x-002-amostras.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# O objetivo deste exercício é testar as técnicas de subamostragem e sobreamostragem utilizando outra base de dados\n",
    "\n",
    "# Carregue a base de dados csv_result-ebay_confianca_completo.csv usado para prever confiança de usuários baseado em traços de personalidade extraídos de textos\n",
    "\n",
    "# A classe é o atributo 'reputation', que pode ser reputação 'boa' ou reputação 'ruim'\n",
    "\n",
    "# Utilize o algoritmo Random Forest e faça os três testes. O algoritmo Naïve Bayes não terá um bom desempenho nesta base de dados, por isso precisamos utilizar o Random Forest que é um algoritmo baseado em árvores de decisão."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Importação das bibliotecas para analise e manipluação de dados\n",
    "\n",
    "import pandas as pd                                                # Analise e tratamento de dados\n",
    "import numpy as np                                                 # Tratamento de dados numéricos\n",
    "import random                                                      # Gerador de números aleatórios\n",
    "import matplotlib.pyplot as plt                                    # Plotagem de gráficos\n",
    "import seaborn as sns                                              # Plotagem de gráficos\n",
    "from sklearn.model_selection import train_test_split               # Divisão de conjuntos de dados de treino e teste\n",
    "from sklearn.ensemble import RandomForestClassifier                # Algoritmo Random Forest(Árvore de decisão)\n",
    "from sklearn.metrics import confusion_matrix, accuracy_score       # Matriz de confusão e métricas de precisão e acurácia\n",
    "from imblearn.under_sampling import TomekLinks                     # Subamostragem aleatória (undersampling)\n",
    "from imblearn.over_sampling import SMOTE                           # Sobreamostragem aleatória (OverSampling)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Carregamento da base de dados em uma variável\n",
    "\n",
    "dados = pd.read_csv('../data/csv_result-ebay_confianca_completo.csv')\n",
    "# 'csv_result-ebay_confianca_completo.csv' é uma base de dados usada para prever confiança de usuários baseado em traços de personalidade extraídos de textos"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(5806, 75)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Visualização da quantidade de registros\n",
    "\n",
    "dados.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>feedbacks</th>\n",
       "      <th>reviews</th>\n",
       "      <th>blacklist</th>\n",
       "      <th>mextraversion</th>\n",
       "      <th>cextraversion</th>\n",
       "      <th>sextraversion</th>\n",
       "      <th>mneuroticism</th>\n",
       "      <th>cneuroticism</th>\n",
       "      <th>sneuroticism</th>\n",
       "      <th>mconscientiousness</th>\n",
       "      <th>...</th>\n",
       "      <th>need_practicaly</th>\n",
       "      <th>need_selfexpression</th>\n",
       "      <th>need_stability</th>\n",
       "      <th>need_structure</th>\n",
       "      <th>value_conservation</th>\n",
       "      <th>value_openess</th>\n",
       "      <th>value_hedonism</th>\n",
       "      <th>value_selfenhancement</th>\n",
       "      <th>value_selftranscendence</th>\n",
       "      <th>reputation</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>49</td>\n",
       "      <td>N</td>\n",
       "      <td>4.181642</td>\n",
       "      <td>0.6</td>\n",
       "      <td>1</td>\n",
       "      <td>2.777591</td>\n",
       "      <td>0.6</td>\n",
       "      <td>0</td>\n",
       "      <td>4.085460</td>\n",
       "      <td>...</td>\n",
       "      <td>0.696359</td>\n",
       "      <td>0.698786</td>\n",
       "      <td>0.756963</td>\n",
       "      <td>0.660119</td>\n",
       "      <td>0.619416</td>\n",
       "      <td>0.746372</td>\n",
       "      <td>0.640073</td>\n",
       "      <td>0.598037</td>\n",
       "      <td>0.828716</td>\n",
       "      <td>Bom</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>56</td>\n",
       "      <td>N</td>\n",
       "      <td>4.007042</td>\n",
       "      <td>0.6</td>\n",
       "      <td>0</td>\n",
       "      <td>2.698650</td>\n",
       "      <td>0.6</td>\n",
       "      <td>0</td>\n",
       "      <td>4.187338</td>\n",
       "      <td>...</td>\n",
       "      <td>0.715300</td>\n",
       "      <td>0.664572</td>\n",
       "      <td>0.728806</td>\n",
       "      <td>0.660740</td>\n",
       "      <td>0.588969</td>\n",
       "      <td>0.735915</td>\n",
       "      <td>0.644465</td>\n",
       "      <td>0.603042</td>\n",
       "      <td>0.809379</td>\n",
       "      <td>Bom</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>50</td>\n",
       "      <td>N</td>\n",
       "      <td>4.538230</td>\n",
       "      <td>0.7</td>\n",
       "      <td>1</td>\n",
       "      <td>2.298492</td>\n",
       "      <td>0.5</td>\n",
       "      <td>1</td>\n",
       "      <td>5.085833</td>\n",
       "      <td>...</td>\n",
       "      <td>0.720150</td>\n",
       "      <td>0.694678</td>\n",
       "      <td>0.669652</td>\n",
       "      <td>0.627962</td>\n",
       "      <td>0.553523</td>\n",
       "      <td>0.766618</td>\n",
       "      <td>0.655470</td>\n",
       "      <td>0.645042</td>\n",
       "      <td>0.826039</td>\n",
       "      <td>Bom</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>72</td>\n",
       "      <td>0</td>\n",
       "      <td>N</td>\n",
       "      <td>4.692854</td>\n",
       "      <td>0.3</td>\n",
       "      <td>0</td>\n",
       "      <td>2.987231</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0</td>\n",
       "      <td>4.831320</td>\n",
       "      <td>...</td>\n",
       "      <td>0.739793</td>\n",
       "      <td>0.637027</td>\n",
       "      <td>0.697221</td>\n",
       "      <td>0.638587</td>\n",
       "      <td>0.675289</td>\n",
       "      <td>0.752234</td>\n",
       "      <td>0.679661</td>\n",
       "      <td>0.674438</td>\n",
       "      <td>0.813391</td>\n",
       "      <td>Bom</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>76</td>\n",
       "      <td>0</td>\n",
       "      <td>N</td>\n",
       "      <td>4.966753</td>\n",
       "      <td>0.3</td>\n",
       "      <td>0</td>\n",
       "      <td>3.048730</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0</td>\n",
       "      <td>4.725294</td>\n",
       "      <td>...</td>\n",
       "      <td>0.718530</td>\n",
       "      <td>0.616852</td>\n",
       "      <td>0.692761</td>\n",
       "      <td>0.646695</td>\n",
       "      <td>0.677245</td>\n",
       "      <td>0.699785</td>\n",
       "      <td>0.648607</td>\n",
       "      <td>0.616075</td>\n",
       "      <td>0.816841</td>\n",
       "      <td>Bom</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 75 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   feedbacks  reviews blacklist  mextraversion  cextraversion  sextraversion  \\\n",
       "2          0       49         N       4.181642            0.6              1   \n",
       "3          0       56         N       4.007042            0.6              0   \n",
       "4          0       50         N       4.538230            0.7              1   \n",
       "5         72        0         N       4.692854            0.3              0   \n",
       "6         76        0         N       4.966753            0.3              0   \n",
       "\n",
       "   mneuroticism  cneuroticism  sneuroticism  mconscientiousness  ...  \\\n",
       "2      2.777591           0.6             0            4.085460  ...   \n",
       "3      2.698650           0.6             0            4.187338  ...   \n",
       "4      2.298492           0.5             1            5.085833  ...   \n",
       "5      2.987231           0.5             0            4.831320  ...   \n",
       "6      3.048730           0.5             0            4.725294  ...   \n",
       "\n",
       "   need_practicaly  need_selfexpression  need_stability  need_structure  \\\n",
       "2         0.696359             0.698786        0.756963        0.660119   \n",
       "3         0.715300             0.664572        0.728806        0.660740   \n",
       "4         0.720150             0.694678        0.669652        0.627962   \n",
       "5         0.739793             0.637027        0.697221        0.638587   \n",
       "6         0.718530             0.616852        0.692761        0.646695   \n",
       "\n",
       "   value_conservation  value_openess  value_hedonism  value_selfenhancement  \\\n",
       "2            0.619416       0.746372        0.640073               0.598037   \n",
       "3            0.588969       0.735915        0.644465               0.603042   \n",
       "4            0.553523       0.766618        0.655470               0.645042   \n",
       "5            0.675289       0.752234        0.679661               0.674438   \n",
       "6            0.677245       0.699785        0.648607               0.616075   \n",
       "\n",
       "   value_selftranscendence  reputation  \n",
       "2                 0.828716         Bom  \n",
       "3                 0.809379         Bom  \n",
       "4                 0.826039         Bom  \n",
       "5                 0.813391         Bom  \n",
       "6                 0.816841         Bom  \n",
       "\n",
       "[5 rows x 75 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Visualização dos primeiros registros do dataframe\n",
    "\n",
    "dados.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>feedbacks</th>\n",
       "      <th>reviews</th>\n",
       "      <th>blacklist</th>\n",
       "      <th>mextraversion</th>\n",
       "      <th>cextraversion</th>\n",
       "      <th>sextraversion</th>\n",
       "      <th>mneuroticism</th>\n",
       "      <th>cneuroticism</th>\n",
       "      <th>sneuroticism</th>\n",
       "      <th>mconscientiousness</th>\n",
       "      <th>...</th>\n",
       "      <th>need_practicaly</th>\n",
       "      <th>need_selfexpression</th>\n",
       "      <th>need_stability</th>\n",
       "      <th>need_structure</th>\n",
       "      <th>value_conservation</th>\n",
       "      <th>value_openess</th>\n",
       "      <th>value_hedonism</th>\n",
       "      <th>value_selfenhancement</th>\n",
       "      <th>value_selftranscendence</th>\n",
       "      <th>reputation</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>5803</th>\n",
       "      <td>49</td>\n",
       "      <td>0</td>\n",
       "      <td>N</td>\n",
       "      <td>6.069133</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0</td>\n",
       "      <td>1.931712</td>\n",
       "      <td>0.5</td>\n",
       "      <td>1</td>\n",
       "      <td>5.616294</td>\n",
       "      <td>...</td>\n",
       "      <td>0.669362</td>\n",
       "      <td>0.602225</td>\n",
       "      <td>0.711554</td>\n",
       "      <td>0.644777</td>\n",
       "      <td>0.666970</td>\n",
       "      <td>0.742045</td>\n",
       "      <td>0.605412</td>\n",
       "      <td>0.650072</td>\n",
       "      <td>0.828953</td>\n",
       "      <td>Bom</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5804</th>\n",
       "      <td>60</td>\n",
       "      <td>0</td>\n",
       "      <td>N</td>\n",
       "      <td>5.056324</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0</td>\n",
       "      <td>3.152737</td>\n",
       "      <td>0.5</td>\n",
       "      <td>1</td>\n",
       "      <td>5.072437</td>\n",
       "      <td>...</td>\n",
       "      <td>0.698962</td>\n",
       "      <td>0.597975</td>\n",
       "      <td>0.690498</td>\n",
       "      <td>0.614929</td>\n",
       "      <td>0.614349</td>\n",
       "      <td>0.747452</td>\n",
       "      <td>0.640656</td>\n",
       "      <td>0.654348</td>\n",
       "      <td>0.801171</td>\n",
       "      <td>Ruim</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5805</th>\n",
       "      <td>78</td>\n",
       "      <td>0</td>\n",
       "      <td>N</td>\n",
       "      <td>5.095499</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0</td>\n",
       "      <td>3.535562</td>\n",
       "      <td>0.5</td>\n",
       "      <td>1</td>\n",
       "      <td>4.692843</td>\n",
       "      <td>...</td>\n",
       "      <td>0.710999</td>\n",
       "      <td>0.692540</td>\n",
       "      <td>0.728816</td>\n",
       "      <td>0.622679</td>\n",
       "      <td>0.708117</td>\n",
       "      <td>0.795567</td>\n",
       "      <td>0.697466</td>\n",
       "      <td>0.673904</td>\n",
       "      <td>0.832073</td>\n",
       "      <td>Bom</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5806</th>\n",
       "      <td>95</td>\n",
       "      <td>0</td>\n",
       "      <td>N</td>\n",
       "      <td>4.780761</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0</td>\n",
       "      <td>3.198620</td>\n",
       "      <td>0.5</td>\n",
       "      <td>1</td>\n",
       "      <td>4.563306</td>\n",
       "      <td>...</td>\n",
       "      <td>0.724970</td>\n",
       "      <td>0.718304</td>\n",
       "      <td>0.722504</td>\n",
       "      <td>0.573924</td>\n",
       "      <td>0.668312</td>\n",
       "      <td>0.805183</td>\n",
       "      <td>0.770677</td>\n",
       "      <td>0.710048</td>\n",
       "      <td>0.819924</td>\n",
       "      <td>Ruim</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5807</th>\n",
       "      <td>53</td>\n",
       "      <td>0</td>\n",
       "      <td>N</td>\n",
       "      <td>4.525406</td>\n",
       "      <td>0.3</td>\n",
       "      <td>0</td>\n",
       "      <td>3.236950</td>\n",
       "      <td>0.5</td>\n",
       "      <td>1</td>\n",
       "      <td>4.488098</td>\n",
       "      <td>...</td>\n",
       "      <td>0.721490</td>\n",
       "      <td>0.601475</td>\n",
       "      <td>0.696407</td>\n",
       "      <td>0.642956</td>\n",
       "      <td>0.656245</td>\n",
       "      <td>0.712400</td>\n",
       "      <td>0.623931</td>\n",
       "      <td>0.633121</td>\n",
       "      <td>0.792092</td>\n",
       "      <td>Ruim</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 75 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      feedbacks  reviews blacklist  mextraversion  cextraversion  \\\n",
       "5803         49        0         N       6.069133            0.5   \n",
       "5804         60        0         N       5.056324            0.5   \n",
       "5805         78        0         N       5.095499            0.4   \n",
       "5806         95        0         N       4.780761            0.4   \n",
       "5807         53        0         N       4.525406            0.3   \n",
       "\n",
       "      sextraversion  mneuroticism  cneuroticism  sneuroticism  \\\n",
       "5803              0      1.931712           0.5             1   \n",
       "5804              0      3.152737           0.5             1   \n",
       "5805              0      3.535562           0.5             1   \n",
       "5806              0      3.198620           0.5             1   \n",
       "5807              0      3.236950           0.5             1   \n",
       "\n",
       "      mconscientiousness  ...  need_practicaly  need_selfexpression  \\\n",
       "5803            5.616294  ...         0.669362             0.602225   \n",
       "5804            5.072437  ...         0.698962             0.597975   \n",
       "5805            4.692843  ...         0.710999             0.692540   \n",
       "5806            4.563306  ...         0.724970             0.718304   \n",
       "5807            4.488098  ...         0.721490             0.601475   \n",
       "\n",
       "      need_stability  need_structure  value_conservation  value_openess  \\\n",
       "5803        0.711554        0.644777            0.666970       0.742045   \n",
       "5804        0.690498        0.614929            0.614349       0.747452   \n",
       "5805        0.728816        0.622679            0.708117       0.795567   \n",
       "5806        0.722504        0.573924            0.668312       0.805183   \n",
       "5807        0.696407        0.642956            0.656245       0.712400   \n",
       "\n",
       "      value_hedonism  value_selfenhancement  value_selftranscendence  \\\n",
       "5803        0.605412               0.650072                 0.828953   \n",
       "5804        0.640656               0.654348                 0.801171   \n",
       "5805        0.697466               0.673904                 0.832073   \n",
       "5806        0.770677               0.710048                 0.819924   \n",
       "5807        0.623931               0.633121                 0.792092   \n",
       "\n",
       "      reputation  \n",
       "5803         Bom  \n",
       "5804        Ruim  \n",
       "5805         Bom  \n",
       "5806        Ruim  \n",
       "5807        Ruim  \n",
       "\n",
       "[5 rows x 75 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Visualização dos últimos registros do dataframe\n",
    "\n",
    "dados.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dados em % :\n",
      " Bom     0.740441\n",
      "Ruim    0.259559\n",
      "Name: reputation, dtype: float64\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Visualização da variável 'reputation' graficamente e em percentual\n",
    "\n",
    "print('Dados em % :\\n', dados['reputation'].value_counts('1'))\n",
    "dados['reputation'].hist();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dados em % :\n",
      " N    0.864967\n",
      "B    0.135033\n",
      "Name: blacklist, dtype: float64\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Visualização da variável 'blacklist' graficamente e em percentual\n",
    "\n",
    "print('Dados em % :\\n', dados['blacklist'].value_counts('1'))\n",
    "dados['blacklist'].hist();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# auteração da variavel 'blacklist' para booleano\n",
    "\n",
    "dados['blacklist'] = dados['blacklist'] == 'N'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# outra opção é retirar a variável 'blacklist' do dataframe\n",
    "\n",
    "# dados.drop('blacklist', axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(5806, 74)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# definição das variaveis para classificação\n",
    "\n",
    "X = dados.iloc[ : , 0:74].values             # variaveis de entrada\n",
    "X.shape                                      # dimensões da variavel X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0, 49, True, ..., 0.6400725899947721, 0.598037041554872,\n",
       "        0.828716381071851],\n",
       "       [0, 56, True, ..., 0.6444645861342442, 0.603041772443901,\n",
       "        0.8093789692710098],\n",
       "       [0, 50, True, ..., 0.6554697130298307, 0.6450420335022421,\n",
       "        0.82603890125472],\n",
       "       ...,\n",
       "       [78, 0, True, ..., 0.6974661634054785, 0.6739035045452172,\n",
       "        0.8320726536961992],\n",
       "       [95, 0, True, ..., 0.7706773179699052, 0.7100477615788761,\n",
       "        0.8199238314070743],\n",
       "       [53, 0, True, ..., 0.6239306346573744, 0.6331205213003834,\n",
       "        0.7920921058167003]], dtype=object)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Visualização da variavel X em formato de array\n",
    "\n",
    "X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(5806,)"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# definição da variavel y de saída\n",
    "\n",
    "y = dados.iloc[ : , 74].values                    # variavel de saida\n",
    "y.shape   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['Bom', 'Bom', 'Bom', ..., 'Bom', 'Ruim', 'Ruim'], dtype=object)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Visualização da variavel y em formato de array\n",
    "\n",
    "y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "# divisão dos dados em conjunto de treino e teste usando a amostragem estratificada\n",
    "\n",
    "X_treino, X_teste, y_treino, y_teste = train_test_split(X, y, test_size=0.2, stratify=y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((4644, 74), (4644,))"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# dimensões do conjunto de treino\n",
    "\n",
    "X_treino.shape, y_treino.shape                              "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((1162, 74), (1162,))"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# dimensões do conjunto de teste\n",
    "\n",
    "X_teste.shape, y_teste.shape       "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array(['Bom', 'Ruim'], dtype=object), array([3439, 1205], dtype=int64))"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Visualização da quantidade de registros por classe treino\n",
    "\n",
    "np.unique(y_treino, return_counts=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array(['Bom', 'Ruim'], dtype=object), array([860, 302], dtype=int64))"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Visualização da quantidade de registros por classe teste\n",
    "\n",
    "np.unique(y_teste, return_counts=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "# classificação com base de dados desbalanceados usando algoritmo Random Forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier()"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# criação do modelo\n",
    "\n",
    "modelo = RandomForestClassifier()                     # definição do modelo na variável modelo\n",
    "modelo.fit(X_treino, y_treino)                        # treinamento do modelo "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# predição dos dados de teste com o modelo treinado\n",
    "\n",
    "previsoes = modelo.predict(X_teste)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['Bom', 'Bom', 'Bom', ..., 'Bom', 'Bom', 'Bom'], dtype=object)"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# visualização dos registros da previsão\n",
    "\n",
    "previsoes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['Ruim', 'Bom', 'Bom', ..., 'Bom', 'Bom', 'Bom'], dtype=object)"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# visualização dos registros reais\n",
    "\n",
    "y_teste"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7409638554216867"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# comparando os registros reais com as previsões\n",
    "\n",
    "accuracy_score(previsoes, y_teste)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[855, 296],\n",
       "       [  5,   6]], dtype=int64)"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# matriz de confusão comparando a qualidade dos registros reais com as previsões\n",
    "\n",
    "cm = confusion_matrix(previsoes, y_teste)\n",
    "cm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# visualização da matriz de confusão graficamente\n",
    "\n",
    "cnn = sns.heatmap(cm, annot=True, fmt='d', cmap='Blues')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7489102005231038"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# percentual de acerto do modelo das pessoas que tem confiança 'boa'\n",
    "\n",
    "859 / (859 + 288)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9333333333333333"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# percentual de acerto do modelo das pessoas que tem confiança 'ruim'\n",
    "\n",
    "14 / (14 + 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Tecnica de subamostragem(undersampling) - Tomek links\n",
    "\n",
    "\n",
    "tl = TomekLinks(sampling_strategy = 'majority')                 # apaga somente registros da classe majoritária(ratio = 'majority')\n",
    "X_under, y_under = tl.fit_resample(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((5417, 74), (5417,))"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Visualização da quantidade de registros\n",
    "\n",
    "X_under.shape, y_under.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array(['Bom', 'Ruim'], dtype=object), array([4299, 1507], dtype=int64))"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Visualização da quantidade de registros por classe da base de dados original\n",
    "\n",
    "np.unique(y, return_counts=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array(['Bom', 'Ruim'], dtype=object), array([3910, 1507], dtype=int64))"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Visualização da quantidade de registros por classe da base de dados com aplicação de subamostragem\n",
    "\n",
    "np.unique(y_under, return_counts=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((4333, 74), (1084,))"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Novo modelo de classificação com nova base de dados\n",
    "X_treino_u, X_teste_u, y_treino_u, y_teste_u = train_test_split(X_under, y_under, test_size=0.2, stratify=y_under)\n",
    "\n",
    "# Visualização da quantidade de registros\n",
    "X_treino_u.shape, y_teste_u.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier()"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# criação do modelo\n",
    "\n",
    "modelo_u = RandomForestClassifier()                         # definição do modelo na variável modelo\n",
    "modelo_u.fit(X_treino_u, y_treino_u)                        # treinamento do modelo com a nova base de dados subamostrada"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "# predição dos dados de teste com o modelo treinado\n",
    "\n",
    "previsoes_u = modelo_u.predict(X_teste_u)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7306273062730627"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# comparando os registros reais com as previsões\n",
    "\n",
    "accuracy_score(previsoes_u, y_teste_u)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[775, 285],\n",
       "       [  7,  17]], dtype=int64)"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# matriz de confusão comparando a qualidade dos registros reais com as previsões\n",
    "\n",
    "cm_u = confusion_matrix(previsoes_u, y_teste_u)\n",
    "cm_u"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# visualização da matriz de confusão graficamente\n",
    "\n",
    "cnn = sns.heatmap(cm_u, annot=True, fmt='d', cmap='Blues')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7272727272727273"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# percentual de acerto do modelo das pessoas que tem confiança 'boa'\n",
    "\n",
    "768 / (768 + 288)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# percentual de acerto do modelo das pessoas que tem confiança 'ruim'\n",
    "\n",
    "14 / (14 + 14)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Tecnica de sobreamostragem(oversampling) - SMOTE\n",
    "\n",
    "smote = SMOTE(sampling_strategy = 'minority')                 # aumenta somente registros da classe minoritária(ratio = 'minority')\n",
    "X_over, y_over = smote.fit_resample(X, y)                     # aplicação da subamostragem na base de dados original"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((8598, 74), (8598,))"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Visualização da quantidade de registros\n",
    "\n",
    "X_over.shape, y_over.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array(['Bom', 'Ruim'], dtype=object), array([4299, 1507], dtype=int64))"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Visualização da quantidade de registros por classe da base de dados original\n",
    "\n",
    "np.unique(y, return_counts=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array(['Bom', 'Ruim'], dtype=object), array([4299, 4299], dtype=int64))"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Visualização da quantidade de registros por classe da base de dados aumetada\n",
    "\n",
    "np.unique(y_over, return_counts=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((6878, 74), (1720,))"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Novo modelo de classificação com nova base de dados\n",
    "X_treino_o, X_teste_o, y_treino_o, y_teste_o = train_test_split(X_over, y_over, test_size=0.2, stratify=y_over)\n",
    "\n",
    "# Visualização da quantidade de registros\n",
    "X_treino_o.shape, y_teste_o.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier()"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# criação do modelo\n",
    "\n",
    "modelo_o = RandomForestClassifier()                         # definição do modelo na variável modelo\n",
    "modelo_o.fit(X_treino_o, y_treino_o)                        # treinamento do modelo com a nova base de dados subamostrada"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "# predição dos dados de teste com o modelo treinado\n",
    "\n",
    "previsoes_o = modelo_o.predict(X_teste_o)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8180232558139535"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# comparando os registros reais com as previsões\n",
    "\n",
    "accuracy_score(previsoes_o, y_teste_o)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[705, 158],\n",
       "       [155, 702]], dtype=int64)"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# matriz de confusão comparando a qualidade dos registros reais com as previsões\n",
    "\n",
    "cm_o = confusion_matrix(previsoes_o, y_teste_o)\n",
    "cm_o"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# visualização da matriz de confusão graficamente\n",
    "\n",
    "cnn = sns.heatmap(cm_o, annot=True, fmt='d', cmap='Blues')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8248792270531401"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# percentual de acerto do modelo das pessoas que tem confiança 'boa'\n",
    "\n",
    "683 / (683 + 145)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8015695067264574"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# percentual de acerto do modelo das pessoas que tem confiança 'ruim'\n",
    "\n",
    "715 / (715 + 177)"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "b3ba2566441a7c06988d0923437866b63cedc61552a5af99d1f4fb67d367b25f"
  },
  "kernelspec": {
   "display_name": "Python 3.8.8 ('base')",
   "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": "3.8.8"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}