classificação_gravidade_acidentes_minas_gerais.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "CEtG6c4koygW"
   },
   "source": [
    "# Projeto #1 - Classificação\n",
    "\n",
    "<p style='text-align: justify;'><font size=3.>Análise de acidentes nas rodovias federais brasileira entre 2007 e 2021.</font></p>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "IGJiaCUD1ncA"
   },
   "source": [
    "### Bibliotecas básicas e outros imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-31T14:02:01.998603Z",
     "start_time": "2022-01-31T14:02:00.270389Z"
    },
    "id": "4bG6r2Bwoyge"
   },
   "outputs": [],
   "source": [
    "import random\n",
    "import warnings\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from scipy import stats\n",
    "import matplotlib.pyplot as plt\n",
    "from xgboost import XGBClassifier\n",
    "from imblearn.over_sampling import SMOTE\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "from sklearn.neural_network import MLPClassifier\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.ensemble import ExtraTreesClassifier, RandomForestClassifier\n",
    "from sklearn.metrics import plot_confusion_matrix, classification_report, accuracy_score\n",
    "\n",
    "%run modules/text.py\n",
    "%matplotlib inline\n",
    "plt.rcParams['font.family'] = 'Arial, Helvetica, sans-serif'\n",
    "\n",
    "warnings.filterwarnings('ignore')\n",
    "pd.set_option('display.max_rows', 1000)\n",
    "pd.set_option('display.max_columns', 200)\n",
    "np.set_printoptions(suppress=True, precision=3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "rjtF8niV1ncF"
   },
   "source": [
    "### Carregamento dos dados"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-31T14:02:08.593186Z",
     "start_time": "2022-01-31T14:02:02.000806Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 134
    },
    "id": "H8Xxqiwcoygh",
    "outputId": "948da79f-e337-4029-b318-a0ef40832d18"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Linhas: 1674465 | Coluna:  30\n"
     ]
    }
   ],
   "source": [
    "df = pd.read_csv('dataset/classifier_acidentes_rodovias_brasileiras_2007_a_2021.csv', low_memory=False, encoding='utf-8')\n",
    "print(f'Linhas: {df.shape[0]} | Coluna:  {df.shape[1]}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Seleção dos dados para o estado MG"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-31T14:02:08.799885Z",
     "start_time": "2022-01-31T14:02:08.594637Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Linhas: 247397 | Coluna:  30\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>id</th>\n",
       "      <th>dia_semana</th>\n",
       "      <th>uf</th>\n",
       "      <th>br</th>\n",
       "      <th>km</th>\n",
       "      <th>municipio</th>\n",
       "      <th>causa_acidente</th>\n",
       "      <th>tipo_acidente</th>\n",
       "      <th>classe</th>\n",
       "      <th>fase_dia</th>\n",
       "      <th>sentido_via</th>\n",
       "      <th>condicao_metereologica</th>\n",
       "      <th>tipo_pista</th>\n",
       "      <th>tracado_via</th>\n",
       "      <th>uso_solo</th>\n",
       "      <th>tipo_veiculo</th>\n",
       "      <th>tipo_envolvido</th>\n",
       "      <th>estado_fisico</th>\n",
       "      <th>idade</th>\n",
       "      <th>sexo</th>\n",
       "      <th>data</th>\n",
       "      <th>mes</th>\n",
       "      <th>ano</th>\n",
       "      <th>feriado</th>\n",
       "      <th>faixa_etaria</th>\n",
       "      <th>porte_veiculo</th>\n",
       "      <th>condicao_clima</th>\n",
       "      <th>estacao_do_ano</th>\n",
       "      <th>regiao_pais</th>\n",
       "      <th>feriado_agrupado</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>83123235.0</td>\n",
       "      <td>quarta</td>\n",
       "      <td>MG</td>\n",
       "      <td>381</td>\n",
       "      <td>491.0</td>\n",
       "      <td>Betim</td>\n",
       "      <td>falta de atencao</td>\n",
       "      <td>colisao traseira</td>\n",
       "      <td>sem vitimas</td>\n",
       "      <td>plena noite</td>\n",
       "      <td>crescente</td>\n",
       "      <td>nublado</td>\n",
       "      <td>dupla</td>\n",
       "      <td>reta</td>\n",
       "      <td>urbano</td>\n",
       "      <td>automovel</td>\n",
       "      <td>condutor</td>\n",
       "      <td>ileso</td>\n",
       "      <td>24</td>\n",
       "      <td>masculino</td>\n",
       "      <td>2014-01-01</td>\n",
       "      <td>1</td>\n",
       "      <td>2014</td>\n",
       "      <td>ano novo</td>\n",
       "      <td>jovem</td>\n",
       "      <td>pequeno porte</td>\n",
       "      <td>nublado</td>\n",
       "      <td>verão</td>\n",
       "      <td>sudeste</td>\n",
       "      <td>feriado</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           id dia_semana  uf   br     km municipio    causa_acidente  \\\n",
       "0  83123235.0     quarta  MG  381  491.0     Betim  falta de atencao   \n",
       "\n",
       "      tipo_acidente       classe     fase_dia sentido_via  \\\n",
       "0  colisao traseira  sem vitimas  plena noite   crescente   \n",
       "\n",
       "  condicao_metereologica tipo_pista tracado_via uso_solo tipo_veiculo  \\\n",
       "0                nublado      dupla        reta   urbano    automovel   \n",
       "\n",
       "  tipo_envolvido estado_fisico  idade       sexo        data  mes   ano  \\\n",
       "0       condutor         ileso     24  masculino  2014-01-01    1  2014   \n",
       "\n",
       "    feriado faixa_etaria  porte_veiculo condicao_clima estacao_do_ano  \\\n",
       "0  ano novo        jovem  pequeno porte        nublado          verão   \n",
       "\n",
       "  regiao_pais feriado_agrupado  \n",
       "0     sudeste          feriado  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "acidentes_minas = df.query('uf==\"MG\"').reset_index(drop=True)\n",
    "acidentes_minas.rename(columns={'classificacao_acidente': 'classe'}, inplace=True)\n",
    "print(f'Linhas: {acidentes_minas.shape[0]} | Coluna:  {acidentes_minas.shape[1]}')\n",
    "acidentes_minas.head(1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Contagem dos dados"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-31T14:02:08.825595Z",
     "start_time": "2022-01-31T14:02:08.801809Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Proporção de dados na base:\n",
      "sem vitimas            51.8\n",
      "com vitimas feridas    44.1\n",
      "com vitimas fatais      4.1\n",
      "Name: classe, dtype: float64\n"
     ]
    }
   ],
   "source": [
    "print('Proporção de dados na base:')\n",
    "print((acidentes_minas['classe'].value_counts(normalize=True)*100).round(1))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Seleção de features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-31T14:02:08.866101Z",
     "start_time": "2022-01-31T14:02:08.826759Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>dia_semana</th>\n",
       "      <th>fase_dia</th>\n",
       "      <th>tipo_pista</th>\n",
       "      <th>sentido_via</th>\n",
       "      <th>tracado_via</th>\n",
       "      <th>condicao_metereologica</th>\n",
       "      <th>tipo_acidente</th>\n",
       "      <th>tipo_veiculo</th>\n",
       "      <th>br</th>\n",
       "      <th>km</th>\n",
       "      <th>sexo</th>\n",
       "      <th>faixa_etaria</th>\n",
       "      <th>classe</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>quarta</td>\n",
       "      <td>plena noite</td>\n",
       "      <td>dupla</td>\n",
       "      <td>crescente</td>\n",
       "      <td>reta</td>\n",
       "      <td>nublado</td>\n",
       "      <td>colisao traseira</td>\n",
       "      <td>automovel</td>\n",
       "      <td>381</td>\n",
       "      <td>491.0</td>\n",
       "      <td>masculino</td>\n",
       "      <td>jovem</td>\n",
       "      <td>sem vitimas</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  dia_semana     fase_dia tipo_pista sentido_via tracado_via  \\\n",
       "0     quarta  plena noite      dupla   crescente        reta   \n",
       "\n",
       "  condicao_metereologica     tipo_acidente tipo_veiculo   br     km  \\\n",
       "0                nublado  colisao traseira    automovel  381  491.0   \n",
       "\n",
       "        sexo faixa_etaria       classe  \n",
       "0  masculino        jovem  sem vitimas  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "colunas = [\n",
    "    'dia_semana',\n",
    "    'fase_dia',\n",
    "    'tipo_pista',\n",
    "    'sentido_via',\n",
    "    'tracado_via',\n",
    "    'condicao_metereologica',\n",
    "    'tipo_acidente',\n",
    "    'tipo_veiculo',\n",
    "    'br',\n",
    "    'km',\n",
    "    'sexo',\n",
    "    'faixa_etaria',\n",
    "    'classe'\n",
    "]\n",
    "dados = acidentes_minas[colunas]\n",
    "dados.head(1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Separação dos dados"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-31T14:02:08.894814Z",
     "start_time": "2022-01-31T14:02:08.867350Z"
    }
   },
   "outputs": [],
   "source": [
    "X = dados.drop(['classe'], axis=1)\n",
    "y = dados['classe']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "lang": "en"
   },
   "source": [
    "### Label encoder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-31T14:02:08.947038Z",
     "start_time": "2022-01-31T14:02:08.896853Z"
    }
   },
   "outputs": [],
   "source": [
    "le = LabelEncoder()\n",
    "y = le.fit_transform(y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Vetorização dos dados categóricos"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-31T14:02:09.411490Z",
     "start_time": "2022-01-31T14:02:08.948378Z"
    }
   },
   "outputs": [],
   "source": [
    "encoder = LabelEncoder()\n",
    "for col in X.columns:\n",
    "    if col != 'br' and col!='km':\n",
    "        X[col] = encoder.fit_transform(X[col].astype(str))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Balanceamento da base"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-31T14:02:12.353764Z",
     "start_time": "2022-01-31T14:02:09.412763Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "384504\n"
     ]
    }
   ],
   "source": [
    "smote = SMOTE(random_state=32)\n",
    "X_smote_res, y_smote_res = smote.fit_resample(X, y)\n",
    "print(len(X_smote_res))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Divisão treino e teste"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-31T14:02:12.405205Z",
     "start_time": "2022-01-31T14:02:12.357118Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape treino: 269152 | Shape teste: 115352\n"
     ]
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X_smote_res, y_smote_res, test_size=0.3, shuffle=True)\n",
    "print(f'Shape treino: {X_train.shape[0]} | Shape teste: {X_test.shape[0]}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Testando famílias de algoritmos"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-31T10:33:48.744378Z",
     "start_time": "2022-01-31T10:01:11.995055Z"
    },
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Resultado para o classificador RandomForestClassifier(max_depth=3, n_estimators=700)\n",
      "0.578654899785006\n",
      "----------------------------------------------\n",
      "Resultado para o classificador ExtraTreesClassifier(max_depth=3, n_estimators=700)\n",
      "0.5543033497468618\n",
      "----------------------------------------------\n",
      "Resultado para o classificador KNeighborsClassifier(n_neighbors=10)\n",
      "0.656000762882308\n",
      "----------------------------------------------\n",
      "Resultado para o classificador XGBClassifier(base_score=0.5, booster=None, colsample_bylevel=1,\n",
      "              colsample_bynode=1, colsample_bytree=1, gamma=0, gpu_id=-1,\n",
      "              importance_type='gain', interaction_constraints=None,\n",
      "              learning_rate=0.5, max_delta_step=0, max_depth=4,\n",
      "              min_child_weight=8, missing=nan, monotone_constraints=None,\n",
      "              n_estimators=100, n_jobs=1, nthread=1, num_parallel_tree=1,\n",
      "              objective='multi:softprob', random_state=0, reg_alpha=0,\n",
      "              reg_lambda=1, scale_pos_weight=None, subsample=0.8,\n",
      "              tree_method=None, validate_parameters=False, verbosity=None)\n",
      "0.7072525833969069\n",
      "----------------------------------------------\n",
      "Resultado para o classificador MLPClassifier(alpha=1e-05, batch_size=100, hidden_layer_sizes=[512, 256, 128],\n",
      "              max_iter=35)\n",
      "0.6268465219502046\n",
      "----------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "estimators = [\n",
    "    RandomForestClassifier(n_estimators=700, max_depth=3),\n",
    "    ExtraTreesClassifier(n_estimators=700, max_depth=3),\n",
    "    KNeighborsClassifier(n_neighbors=10),\n",
    "    XGBClassifier(learning_rate=0.5, max_depth=4, min_child_weight=8, nthread=1, subsample=0.8),   \n",
    "    MLPClassifier(hidden_layer_sizes=[512, 256, 128], activation='relu', solver='adam', alpha=1e-5, max_iter=35, batch_size=100, shuffle=True, verbose=False)\n",
    "]\n",
    "\n",
    "for clf in estimators:\n",
    "    clf.fit(X_train, y_train)\n",
    "    y_pred = clf.predict(X_test)\n",
    "    print(f'Resultado para o classificador {clf}')\n",
    "    print(accuracy_score(y_test, y_pred))\n",
    "    print('----------------------------------------------')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Teste pareado Student’s t-test \n",
    "> Para comparação do desempenho dos modelos de ML\n",
    "\n",
    "\n",
    "**H0**: Não há diferença entre o desempenho dos dois modelos\n",
    "\n",
    "**H1**: Há diferença entre o desempenho dos dois modelos"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "start_time": "2022-01-31T10:00:59.176Z"
    }
   },
   "outputs": [],
   "source": [
    "clf_mlp = MLPClassifier(hidden_layer_sizes=[512, 256, 128], max_iter=35, activation='relu', solver='adam', alpha=1e-5, batch_size=100, shuffle=True, verbose=False)\n",
    "clf_xgb = XGBClassifier(learning_rate=0.5, max_depth=4, min_child_weight=8, nthread=1, subsample=0.8)\n",
    "\n",
    "XGB_score, RF_score = [], []\n",
    "\n",
    "for i in range(10):\n",
    "    print(f'Iniciando idx: {i}')\n",
    "    random_state = random.randint(42, 950)\n",
    "    print(f'Random state: {random_state}')\n",
    "    \n",
    "    X_train, X_test, y_train, y_test = train_test_split(X_smote_res, \n",
    "                                                        y_smote_res, \n",
    "                                                        test_size=0.3,\n",
    "                                                        shuffle=True,\n",
    "                                                        random_state=random_state)\n",
    "    \n",
    "    scores_mlp = cross_val_score(clf_mlp, X, y, cv=5, scoring='accuracy', verbose=False, n_jobs=2)\n",
    "    RF_score.append(scores_mlp.mean())\n",
    "    scores_xgb = cross_val_score(clf_xgb, X, y, cv=5, scoring='accuracy', verbose=False, n_jobs=2)\n",
    "    XGB_score.append(scores_xgb.mean())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-28T17:53:51.217081Z",
     "start_time": "2022-01-28T17:53:51.211118Z"
    }
   },
   "source": [
    "Comparando o desempenho"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "start_time": "2022-01-31T10:00:59.178Z"
    }
   },
   "outputs": [],
   "source": [
    "alpha = 0.05\n",
    "_, p_value = stats.ttest_rel(XGB_score, MLP_score)\n",
    "print(f'Teste pareado Student’s t-test: {p_value}')\n",
    "\n",
    "if p_value < alpha:\n",
    "    print(f'Rejeita H0. Temos evidências estatísticas de que a floresta aleatória e o XGB têm desempenho diferente.')\n",
    "else:\n",
    "    print(f'Não rejeita H0. Não há diferença de desempenho entre os modelos.')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Modelagem dos dados"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Modelo"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-31T14:02:13.601763Z",
     "start_time": "2022-01-31T14:02:13.598517Z"
    }
   },
   "outputs": [],
   "source": [
    "clf = XGBClassifier(learning_rate=0.5, max_depth=4, min_child_weight=8, nthread=1, subsample=0.8)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Validação cruzada"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-31T14:04:51.220812Z",
     "start_time": "2022-01-31T14:02:13.603318Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=2)]: Using backend LokyBackend with 2 concurrent workers.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.69 (+/- 0.06)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=2)]: Done   5 out of   5 | elapsed:  2.6min finished\n"
     ]
    }
   ],
   "source": [
    "scores = cross_val_score(clf, X_smote_res, y_smote_res, cv=5, scoring='accuracy', verbose=True, n_jobs=2)\n",
    "print('Accuracy: %0.2f (+/- %0.2f)' % (scores.mean(), scores.std() * 2))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Relatório de classificação"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-31T14:05:34.163508Z",
     "start_time": "2022-01-31T14:04:51.223107Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                     precision    recall  f1-score   support\n",
      "\n",
      " com vitimas fatais       0.81      0.82      0.82     38336\n",
      "com vitimas feridas       0.65      0.51      0.57     38465\n",
      "        sem vitimas       0.66      0.79      0.72     38551\n",
      "\n",
      "           accuracy                           0.71    115352\n",
      "          macro avg       0.71      0.71      0.70    115352\n",
      "       weighted avg       0.71      0.71      0.70    115352\n",
      "\n"
     ]
    }
   ],
   "source": [
    "clf.fit(X_train, y_train)\n",
    "y_pred = clf.predict(X_test)\n",
    "print(classification_report(y_test, y_pred, target_names=le.classes_.tolist()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Matriz de confusão"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-31T14:05:35.956713Z",
     "start_time": "2022-01-31T14:05:34.165426Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAb0AAAFbCAYAAABBM1DbAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAAA91ElEQVR4nO3deZwU1bnG8d8zwyoqIDsCol5cEBQREDfEXdRoogZFs5jEePVqzDUxZr2KxiXRJCbuS2IwajRxR0Vxxd0IIrKKIoqAKKCAoijLvPePqhl7xll6pmdr+vn66Y9dVadOvVXT9Nvn1KkqRQRmZmaFoKipAzAzM2ssTnpmZlYwnPTMzKxgOOmZmVnBcNIzM7OC4aRnZmYFw0nPvkLSOEkP1nOdMyWNrc86G4KkTSTdJWmVpJDUt57qPV3SSklnSTpe0kX1UW++qfjZaojPWiXbfFDSuIbchuUPJ71mKP0iCEl/q2TZ79NltfqiSNc5NsviPwa+VZv6NyLfB0YAewM9gIX1VO8xwGHAnsD5wG31VG++K+TPmjWBFk0dgFVpITBa0pkR8SmApBbAd4B3G2KDaf0bImJVQ9SfJ/4LmBMRM+qz0ojYP337Qn3WW98yPgONcteKAv+sWRNwS6/5mg68CYzOmHc48DkwKbOgpKGSHpW0XNLHkp6TtEfG8nfSt3emLb530vlj027HkyS9BXwBtMvscpI0Ml2n4qtcDBXi6SrpfklrJC2Q9P1KyrSXdIOkpZI+kfS0pCHVHRBJrSRdnNb5haT5ks7MWD5C0n8kfS7pA0mXS2qVsXySpGvSOpan2/6DpKLS5SQtjxGZ+yjpHUlnV4hlkqSrMqaPljQ93eeP0v3pli7bNj0e70v6VNJUSUdUqK+jpJslrUjreFzSTjUcj3ck/UbS9enffZGkn1Uo00fSvekx/kTSPZJ6ZSyv6jMQkk5L4/5M0huS9pPUS9LEdD+mSRqcUVcnSbencayRNEvS92rYh6w/a9nUr6R7epyk1eln4FeVbPNbkianx2OppDslbVldnLbxcNJr3v5G0t1W6vvA34GKv8I3A24B9gGGAdOACZI6pcuHpv//IUmX3dCMdbcGTgC+CexCklQzvZCuU/oaAqykQuKtYBxJi+lA4OskrdO+pQslCXgI2BI4AtgVeAZ4UlKPauq9Oa3rJ8COwA/SWEi/tB4GXk3r+wEwBrikQh0nAutJuhnPAP4XOC5ddjTJ8X0x3dejq4mljKTuwB1pfDuSdI/eklFk0zS2g0iO8d3APZJ2yCgzDtgdOIrkb/gZ8IiktjVs/ixgBjAY+D1wqdIfPGkyvx/oBuyXvnoC96V/g1JVfQZ+k+7XLsCU9P3fgGtIjvF7adyl2gBTSf6mOwF/Aa6XdEAN+1Cqps9aNvX/geQ4HwMckMY5osJ2WgHnpft1BNAZuD3LGC3fRYRfzexF8kXyINARWAP0A7qT/ArvU7q8mvUFLAG+lTEvgGMrlBsLrAO6Vbb9SuptS/Lldw+gKra9XbqtvTLmbQVsAMam0/sDq4G2FdadBpxTRb390noPrWL5RSQt46KMeSelx2yTdHoS8GKF9R4D/poxfRUwqUKZd4CzK8ybBFyVvh+cxrZVLf7GLwG/qbBvIzKWtwdWASdXU8c7wO0V5r2ZUe9B6XHvm7F8G6AEOLCGz0AAl2RMD0jn/SRj3sh0XudqYryjwvEt99nK5bNWsX6SHxdfACdmLN+UJHGOq6aOHdL96FWbf6d+5efLLb1mLCJWAPeStPC+S/Jl/JXzeUq6E69Pu6BWAZ8AXUkSZE0WRcQHNRVKWwbjgGLg25F+W1RiR5Iv1Zcz9mMBSaug1G7AJsCytBtqtaTVJF+s21ZR765pvU9Vs92XIqIkY95zJL/q/ytj3vQK671Hcqxy8RrwODBT0t1pt2CX0oWS2km6VNLstPtyNUkrpvTvU3rMXixdJ5JzXTOA/jVsu7r92RF4LyLeyah3floms96qPgOZdZcun1HJvK4Akool/Trt5v0w3c+jye5zWKaqz1oW9W9L8vfOPI6rK8SMpMFpt+0CSZ+QJFdqG6flJw9kaf5uIuk2Ww2cW0WZm0m6sM4i+fX/BfAEyRdATT7NMo5zSbqJhkY6sKYG1Q2EKCL5wtynkmUfZxlPbWTGsq6SZTX9+CshaT1nallWQcQGSQcDw4GDSbpWL5G0b0S8RtLldihwNklL7DPgH2T396lpQEld9qdivVX9PTPrjmrmlW7vbOCnJOdFZ5B8Zi+m9j8qqvqs5Vy/pHbARJIfKd8GlpJ0bz5Ldn8Py3Nu6TV/TwBrSf5h3ldFmb2BKyPioYiYRdLSq3hubB3JL+daU3KpwznAURGxqIbir5N8roZlrN+H5FxSqakkSbokIuZVeC2tot5pab37VbF8DjA8PY9Vam+SY/dWDTHXZBkZx1NSG5IusTKReDEizic5Z/oeX54r3Bv4R0TcHRHTgUWUb9HOIdm3zMFHmwMDgdk5xD0H6KmMaw0lbUPyt8il3qrsDTwQEbdExDSS475dbSqo4bNWU/1vkXzOh2fU146kB6HUDiT/ln4VEc9ExOvk3tK3POKk18ylXTs7A1tHxBdVFHsD+Jak/pKGkpznWFuhzDvAAZK6S+qY7fYlDSBpSf4KeDddv7ukLaqIdy7wCMkAgz0kDSLpqlqTUexx4HngfkmjJG2dlj1fUmWtPyLiDeDfwF8lHZOus4+kb6dFriH5Mr9G0o6SDgd+R3Le7bNs97cKTwInpqMLdyJpfZf1kkgarmQU5dA0wR8J9ObLxPIG8I20W20gcCvJoIzSfXuTZMDJ9ek+lZb5GPhnDnE/TtJFeZukIUpGx95G8qPjyRzqrcobJJ+xvdNBOleRDJLJShaftWrrT7sy/wb8XtJBGX+rzB9775L0hJwhaZv0c/Lbuu6w5R8nvTwQEZ9ERHXdft8nOWH/CknCu4kkyWX6KUkraSHJCMdsDSE5//ZnksExpa97qlnnJOBtki/WB0i+uMviSRP5YenyG4G5JAlte8qf+6voO2ldV5C0KMeRDPggIhYDo0jO/U0jOQa3k3yB5uqSNNb7gUdJzhVmHsNVwF4kg4/eBP4I/DYibk2X/4SkG+1ZklGcL6XvM32P5Dzo+PT/m5AM2llDHaXH+SiSlupT6et94OvVnJPNxYUksT9MMhr3U2p3EX5Nn7Vs6j+bZD/vTf8/My0LQEQsIzk//nWSHyXnkfx9rECoYT77ZmZmzY9bemZmVjCc9MzMrGA46ZmZWcFw0jMzs4LhpGdmZgXDSc/MzAqGk56ZmRUMJz0zMysYTnpmZlYwnPTMzKxgOOmZmVnBcNIzM7OC4aRnZmYFw0nPzMwKhpOemZkVDCc9MzMrGE56ZmZWMJz0zMysYLRo6gCscmrRNtRqs6YOoyAM3L53U4dQEIqkpg6hYEx79ZXlEdGlsbZXvPlWEevX1Hq9WLNsYkQc2gAhVclJr5lSq81ovf3opg6jIDz81J+aOoSCsEmr4qYOoWB0bNdiQWNuL9Z/Tusdjq/1ep+/emXnBginWk56ZmaWGwF50pJ30jMzs9wpP4aI5EeUZmbWvEm1f9VYpQ6VNFfSPEm/qGR5H0lPSXpV0nRJh9VUp5OemZnlSElLr7av6mqUioGrgVFAf2CMpP4Viv0G+HdE7AocD1xTU6ROemZmlrv6b+kNA+ZFxPyIWAvcARxVoUwAm6fv2wPv1VSpz+mZmVluRF3P6XWWNCVj+oaIuCF9vyWwMGPZImD3CuuPBR6V9COgHXBgTRt00jMzsxxld46uEssjYkgOGx4DjIuIP0raA7hF0oCIKKlqBXdvmplZc7QYyLxzRK90XqYfAP8GiIgXgTZAtdf+OemZmVnu6nkgCzAZ6Cdpa0mtSAaqjK9Q5l3gAABJO5IkvWXVVeruTTMzy109X5weEeslnQFMBIqBmyJilqQLgCkRMR74KXCjpLNIBrWcFBFRXb1OemZmliM1yMXpETEBmFBh3rkZ72cDe9WmTic9MzPLjW9DZmZmBSVPbkPmpGdmZjlqmO7NhuCkZ2ZmuSty96aZmRWCut+RpdE56ZmZWe48kMXMzAqDz+mZmVkhcUvPzMwKhlt6ZmZWELJ8Enpz4KRnZma5c0vPzMwKhlt6ZmZWGPJn9GZ+RGlmZlYP3NIzM7PcuXvTzMwKgm9DZmZmhSN/zuk56ZmZWe7cvWlmZgXDLT0zMysYbumZmVlBkM/pmZlZIXFLz8zMCoWc9MzMrBAIJz0zMysUSl95wEnPzMxypLxp6eXHcBszM2vWJNX6lUWdh0qaK2mepF9UsvxySdPS1xuSVtZUp1t6ZmaWs/pu6UkqBq4GDgIWAZMljY+I2aVlIuKsjPI/Anatqd5mlfQkdQDGRMS1TR2LlXfAHjtyyU+PpbioiFvuf4E/3/xYueW9unXkmrHfpv1mbSkuKuL8q+7nsRdmM3LYDpx3xpG0atmCtevWc+4V9/HslDeaaC/yw9Mvz+G3V93Hhg0lHHf4cE494YByy19+7S0uvPo+Xn9rCX8599uM2neXsmUnnXM902YvYMjAbfjrJSc3dujN3pMvzeHcP9/DhpISTvjacH707YPKLf9i7XrO/O2tTJ+7kI7t23H9Bd+ld49O3D1xCtf+88mycrPfeo9HbzqbAdv14pLrH+SuRyaz8pPPeOvxyxp7l5qNBujeHAbMi4j5af13AEcBs6soPwY4r6ZKm1v35hXATEmDJB1WOlPSkZU1bRuCpNaSHk+by8dVU+4kST2zqO8CSQfWb5SNq6hIXHbOaL7542sYPvpCjjl4N7bfunu5Mj/9waHc9/hU9v3W7/nBr//OH36eHLoPV65mzE+uZ68xF/M/59/Cded/pyl2IW9s2FDC2L/cw02/O4WJ437OA09M5c133i9Xpme3jlz68zF87YDBX1n/h8ftxx9/dWJjhZtXNmwo4Vd/vJPb/vjfPH3bL7nv8anMfbv8sb39wRdpv1lbXvz3/3HKcSO58JoHADjmkCE8fvM5PH7zOVx57rfo02MLBmzXC4CD9xrAhBt/0uj706yoji/oLGlKxuuUjFq3BBZmTC9K531189JWwNbAk5Utz9Rskp6kbsBdEfEsMAgoS3oRMT4iftdIoeyabnNQRPyrmnInATUmvYg4NyIer6fYmsRuO/Vl/sLlLFj8IevWb+Cex6Zy2L47ly8UwWbt2gCw+aZteX/5KgBmvLGo7P2ct5bQtnVLWrVsVh0Mzcprr7/LVj0706dnJ1q1bMER++/K48/PLFemV/ct2GHbnhQVffWX9V67bUe7TVo3Vrh55dU5C+jbqwtbbdmZVi1bcNQBg5n47IxyZR55diajDxsGwBEjd+HZV94gIsqVufexVzjqwC9/cOw2oC/dOrdv+B3YOC2PiCEZrxvqWM/xJPljQ00FGz3pSfqdpNMzpsdKOhtoC1wsqRVwAXBcaWsrbVVdlZYfJ+laSS9Jmi9ppKSbJM2RNC6j3mvTXw6zJJ1fYfuzJU2X9IcKsXUFbgWGptveVtK5kiZLminpBiWOBYYAt6Xl2lZWLiPeY2vadnPWo0t7Fn+womz6vQ9W0KNL+X/kv7thAqNHDWPmg7/l338+jXMuu/Mr9Ry5/yBem7uQtevWN3jM+eqD5avo0bVD2XT3Lh34IP3RYLl5f9kqtsw4tj26duD9ZasqlFlJz64dAWjRopjN27Xho1Wflisz/olX+cZBX21lFzJR+0EsWXSHLgZ6Z0z3SudV5njg9mxibYqW3r+A0RnTo9N5AETEWuBc4F/VtLY6AnsAZwHjgcuBnYCBkgalZX4dEUOAnYF9Je0sqRPwDWCniNgZuDCz0ohYCpwMPJtu+y3gqogYGhEDSBLzERFxFzAFODEtt6aycpl117TttMwppc38WL+m+qPYzBxzyBD++eBLDDji/xj9v9dy3fnfKfeh3mGb7oz90VGcdfEdTRilWW6mznqHtm1ascM2NXbyFJwGSHqTgX6Stk4bQ8eTfN9X3O4OJDnhxWzibPSkFxGvAl0l9ZS0C7AiIhbWtF4FD0TS5zAD+CAiZkRECTAL6JuWGS1pKvAqSULsD6wCPgf+Julo4LMstrWfpP9ImgHsn9ZVl3I1bjsibiht5qtF2yxCaxxLlq1iy24dy6Z7duvIkgq/kL911B7c9/hUACbPeJs2rVvSqUO7pHzXDtxy6Smcdt4tvLN4eeMFnoe6dW7PkqUry6bfX7bSXWf1pHuX9izOOLZLlq6ke4Uei+5dOvDe0qRXY/36DXz86eds0b5d2fL7Hp/K1w90K68y9Z30ImI9cAYwEZgD/DsiZqXjJI7MKHo8cEdU7IeuQlOd07sTOBY4joxWXi18kf6/JON96XQLSVsDZwMHpK2qh4A26UEcBtxF0hJ7pLqNSGoDXAMcGxEDgRuBNnUpV9ttNydTZy9g2z5d6NOzEy1bFHP0QYN5+Jnp5cosfv8jRgzdHoDt+najdauWLF+xms03bcu/Lj+V86++n/9Mn98U4eeVnXfozTuLl7FwyYesXbeeB598lQP2HNDUYW0UBu3Qh7cXLePd95Jje/8TUzlk7/LH9pC9B/DvCS8D8OCk19h7t35lX84lJSU88OQ0J70qNEBLj4iYEBHbRcS2EXFROu/ciBifUWZsRGQ90LGpRhT8iyQxdAb2rWT5J8BmOdS/OfApsErJAJlRwCRJmwKbRMQESc8DNX0Llyau5em6x5IkrYoxVlcOgDpsu9nYsKGEcy79N3dfcTrFxeK28S/x+vz3+eV/H860Oe/y8DMz+M2f7+Uvvx7D/4zZjwBOP/8WAH44egRb9+7COSeP4pyTRwFw9BlXsXzF6ibco+arRXEx5515NCedcwMlJSUcO2oY223dnctvepiB2/fmwL0GMP31dznt//7OqtVrePLFWfzl74/wyLifA3DcmVcy/92lfLrmC/b65vlc8rPjGDFshybeq+ahRYtiLj7rGMb85Fo2bCjh+COGs/02Pbj0xgnsskNvDtlnIGOOGM6Pfnsre4z+LR0234Trzv9u2fovTXuLnl07sNWWncvV+9ur7+fex15hzefrGPz1cznha3tw9g9GNfbuNa08ug2ZsmwR1v+Gk27A5RGxXzrdF3gwIgZI2oKkSdsSuITkHNmQiDhDyWCVByPirsx10joyl40D9iQZ8rqKpC94InA/SZIS8IeIuLlCXCOBsyPiiHT6QpLrP94H3gAWRMRYSccAFwNrSM4v/rqKcuOAB4Hna9p2pqJNukbr7UdXtdjq0VtP/ampQygIm7QqbuoQCkbHdi1eScc0NIoWnbeJDkdcXOv1Prx5TKPGCU2Y9Kx6TnqNx0mvcTjpNZ7GTnotO28bHb5W+6S3fNzxjZ70fMGUmZnlLJtzdM2Bk56ZmeUuP3Kek56ZmeVIbumZmVkBcdIzM7OC4aRnZmYFQXn05HQnPTMzy11+5DwnPTMzy5EHspiZWSHJl6TXbB4ia2Zm1tDc0jMzs5zlS0vPSc/MzHKXHznPSc/MzHLnlp6ZmRWEbB8K2xw46ZmZWc6c9MzMrGA46ZmZWeHIj5znpGdmZrlzS8/MzAqDb0NmZmaFQkCe5DwnPTMzy5UvWTAzswKSJznPSc/MzHKXLy09P2XBzMxyo6SlV9tXjdVKh0qaK2mepF9UUWa0pNmSZkn6Z011uqVnZmY5EVBUVL8tPUnFwNXAQcAiYLKk8RExO6NMP+CXwF4RsUJS15rqdUvPzMxy1gAtvWHAvIiYHxFrgTuAoyqU+SFwdUSsAIiIpTVV6qRnZmZNpbOkKRmvUzKWbQkszJhelM7LtB2wnaTnJb0k6dCaNujuTTMzy1kdB7Isj4ghOWy2BdAPGAn0Ap6RNDAiVla1glt6ZmaWm4YZyLIY6J0x3Sudl2kRMD4i1kXE28AbJEmwSk56ZmaWk+SOLKr1qwaTgX6StpbUCjgeGF+hzH0krTwkdSbp7pxfXaXu3jQzsxzV/x1ZImK9pDOAiUAxcFNEzJJ0ATAlIsanyw6WNBvYAPwsIj6srl4nPTMzy1lDXJseEROACRXmnZvxPoCfpK+sOOmZmVnO8uWOLE56ZmaWmyzvsNIcOOmZmVlOSgey5AMnPTMzy1me5DwnPTMzy51bemZmVjDyJOc56ZmZWY7klp7laKftenHvxMuaOoyC8IPbX23qEArC13ft1tQhWANJBrI0dRTZcdIzM7Mc1f8dWRqKk56ZmeUsT3Kek56ZmeXOLT0zMysMeXRHFj9ayMzMCoZbemZmlhPfhszMzAqKk56ZmRWMPMl5TnpmZpY7t/TMzKww5NHoTSc9MzPLiXxHFjMzKyR5kvOc9MzMLHdFeZL1nPTMzCxneZLznPTMzCw38vP0zMyskBTlR85z0jMzs9zlS0vPN5w2M7OcSbV/1VynDpU0V9I8Sb+oZPlJkpZJmpa+Tq6pTrf0zMwsJyK5Vq9e65SKgauBg4BFwGRJ4yNidoWi/4qIM7Kt1y09MzPLWZFq/6rBMGBeRMyPiLXAHcBROceZawVmZlbglNyRpbavGmwJLMyYXpTOq+gYSdMl3SWpd02VOumZmVlT6SxpSsbrlFqu/wDQNyJ2Bh4Dbq5pBZ/TMzOznNVx8ObyiBhSxbLFQGbLrVc6r0xEfJgx+Vfg0po2WGXSk3QlEFUtj4gza6rczMw2fqJBbkM2GegnaWuSZHc8cEK57Uo9ImJJOnkkMKemSqtr6U2pY6BmZlZg6jvnRcR6SWcAE4Fi4KaImCXpAmBKRIwHzpR0JLAe+Ag4qaZ6q0x6EVGub1TSJhHxWQ77YGZmG6mGuDg9IiYAEyrMOzfj/S+BX9amzhoHskjaQ9Js4PV0ehdJ19RmI2ZmtvGqy4XpTXUDl2xGb/4ZOAT4ECAiXgNGNGBMZmaWZ4qkWr+aQlajNyNiYYWm64aGCcfMzPJRftx5M7ukt1DSnkBIagn8mCxGyJiZWeHIlxtOZ5P0TgX+QnIl/HskI2lOb8igzMwsfySXLDR1FNmpMelFxHLgxEaIxczM8lF2txVrFrIZvbmNpAfSxzcslXS/pG0aIzgzM8sPG9PozX8C/wZ6AD2BO4HbGzIoMzPLLw1ww+kGkU3S2yQibomI9enrVqBNQwdmZmb5ofScXj0/WqhBVHfvzS3Stw+nT6y9g+RenMdR4Qp5MzMrbPlyTq+6gSyvkCS50j3574xlQS1v/WJmZhuv/Eh51d97c+vGDMTMzPKT1CBPWWgQWd2RRdIAoD8Z5/Ii4h8NFZSZmVlDqDHpSToPGEmS9CYAo4DnACc9MzMDmu4ShNrKZvTmscABwPsR8T1gF6B9g0ZlZmZ5JV8uWcime3NNRJRIWi9pc2Ap5R/hXiuSOgBjIuLautZhje+5ya/z++vGs2FDCUePGsbJx+1fbvmUGfO59LrxvDF/CZf+6kQO3mfnsmW7jDqHfn27A9Cja0euPP97jRp7vtm1V3t+sMdWFEk8Pncp97y2pNzy/fp15ru79+Gjz9YCMGHWBzw+dxkA/3fo9mzfdVPmfPAJF018o9FjzydzZr3NPf9+kohg+F4DOfCQ3cstf/6ZaTz39DRUJFq3bsVxJx5E9x6dy5av+OhjLrng7xx6+J7sf9DQxg6/2cmXll42SW9KmqhuJBnRuRp4MYdtXgHcKGkQ0DN9SCDp02/7R8Tvcqg7K5JaAw8BnYFLIuJfWa53KvBZxfOZkvoCD0bEgPqOtTnYsKGEi66+lxsuOYXundtz/I+uYL/hO7HtVt3KyvTo0oHf/nQ0N9/19FfWb92qJXdd+5PGDDlvFQlO2asvYye8zoefruXSr+/EywtWsmjlmnLlnp//ITe+sOAr6983fQmtWxRxyI5dGyvkvFRSUsJddzzOaWd+kw4dN+NPv7uVATtvWy6p7TZ0R/YaMQiAma/N4767JnHqj44tW37fXU+x404e7wcgmu5RQbWVzb03/yd9e52kR4DNI2J6XTYmqRtwV0Q8K+kkYAjpNX/po9/H16XeOtg13eagbFeQ1CIirmuwiJqxGXPfpU/PzvTu0QmAUSMH8dSLs8olvS27J5d1Kl/uOttM9euyKUs+/pwPPvkCgOfe+ohhW3X8StKryoz3PmanHps1ZIgbhQXvvE/nLh3p3KUDALsO2YEZr71VLum1adu67P0Xa9eVW3/6tDfZolN7WrVu2SjxNntNeFux2qru4vTB1S2LiKnVLP8dsDAirk6nx5K0EO8CLk6T5wVAW0l7A5cAbYEhEXGGpHHAGpLk1BX4PvAdYA/gPxFxUlrvtcDQdN27IuK8jO0fCawHHo2IszNi6wrcCnSRNA04BugA/AnYFFgOnBQRSyRNAqYBewO3S9oMWB0Rf5C0G3BTWu2jGfX3BW4B2qWzzoiIFyT1AP4FbE5y3E+LiGerOobNydIPP6Z7+uUA0K1ze6a//m7W669du57jzvgLLYqL+P5x+3HAnhtlg7hebNGuFctXry2b/vDTtWzXtd1Xyg3fegv6d9+c91Z9zk0vLeDDT9d+pYxVbdXKT+jY8csfBx06bsqCt5d8pdyzk15l0hNT2LChhNP/dzQAX3y+licefZn/OfObPPn45EaLubnbGC5O/2M1ywLYv5rl/yJ54vrV6fRokqevFwNExFpJ55ImOYC05ZepI0mSO5KkBbgXcDIwWdKgiJgG/DoiPpJUDDwhaWdgMfANYIeIiLRr9svAI5ZKOhk4OyKOSJ8ReAtwVEQsk3QccBFJogVoFRFD0hjHZlT1d5KE9oykyzLmLwUOiojPJfUjuU/pEOAEYGJEXJTGu0nFgybpFOAUgJ696nzatNmZeMuv6Na5PQuXfMjJP7+e7fp2p3fPzjWvaJWa8u5Knn3rQ9aXBAfv0JUfj9yGcx96vanD2ijtM3JX9hm5K6+8PIdHJ7zIiScdxiMPvcDIA3ajdZtWTR1es5LNqMjmoLqL0/era6UR8aqkrpJ6Al2AFenT1/vWopoH0qQ1A/ggImYASJoF9CVpgY1OE0ULkhti9wdmA58Df5P0IPBgDdvZHhgAPJb+UikGMn/yfeV8X5pIO0TEM+msW0gu5QBoCVyVnrPcAGyXzp8M3JQm2fvSpF1ORNwA3AAwcNDgqCHuRtO10+a8v2xl2fQHy1fRrXP2A3hLy/bu0YkhO2/DnLfec9KrwkefrqXzpl9+mXZq14oPPy3ftfbJF+vL3j8+dynf2X3j+YHUWNp32IwVKz4pm165YjXtO1TdLbzrkB248/bHAFjw9hKmTX2D8fc8w5o1X1Ak0bJlMfuMrLJzbKMnNo6WXq7uJLncoTuVJI4sfJH+vyTjfel0C0lbA2cDQyNiRdol2iYi1ksaRnKZxbHAGVTfKhUwKyL2qGL5p7WM+yzgA5JLO4pIEjBpi3AEcDgwTtKf8uUC/wHb92bB4uUsev8junXanIcnTeP3vzghq3VXffIZbVu3olWrFqxY9SnTZi3g+9+s8++pjd6by1bTY/M2dN2sNR99upa9t92Cy596q1yZjm1bsmJNkgiHbtWRRSs+b4pQ81qfrbqzfOkKPly+kvYdNuPVKa/z7e8fXq7MsqUr6NK1IwCzZ84ve3/m2WPKyjz84PO0bt2qoBNeqXw5nd+QSe9fJCM+OwP7VrL8EyCXM+6bkySkVekAmVHAJEmbkjwZYoKk54H5NdQzl+T83h4R8WLaEtsuImZVtUJErJS0UtLeEfEc5R+y2x5YlF7m8V3SLl1JW6Xzb0xHjw4mTy7wb1FczK9O/zqn/upGNpSU8I2Dh/Fffbtz1c0T2Wm7Xuy3x07MnLuQH19wM5988hlPvzSHa/7xKPfdeDZvv7uU86+4myKJkgh+cNx+5QbAWHklATe+8A7njdqeIokn5i5j4Yo1jNltS+Yt+5TJ767k8AHdGbpVBzaUBKu/2MCVT3+ZFC/62o5s2b4tbVoWc+OYXbn62flMW7SqCfeoeSouLuKY4w/guivvpqSkhN33HEiPnp2Z8MBz9OnTnQG7/BfPTnqVN15fQFFxEZts0oYTvjuq5ooLWMEnvYiYlQ78WBwRXz1DDE8Bv0gHk1xSh/pfk/Qq8DqwEHg+XbQZcL+kNiStuGrHyqfnF48FrpDUnuSY/BmoMumlvkfSXRlkDGQBrgHulvQd4BG+bCmOBH4maR3JoJ7vZLOfzcWIYTsyYtiO5ead8d1Dyt4P2L43T9z2m6+sN2invtx7/U8bPL6NydSFq5i6sPwA6dtfWVz2/tbJC7l18sJK1/31A3MaNLaNSf8B29B/QPnnYR/2tb3L3h89uroOosSoI/aq97jyUfJQ2PzIeoqo/tSRkj05EdgmIi6Q1AfoHhEvN0aAhWrgoMFx76PP11zQcvaju+t0BY7V0td3dQu/sZy6x9avlA7Aawzd+w2Ib19+d63X+8PXdmjUOCG7ATfXkIyiLO3I/oQvR2WamZmlrb3avZpCNklv94g4nS8HZKwAPFbXzMyA0ienq9avGuuVDpU0V9K89GHmVZU7RlJIqrHVmE3SW5deVxZp5V1IRlCamZkBSTKp7as6ad65mmSQYn9gjKT+lZTbDPgx8J9s46zJFcC9QFdJF5E8VujibCo3M7PC0ADdm8OAeRExPyLWAncAR1VS7rfA70l7I2uSzb03b5P0Csl1bwK+HhEeImZmZrnqLGlKxvQN6U06ALYkGZlfahFQ7lEY6e0ye0fEQ5J+ls0Gs3mIbB/gM+CBzHkRkf3NF83MbKOlLM/RVWJ5XUdvSioiuWfySbVZL5vr9B4iOZ8noA2wNckF3TvVLkQzM9tYNcBozMWUf3Zrr3Reqc1IbiE5Kb1GsDswXtKREZHZeiwnm+7NgZnTaXPyf6oobmZmBagB7sgyGeiX3nJyMXA8yY37AYiIVSR3/AIgfSrO2dUlPKjDHVkiYqqk3WsuaWZmhaD0koX6lN5H+QxgIsntHG9K7/R1ATAlfQZrrWVzTi/zNl5FJPeMfK8uGzMzs41TQ1xsHhETSB80njHv3CrKjsymzmxaepk3hV5Pco6v9vebMTOzjZM2khtOpxcHbpb55HEzM7OKRH5kvSqTnqQWaZ+qbyNuZmZVSs7pNXUU2amupfcyyfm7aZLGkzwUtuyBqhFxTwPHZmZmeWJjSHql2gAfkjx9vPR6vQCc9MzMDMif5+lVl/S6piM3Z/JlsitV/UP4zMysYGws3ZvFwKZQ6dlJJz0zM0s04fPxaqu6pLckIi5otEjMzCxv1ffF6Q2luqSXH3tgZmZNamPp3jyg0aIwM7O8licNvaqTXkR81JiBmJlZvhJFedI5mM2T083MzDYKtX7KgpmZWSaxEXRvmpmZZWVjueG0mZlZNjaGSxbMzMxq5O5NMzMrKG7pmZlZwciTnOekZ2ZmuRH5c/2bk56ZmeVGG8ejhczMzLKSHynPSc/MzHKU3HA6P9Kek56ZmeUsP1Kek56ZmdWDPGnoOemZmVmu5IEsZmZWGPLpkoV8idPMzJoxSbV+ZVHnoZLmSpon6ReVLD9V0gxJ0yQ9J6l/TXU66ZmZWc5Uh1e19UnFwNXAKKA/MKaSpPbPiBgYEYOAS4E/1RSnk56ZmeVGDdLSGwbMi4j5EbEWuAM4KrNARHycMdkOiJoq9Tm9ZqpVcRG9tmjb1GEUhD8dNaCpQygIg795cVOHYM1PZ0lTMqZviIgb0vdbAgszli0Cdq9YgaTTgZ8ArYD9a9qgk56ZmeUkh4EsyyNiSC7bjoirgaslnQD8BvhudeWd9MzMLGcNcMnCYqB3xnSvdF5V7gCuralSn9MzM7Oc1fdAFmAy0E/S1pJaAccD48ttU+qXMXk48GZNlbqlZ2ZmOavvhl5ErJd0BjARKAZuiohZki4ApkTEeOAMSQcC64AV1NC1CU56ZmaWo+ScXv3fkSUiJgATKsw7N+P9j2tbp5OemZnlLE/uQuakZ2ZmuRLKk+csOOmZmVnO3NIzM7OC0FDn9BqCk56ZmeVGbumZmVkBcdIzM7OC4YEsZmZWEAQU5UfOc9IzM7PcuaVnZmYFw+f0zMysYORLS89PWTAzs4Lhlp6ZmeXEA1nMzKyA+N6bZmZWKHxHFjMzKyR5kvOc9MzMLDfJOb38SHtOemZmlrP8SHlOemZmVh/yJOs56ZmZWc48etPMzApGnpzSc9IzM7Pc5UnOc9IzM7N6kCdZz0nPzMxyInxOz8zMCoXvyGJmZoUkT3KeHy1kZmb1QHV41VSldKikuZLmSfpFJct/Imm2pOmSnpC0VU11OumZmVmOVKf/qq1RKgauBkYB/YExkvpXKPYqMCQidgbuAi6tKVInPTMza46GAfMiYn5ErAXuAI7KLBART0XEZ+nkS0Cvmir1OT0zM8tZHQeydJY0JWP6hoi4IX2/JbAwY9kiYPdq6voB8HBNG3TSMzOznGR5iq4yyyNiSM7bl74FDAH2ramsk56ZmeWu/odvLgZ6Z0z3SueV36x0IPBrYN+I+KKmSn1Oz8zMclbfA1mAyUA/SVtLagUcD4wvt01pV+B64MiIWJpNnG7pmZlZzur74vSIWC/pDGAiUAzcFBGzJF0ATImI8cBlwKbAnUoCeDcijqyuXic9MzPLWUNcnB4RE4AJFeadm/H+wNrW6aSXktQSOB24MiI2NHU8zc3jL8zml3+8iw0lJXz7qD0566SDyy3/Yu06TjvvFqa9/i5btG/HTRd/nz49O5UtX/j+R+wx+kJ+/sPD+NG3a/05LSjPT5nLH24Yz4aS4BsHD+V7o/crt/yVmfP54w0P8Obb73PJz8dw4N47AzD5tbf4440PlJV7Z9EyLvn5Cey3x06NGn++OGBYPy750eEUFxVxy0NT+PM/nym3/KLTD2OfXbcBoG2blnTp0I6+R1wIwNj/PoSDh28PwGX/eIp7n5rRuME3NzmMZGlsTnpfugB4Degtac+I+CeApCHAdyLizCaNrglt2FDCzy79N/dedQY9u3Vg/+9exqgRA9lhmx5lZW65/0Xab96WqfeO5e5HpzD2yvu56ZLvly3/zeX3cOCe/vKtyYYNJfz+2vu45sKT6da5Pd866yr2Hd6fbfp0KyvTo0sHxp41mlvuKf8lPXSXbbnjqv8FYNUnn3HUyZcyfNd+jRl+3igqEpf979f4xk//znvLPubJ60/j4efnMHfBsrIyv776ywbGD48ezs79egJw8PDt2Xm7nuxz8lW0blnMA385mcf/8waffFbjGIqNWr7ccNoDWQBJbYGZEXEH0Bc4oXRZREwp5IQH8Mqsd9imd2f69upMq5YtOPqgwUx4enq5Mg8/M50xhyeX0By1/648PXkuEQHAQ5Neo0/PTuywTfdGjz3fzHxjIb16dqJXj060bNmCQ0bswqSXZpcr07PbFmy3dQ+KqjmJ8vhzM9hryPa0bdOqoUPOS7vt2Iv5iz9iwZIVrFu/gXuenM5he+9YZfljD9iZu594DYDt+3bhhdfeYcOGEj77fB2z3nqfA3Yv7B8XIjmnV9tXU2hWSU9SO0kPSXpN0kxJx6Xzd5P0tKRXJE2U1COdP0nS5ZKmSJojaaikeyS9KenCSuo/VdJlGdMnSboqItaQjAAC+B2wj6Rpks6SNFLSg2n5sZJulvSspAWSjpZ0qaQZkh5Ju0iRdK6kyek+3KD0DKukMzPuE3dHgx7MerRk2Sq27NaxbLpnt44sWbaqXJn3ln5ZpkWLYjbftC0frfqU1Z99wV/+8Rg//+FhjRpzvlr24Sq6d+5QNt21c3uWfriq6hWqMPGZ1zhk30H1F9hGpkfnzVm89Mvj+t6yj+nRuX2lZXt360CfHlvwzNT5AMyc9z4HDutH29Yt2aL9Juyz6zZs2aXydQtJA9x6s0E0q6QHHAq8FxG7RMQAoDSRXAkcGxG7ATcBF2Wssza9uPE64H6S83IDgJMkdSpfPXcD38iYPo7k1jaZfgE8GxGDIuLySmLcFtgfOBK4FXgqIgYCa4DD0zJXRcTQdB/aAkdk1L1rep+4U7M4Hnnv9zc8xGlj9mfTTVo3dSgFY9lHHzPvnffZY/B2TR3KRuHo/Qcy/umZlJQkPRdPTZnHYy+9wcSr/5u/nnsck2e9y4Z0WUHLk6zX3M7pzQD+KOn3wIMR8aykASRJ7LG0wVQMLMlYZ3zGurMiYgmApPkkFzZ+WFowIpZJmi9pOPAmsAPwfC1jfDgi1kmakcbySMb2+6bv95N0DrAJsAUwC3gAmA7cJuk+4L6KFUs6BTgFoHefPrUMq+H06NKexR+sKJt+74MV9Kjwy7Zn16TMlt06sn79Bj5evYYt2rdjyqwF3P/kNM678j5WfbKGoiLRunVLThld440TClKXTu15f/nKsumly1fRtVPtWhGPPTud/fbYiZYtius5uo3HkuUfs2XXL49rzy6bs2R55S3qow/YmZ9d/kC5eX+8dRJ/vHUSADf+32jeWri8wWLNFz6nVwcR8QYwmCSBXCjpXJLfA7PSltegiBgYEZlDB0vPHpdkvC+driyp3wGMBo4B7o3SE0/Z+yKNtQRYl7F+CdBCUhvgGpKW6UDgRqBNWuZwkruGDwYmSyoXX0TcEBFDImJIl85dahlWwxncfyveencZCxYvZ+269dzz2FRGjdi5XJlD9xnI7Q/9B4D7n3yVEUO3QxIP33gW08dfwPTxF3DamJH85KSDnfCqsdN2vVi4+EMWv/8R69atZ+Izr7Hv7lWfa6rMI09P41B3bVZr6uuL2bZXJ/p070jLFsUcvf/OPPz8618p169PZzps2paXZ71bNq+oSHTcvC0AO23TjZ226c6TU+Y1WuzNVb6c02tWLT1JPYGPIuJWSSuBk0nOsXWRtEdEvJh2d24XEbPquJl7SW5Zsyvw80qWfwJsVse64csEt1zSpsCxwF2SioDeEfGUpOdI7i6wKbAyh201ihYtirn0nNEcc+bVbNgQnHjkcHbctgcXX/cgg3bsw2H77sy3j9qTU8/7B4O/MZaOm7fjbxd9r6nDzkstiov5+WlHcfr//Y2SkhKOPGgo227VnWtveZT+/Xqx7/D+zHpjIT+98B98vHoNz7w8h+tue4y7rv0pAO998BEfLF/FbgO3buI9ad42bCjhnD8/wN1/OIniInHbhKm8/s5Sfvn9A5j2+mIefiFJgEfvvzP3PFl+0FbLFsVMuPIUAD759HNOuehONmwoafR9aG7yo50Hqn1Dp+FIOoTkCvsSYB1wWkRMkTQIuAJoT5Ko/xwRN0qaBJydlhmZvj8iratsWSXbeRDoHxHbZMxbHRGbpkl1ItAJGEfyvKazI+IISWOB1RHxh8x10vdly9JBNGOA94E3gAUk5yGfSvdBwK0R8buqjsVuuw2J5//zldCtAby99NOmDqEgDP7mxU0dQsH4/IWLX6mPGzlna8Aug+OeR5+r9Xrbd2/XqHFCM2vpRcREkoRTcf40YEQl80dmvJ8ETKpsWSXrHVHJvE3T/68jGaiSaVK6bGxl61RcFhG/AX5Tyab3riomM7N8lYxLyY+2XrM6p2dmZtaQmlVLz8zM8lATDkypLSc9MzPLWZ7kPCc9MzOrB3mS9Zz0zMwsR1k9FLZZcNIzM7Oc+ZyemZkVhDx6nJ6TnpmZ1YM8yXpOemZmljOf0zMzs4Lhc3pmZlYw8iTnOemZmVmOfEcWMzMrLPmR9Zz0zMwsJ8ItPTMzKyB5kvOc9MzMLHf50tLz8/TMzCxnqsN/NdYpHSpprqR5kn5RyfIRkqZKWi/p2GzidNIzM7NmR1IxcDUwCugPjJHUv0Kxd4GTgH9mW6+7N83MLHf13705DJgXEfMBJN0BHAXMLi0QEe+ky0qyrdQtPTMzy5nq8AI6S5qS8Tolo8otgYUZ04vSeTlxS8/MzHKiul+cvjwihtRzONVy0jMzs5w1wA2nFwO9M6Z7pfNy4u5NMzPLXR37N6sxGegnaWtJrYDjgfG5humkZ2ZmOavvnBcR64EzgInAHODfETFL0gWSjgSQNFTSIuCbwPWSZtUUp7s3zcwsZw1xcXpETAAmVJh3bsb7ySTdnllz0jMzsxxld7F5c+CkZ2ZmOcmnG077nJ6ZmRUMt/TMzCxn+dLSc9IzM7Oc+ZyemZkVhrrfkaXROemZmVlOsrvWvHlw0jMzs9zlSdZz0jMzs5z5nJ6ZmRWMfDmn5+v0zMysYLilZ2ZmOcuThp6TnpmZ1YM8yXpOemZmljMPZDEzs4KQTzecVkQ0dQxWCUnLgAVNHUcddAaWN3UQBcDHufHk47HeKiK6NNbGJD1Ccpxqa3lEHFrf8VTHSc/qlaQpETGkqePY2Pk4Nx4f642LL1kwM7OC4aRnZmYFw0nP6tsNTR1AgfBxbjw+1hsRn9MzM7OC4ZaemZkVDCe9AiCpg6TTmjqOjZGPbWGQ1FLS/0oqbupYLDdOeoXhCmCmpEGSDiudKelISb9ojAAktZb0uKRpko6rptxJknpmUd8Fkg6s3yjrJG+ObSXrnSrpO5XM7ytpZv1GmfcuAN4Heks6oXSmpCGSrmi6sKy2fE5vIyepG7B7RIyXdBIwJCLOaII4hgMXRkS1iUrSJODsiJjSKIHlIN+ObYV1WkTE+iqW9QUejIgB9RRiXpPUFjg6Im6TNJLk83lE00ZldRYRfm0EL+B3wOkZ02OBs4G+wEygFfAusAyYBhwHnARclZYfB1wLvATMB0YCNwFzgHEZ9V4LTAFmAedX2P5sYDrwhwqxdQXmAavSbW8LnAtMTmO7geRORscCq4G5abm2lZXLiPfYmrZdoMd2N+Bp4BVgItAjLTsJ+HO6jZ+W7ke6bDfgtfR1GTAznd8XeBaYmr72TOf3AJ5JtzkT2KcRPuPtgIfSGGcCx2XEXtX+Xp7u7xxgKHAP8CbJj4SK9Z8KXJYxnfk3XJ3+/6WMY31W+rd8MONzcXN6vBYARwOXAjOAR4CWabmqPtNnZvyd72jq75SN9dXkAfhVT39I2BV4OmN6NtA7/dIq/QIr+0dccZrki/kOkuRzFPAxMJCkC/wVYFBabov0/8Xpl8rOQCeSRFX6j7dDJfGVfTlk1pO+vwX4Wvp+EkmLqaZy40iSZI3bLqRjC7QEXgC6pNPHATdlHNtrMtYby5dJbzowIn2fmfQ2Adqk7/sBU9L3PwV+nRHvZo3wGT8GuDFjun0W+/v79P2PgfdIknVrYBHQqUL9XYB5GdMPA3un70uTXtmxruTYjwWeS2PaBfgMGJUuuxf4eg2f6feA1g31OfYrefmc3kYiIl4FukrqKWkXYEVELKxlNQ9E8i9uBvBBRMyIiBKSlkfftMxoSVOBV4GdgP4kv3w/B/4m6WiSf+w12U/SfyTNAPZP66pLubpsu1by7NhuDwwAHpM0DfgN0Ctj+b8qriCpA8mX7DPprFsyFrcEbkyP/51pTJC0VL4naSwwMCI+qSGu+jADOEjS7yXtExGrqHl/x2esOysilkTEFyQt7t6ZlUfEMmC+pOGSOgE7AM/XMsaHI2Jdur1ikhZe6fb7pu+r+kxPB26T9C2g0q5ny52fsrBxuZOk9dOdSr7csvBF+v+SjPel0y0kbU3SrTc0IlZIGkfSClgvaRhwQLr9M0j+MVdKUhvgGpIW3cL0i7NNXcrVdts5yItjS9KanBURe1Sx/NNaxn0W8AFJy6WIJAETEc9IGgEcDoyT9KeI+Ect666ViHhD0mDgMOBCSU+QtKCq299qj3sl5e8ARgOvA/emP1Rq44s01hJJ6zLWL/07V/eZPhwYAXwN+LWkgVHFeVerO7f0Ni7/Ao4n+XK8s5LlnwCb5VD/5iRfmqvSQRyjACRtCrSPiAkkX5K71FBP6T/y5em6x1YRY3XlqOO26ypfju1coIukPdL1W0qqqhUNQESsBFZK2juddWLG4vbAkrRV+m2S1guStiJpsd4I/BUYnO2O1lU6qveziLiVpAt2MHXY3xrcS9IFPYYkAVaU69+50s+0pCKgd0Q8Bfyc5LhvmsN2rApu6W1EImKWpM2AxRGxpJIiTwG/SLuBLqlD/a9JepXkV/BCvuz62Qy4P/0VK+AnNdSzUtKNJCfy3yfpKis1DrhO0hpgD6CqcqVqte26yqNju1bSscAVktqT/Bv/M0k3anW+B9wkKYBHM+ZfA9ydXtrwCF+2FEcCP5O0jmTw0VcufWgAA4HLJJUA64DTctjfSqWt7DlA/4h4uZIi04ENkl4j+ay+Wsv6q/rsFwO3pvsg4Ir0x4jVM1+yYGZmBcPdm2ZmVjCc9MzMrGA46ZmZWcFw0jMzs4LhpGdmZgXDSc8sC5I2pE8xmCnpTkmb5FDXuHSYPZL+Kql/NWVHStqzDtt4R1LnbOdXKLO6ltsaK+ns2sZo1hSc9MyysyYiBkXy5IG1JDcnLiOpTte8RsTJETG7miIjgVonPTOrnJOeWe09C/xX2gp7VtJ4YLakYkmXSZosabqk/wZQ4ipJcyU9TvJkBNJlkyQNSd8fKmmqpNckPaHkET+nAmelrcx9JHWRdHe6jcmS9krX7STpUUmzJP2V5ALnakm6T9Ir6TqnVFh2eTr/CUld0nnbSnokXedZSTvUy9E0a0S+I4tZLaQtulF8eSPhwcCAiHg7TRyrImKopNbA85IeJXlKw/YkN2vuRvKUhpsq1NuF5O4zI9K6toiIjyRdR3KH/z+k5f4JXB4Rz0nqQ/IonR2B84DnIuICSYcDP8hid76fbqMtMFnS3RHxIckjfKZExFmSzk3rPoPkMTinRsSbknYnuVtLQ9zn1KzBOOmZZadteosxSFp6fyPpdnw5It5O5x8M7Fx6vo7k/on9SG4ifHtEbADek/RkJfUPB54prSsiPqoijgOB/lJZQ27z9B6OI0ie30ZEPCRpRRb7dKakb6Tve6exfkhyc+TSm2rfCtyTbmNP4M6MbbfOYhtmzYqTnll21kTEoMwZ6Zd/5lMLBPwoIiZWKHdYPcZRBAyPiM8riSVrSp4AfiCwR0R8puSJ9V950kUq0u2urHgMzPKNz+mZ1Z+JwGmSWgJI2k5SO5InjB+XnvPrAexXybovASOUPGIISVuk8yve1f9R4EelE5IGpW+fAU5I540COtYQa3uS5wJ+lp6bG56xrIgvn2hxAkm36cfA25K+mW5DSp4taJZXnPTM6s9fSc7XTZU0E7iepDflXuDNdNk/gBcrrpg+wPQUkq7E1/iye/EB4BulA1mAM4Eh6UCZ2Xw5ivR8kqQ5i6Sb890aYn2E5Pluc4DfkSTdUp8Cw9J92B+4IJ1/IvCDNL5ZJI/gMcsrfsqCmZkVDLf0zMysYDjpmZlZwXDSMzOzguGkZ2ZmBcNJz8zMCoaTnpmZFQwnPTMzKxj/Dy+T8H0CzwnLAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 720x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "labels = ['vítimas fatais', 'vítimas feridas', 'sem vítimas']\n",
    "np.set_printoptions(precision=2)\n",
    "plt.rcParams['figure.figsize'] = [10, 5]\n",
    "disp = plot_confusion_matrix(clf, X_test, y_test,\n",
    "                                 display_labels=labels,\n",
    "                                 cmap=plt.cm.Blues,\n",
    "                                 normalize='true'\n",
    "                                 )\n",
    "disp.ax_.set_title('Matriz de confusão normalizada', fontsize=14, y=1.05, loc='left', x=-0.01)\n",
    "plt.grid(False)\n",
    "plt.savefig('matriz_confusao_acidentes_minas_gerais.png')\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "collapsed_sections": [],
   "name": "regra_associação.ipynb",
   "provenance": []
  },
  "file_extension": ".py",
  "hide_input": false,
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.9"
  },
  "mimetype": "text/x-python",
  "name": "python",
  "nbTranslate": {
   "displayLangs": [
    "*"
   ],
   "hotkey": "alt-t",
   "langInMainMenu": true,
   "sourceLang": "en",
   "targetLang": "fr",
   "useGoogleTranslate": true
  },
  "npconvert_exporter": "python",
  "pygments_lexer": "ipython3",
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  },
  "version": 3
 },
 "nbformat": 4,
 "nbformat_minor": 1
}