464\n",
      "Epoch 179/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5972 - f1_weighted: 0.7495 - accuracy: 0.7449\n",
      "Epoch 180/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5982 - f1_weighted: 0.7449 - accuracy: 0.7475\n",
      "Epoch 181/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5909 - f1_weighted: 0.7486 - accuracy: 0.7445\n",
      "Epoch 182/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5970 - f1_weighted: 0.7419 - accuracy: 0.7500\n",
      "Epoch 183/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5856 - f1_weighted: 0.7540 - accuracy: 0.7507\n",
      "Epoch 184/200\n",
      "56/56 [==============================] - 0s 852us/step - loss: 0.5881 - f1_weighted: 0.7430 - accuracy: 0.7488\n",
      "Epoch 185/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5993 - f1_weighted: 0.7409 - accuracy: 0.7422\n",
      "Epoch 186/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6065 - f1_weighted: 0.7370 - accuracy: 0.7389\n",
      "Epoch 187/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5932 - f1_weighted: 0.7464 - accuracy: 0.7468\n",
      "Epoch 188/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5874 - f1_weighted: 0.7408 - accuracy: 0.7495\n",
      "Epoch 189/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6020 - f1_weighted: 0.7402 - accuracy: 0.7419\n",
      "Epoch 190/200\n",
      "56/56 [==============================] - 0s 853us/step - loss: 0.5898 - f1_weighted: 0.7478 - accuracy: 0.7471\n",
      "Epoch 191/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5860 - f1_weighted: 0.7290 - accuracy: 0.7478\n",
      "Epoch 192/200\n",
      "56/56 [==============================] - 0s 853us/step - loss: 0.5886 - f1_weighted: 0.7459 - accuracy: 0.7450\n",
      "Epoch 193/200\n",
      "56/56 [==============================] - 0s 971us/step - loss: 0.5935 - f1_weighted: 0.7477 - accuracy: 0.7479\n",
      "Epoch 194/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5950 - f1_weighted: 0.7442 - accuracy: 0.7441\n",
      "Epoch 195/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5970 - f1_weighted: 0.7501 - accuracy: 0.7467\n",
      "Epoch 196/200\n",
      "56/56 [==============================] - 0s 1ms/step - loss: 0.6004 - f1_weighted: 0.7405 - accuracy: 0.7429\n",
      "Epoch 197/200\n",
      "56/56 [==============================] - 0s 796us/step - loss: 0.5918 - f1_weighted: 0.7421 - accuracy: 0.7495\n",
      "Epoch 198/200\n",
      "56/56 [==============================] - 0s 852us/step - loss: 0.5905 - f1_weighted: 0.7358 - accuracy: 0.7464\n",
      "Epoch 199/200\n",
      "56/56 [==============================] - 0s 991us/step - loss: 0.5917 - f1_weighted: 0.7473 - accuracy: 0.7477\n",
      "Epoch 200/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "56/56 [==============================] - 0s 1ms/step - loss: 0.5841 - f1_weighted: 0.7468 - accuracy: 0.7476\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\lenovo\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\keras\\engine\\sequential.py:450: UserWarning: `model.predict_classes()` is deprecated and will be removed after 2021-01-01. Please use instead:* `np.argmax(model.predict(x), axis=-1)`,   if your model does multi-class classification   (e.g. if it uses a `softmax` last-layer activation).* `(model.predict(x) > 0.5).astype(\"int32\")`,   if your model does binary classification   (e.g. if it uses a `sigmoid` last-layer activation).\n",
      "  warnings.warn('`model.predict_classes()` is deprecated and '\n"
     ]
    }
   ],
   "source": [
    "scores_ann = cross_validate(ann, X_train, y_train, cv=5,\n",
    "                    scoring=('precision_weighted','recall_weighted','balanced_accuracy','f1_weighted','roc_auc_ovr_weighted'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "id": "7e7dbd51",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cross-validation performance : 0.715\n"
     ]
    }
   ],
   "source": [
    "print(\"Cross-validation performance : {:.3f}\".format(scores_ann['test_f1_weighted'].mean()))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "480ba955",
   "metadata": {},
   "source": [
    "Le réseau de neurones n'est pas bien implémenté, il nécessite des modification pour être plus performant que les autres. On peut notamment utiliser tout le dataset original pour l'entraîner et avoir de meilleurs résultats."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b31239a1",
   "metadata": {},
   "source": [
    "## 4. Sélection de modèle :\n",
    "\n",
    "Nous allons maintenant choisir le modèle avec la meilleure performance par cross-validation, qui est le modèle XGBoost."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "ee84cf01",
   "metadata": {},
   "outputs": [],
   "source": [
    "model = xgb_random"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a045670a",
   "metadata": {},
   "source": [
    "## 5. Evaluation du modèle :\n",
    "\n",
    "On évalue maintenant le modèle sélectionné sur le **Test Set**."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "id": "1fdf67ef",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test set score : 0.74\n"
     ]
    }
   ],
   "source": [
    "print(\"Test set score : {:.2f}\".format(model.score(X_test, y_test)))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "47efe74c",
   "metadata": {},
   "source": [
    "Le modèle nous donne un score de 74% sur le test set ce qui est un très bon début. Le modèle se généralise bien aux données qu'il n'a jamais vu."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cc5e9816",
   "metadata": {},
   "source": [
    "### Rapport de classification et matrice de confusion :\n",
    "\n",
    "Nous regardons ci-dessous que le modèle rencontre des difficultés à différencier les crimes de la catégorie ODATP de la catégorie ODAP et se trompe très souvent."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "id": "43f8f488",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "       ODATP       0.79      0.52      0.63      2204\n",
      "        ODAP       0.74      0.91      0.82      3157\n",
      "      OAPHSD       0.72      0.75      0.73       639\n",
      "\n",
      "    accuracy                           0.75      6000\n",
      "   macro avg       0.75      0.73      0.73      6000\n",
      "weighted avg       0.76      0.75      0.74      6000\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn import metrics\n",
    "\n",
    "y_pred = model.predict(X_test)\n",
    "target_names = ['ODATP', 'ODAP', 'OAPHSD']\n",
    "print(metrics.classification_report(y_test, y_pred, target_names=target_names))\n",
    "sns.heatmap(metrics.confusion_matrix(y_test,y_pred),annot=True,fmt='3.0f',cmap='magma')\n",
    "plt.title('Confusion Matrix', y=1.05, size=15)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e8680593",
   "metadata": {},
   "source": [
    "Mais notre modèle reste quand même beaucoup plus performant qu'un *dummy classifier* qui prédit les classes uniformément."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "id": "959f18ff",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.dummy import DummyClassifier\n",
    "dum = DummyClassifier(strategy='uniform', random_state=0)\n",
    "dum = dum.fit(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "id": "b922b3c1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'matrice de confusion')"
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_pred_dum = dum.predict(X_test)\n",
    "sns.heatmap(metrics.confusion_matrix(y_test, y_pred_dum),annot=True,fmt='3.0f',cmap=\"magma\")\n",
    "plt.title('matrice de confusion')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "19aaee2d",
   "metadata": {},
   "source": [
    "## 6. Mise en production :\n",
    "\n",
    "Nous pouvons à présent donner à notre modèle un 'Location Description' (codé selon le Ordinal Encoder), si une arrestation a eu lieu ou pas, si le crime est domestique, l'heure, le jour du mois, le mois, l'année et la latitude et la longitude et il va nous prédire la catégorie du crime qui se produira avec la fonction `predict()`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "56b1a9fa",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_to_predict = []\n",
    "model.predict(y_to_predict)\n",
    "print(y_to_predict)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}