Arvores de Decisão.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "<h1 style=\"text-align:center;\">Minerando Dados </h1>\n",
    "<h2 style=\"text-align:center;\"> Árvore de Decisão na prática</h2>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " <img src=\"tabela.png\" width=\"500\" height=\"400\" align=\"center\"/>,     \n",
    "<img src=\"tree.png\" width=\"600\" height=\"500\" align=\"center\"/>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Construção da Árvode\n",
    "- Começa no top node e se baseia no índice de information gain para determinar a melhor divisão \n",
    "- Realiza loop para verificar melhor feature e threshold para realizar a divisão\n",
    "- determina melhor feature e melhor threshold para divisão\n",
    "- Constroi a Árvore de forma recursiva \n",
    "- Determinar critério para determinar até quando (onde) a árvode cresce\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def read_df(file_path):\n",
    "    index_col = 0\n",
    "    # Leitura do csv file pulando primeira coluna\n",
    "    df = pd.read_csv(file_path, index_col = index_col)\n",
    "    #seleciona somente valores não nulos\n",
    "    df = df[df['Y'].notnull()]\n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>brand</th>\n",
       "      <th>price</th>\n",
       "      <th>cartype</th>\n",
       "      <th>model</th>\n",
       "      <th>gearbox</th>\n",
       "      <th>regdate</th>\n",
       "      <th>mileage</th>\n",
       "      <th>motorpower</th>\n",
       "      <th>fuel</th>\n",
       "      <th>car_steering</th>\n",
       "      <th>carcolor</th>\n",
       "      <th>exchange</th>\n",
       "      <th>version</th>\n",
       "      <th>doors</th>\n",
       "      <th>financial</th>\n",
       "      <th>extra</th>\n",
       "      <th>Y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ford</td>\n",
       "      <td>31900</td>\n",
       "      <td>hatch</td>\n",
       "      <td>ka</td>\n",
       "      <td>manual</td>\n",
       "      <td>2015</td>\n",
       "      <td>39869</td>\n",
       "      <td>1</td>\n",
       "      <td>flex</td>\n",
       "      <td>elétrica</td>\n",
       "      <td>preto</td>\n",
       "      <td>sim</td>\n",
       "      <td>ka 1.0 se se plus tivct flex 5p</td>\n",
       "      <td>4 portas</td>\n",
       "      <td>ipva pago</td>\n",
       "      <td>vidro elétrico, air bag, trava elétrica, ar co...</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>vwvolkswagen</td>\n",
       "      <td>26000</td>\n",
       "      <td>passeio</td>\n",
       "      <td>gol</td>\n",
       "      <td>manual</td>\n",
       "      <td>2013</td>\n",
       "      <td>109000</td>\n",
       "      <td>1.6</td>\n",
       "      <td>flex</td>\n",
       "      <td>hidráulica</td>\n",
       "      <td>prata</td>\n",
       "      <td>sim</td>\n",
       "      <td>gol 1.6 mi plus total flex 8v 4p</td>\n",
       "      <td>4 portas</td>\n",
       "      <td>0</td>\n",
       "      <td>vidro elétrico, trava elétrica, ar condicionad...</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>vwvolkswagen</td>\n",
       "      <td>43990</td>\n",
       "      <td>hatch</td>\n",
       "      <td>fox</td>\n",
       "      <td>manual</td>\n",
       "      <td>2017</td>\n",
       "      <td>53000</td>\n",
       "      <td>1.6</td>\n",
       "      <td>flex</td>\n",
       "      <td>elétrica</td>\n",
       "      <td>branco</td>\n",
       "      <td>sim</td>\n",
       "      <td>fox pepper 1.6 flex 16v 5p</td>\n",
       "      <td>4 portas</td>\n",
       "      <td>0</td>\n",
       "      <td>vidro elétrico, air bag, trava elétrica, ar co...</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ford</td>\n",
       "      <td>23900</td>\n",
       "      <td>0</td>\n",
       "      <td>fiesta</td>\n",
       "      <td>manual</td>\n",
       "      <td>2013</td>\n",
       "      <td>70000</td>\n",
       "      <td>1</td>\n",
       "      <td>flex</td>\n",
       "      <td>hidráulica</td>\n",
       "      <td>preto</td>\n",
       "      <td>sim</td>\n",
       "      <td>fiesta class 1.0 4p</td>\n",
       "      <td>4 portas</td>\n",
       "      <td>ipva pago</td>\n",
       "      <td>0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          brand  price  cartype   model gearbox  regdate  mileage motorpower  \\\n",
       "0          ford  31900    hatch      ka  manual     2015    39869          1   \n",
       "1  vwvolkswagen  26000  passeio     gol  manual     2013   109000        1.6   \n",
       "0  vwvolkswagen  43990    hatch     fox  manual     2017    53000        1.6   \n",
       "3          ford  23900        0  fiesta  manual     2013    70000          1   \n",
       "\n",
       "   fuel car_steering carcolor exchange                           version  \\\n",
       "0  flex     elétrica    preto      sim   ka 1.0 se se plus tivct flex 5p   \n",
       "1  flex   hidráulica    prata      sim  gol 1.6 mi plus total flex 8v 4p   \n",
       "0  flex     elétrica   branco      sim        fox pepper 1.6 flex 16v 5p   \n",
       "3  flex   hidráulica    preto      sim               fiesta class 1.0 4p   \n",
       "\n",
       "      doors  financial                                              extra    Y  \n",
       "0  4 portas  ipva pago  vidro elétrico, air bag, trava elétrica, ar co...  1.0  \n",
       "1  4 portas          0  vidro elétrico, trava elétrica, ar condicionad...  1.0  \n",
       "0  4 portas          0  vidro elétrico, air bag, trava elétrica, ar co...  0.0  \n",
       "3  4 portas  ipva pago                                                  0  1.0  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "file_path = 'car_information.csv'\n",
    "df = read_df(file_path)\n",
    "df.head(4)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Explorando o dataframe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(4554, 17)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# verificar shape do dataset\n",
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "brand            object\n",
       "price             int64\n",
       "cartype          object\n",
       "model            object\n",
       "gearbox          object\n",
       "regdate           int64\n",
       "mileage           int64\n",
       "motorpower       object\n",
       "fuel             object\n",
       "car_steering     object\n",
       "carcolor         object\n",
       "exchange         object\n",
       "version          object\n",
       "doors            object\n",
       "financial        object\n",
       "extra            object\n",
       "Y               float64\n",
       "dtype: object"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "brand           0\n",
       "price           0\n",
       "cartype         0\n",
       "model           0\n",
       "gearbox         0\n",
       "regdate         0\n",
       "mileage         0\n",
       "motorpower      0\n",
       "fuel            0\n",
       "car_steering    0\n",
       "carcolor        0\n",
       "exchange        0\n",
       "version         0\n",
       "doors           0\n",
       "financial       0\n",
       "extra           0\n",
       "Y               0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Verificar valores na\n",
    "df.isna().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f08fcfbf6d8>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df.model.value_counts().plot(kind='bar', xlabel = 'Modelo do carro', ylabel ='Quantidade', title = 'Quantidade de Carros')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f08fcea28d0>"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df.regdate.value_counts().plot(kind='bar', xlabel = 'Ano', ylabel ='Quantidade', title = 'Quantidade de Carros por ano de registro')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f08fce14550>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df.groupby('regdate').mean()['price'].plot(kind='bar')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "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>brand</th>\n",
       "      <th>price</th>\n",
       "      <th>cartype</th>\n",
       "      <th>model</th>\n",
       "      <th>gearbox</th>\n",
       "      <th>mileage</th>\n",
       "      <th>motorpower</th>\n",
       "      <th>fuel</th>\n",
       "      <th>car_steering</th>\n",
       "      <th>carcolor</th>\n",
       "      <th>exchange</th>\n",
       "      <th>version</th>\n",
       "      <th>doors</th>\n",
       "      <th>financial</th>\n",
       "      <th>extra</th>\n",
       "      <th>Y</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>regdate</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2013</th>\n",
       "      <td>vwvolkswagen</td>\n",
       "      <td>75000</td>\n",
       "      <td>vanutilitário</td>\n",
       "      <td>voyage</td>\n",
       "      <td>semiautomático</td>\n",
       "      <td>51000000</td>\n",
       "      <td>3.03.9</td>\n",
       "      <td>flex</td>\n",
       "      <td>mecnica</td>\n",
       "      <td>vermelho</td>\n",
       "      <td>sim</td>\n",
       "      <td>voyage trendline 1.6 t.flex 8v 4p</td>\n",
       "      <td>4 portas</td>\n",
       "      <td>ipva pago</td>\n",
       "      <td>vidro elétrico, trava elétrica, direção hidráu...</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2014</th>\n",
       "      <td>vwvolkswagen</td>\n",
       "      <td>110000</td>\n",
       "      <td>vanutilitário</td>\n",
       "      <td>voyage</td>\n",
       "      <td>semiautomático</td>\n",
       "      <td>9999999999</td>\n",
       "      <td>4.0oumais</td>\n",
       "      <td>flex</td>\n",
       "      <td>mecnica</td>\n",
       "      <td>vermelho</td>\n",
       "      <td>sim</td>\n",
       "      <td>voyage trendline 1.6 t.flex 8v 4p</td>\n",
       "      <td>4 portas</td>\n",
       "      <td>ipva pago</td>\n",
       "      <td>vidro elétrico, trava elétrica, direção hidráu...</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015</th>\n",
       "      <td>vwvolkswagen</td>\n",
       "      <td>29900000</td>\n",
       "      <td>vanutilitário</td>\n",
       "      <td>voyage</td>\n",
       "      <td>semiautomático</td>\n",
       "      <td>76800000</td>\n",
       "      <td>2.02.9</td>\n",
       "      <td>flex</td>\n",
       "      <td>mecnica</td>\n",
       "      <td>vermelho</td>\n",
       "      <td>sim</td>\n",
       "      <td>voyage trendline 1.6 t.flex 8v 4p</td>\n",
       "      <td>4 portas</td>\n",
       "      <td>ipva pago</td>\n",
       "      <td>vidro elétrico, trava elétrica, ar condicionad...</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016</th>\n",
       "      <td>vwvolkswagen</td>\n",
       "      <td>102999</td>\n",
       "      <td>vanutilitário</td>\n",
       "      <td>voyage</td>\n",
       "      <td>semiautomático</td>\n",
       "      <td>66000000</td>\n",
       "      <td>2.02.9</td>\n",
       "      <td>flex</td>\n",
       "      <td>mecnica</td>\n",
       "      <td>vermelho</td>\n",
       "      <td>sim</td>\n",
       "      <td>voyage trendline 1.6 t.flex 8v 4p</td>\n",
       "      <td>4 portas</td>\n",
       "      <td>ipva pago</td>\n",
       "      <td>vidro elétrico, trava elétrica, ar condicionad...</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017</th>\n",
       "      <td>vwvolkswagen</td>\n",
       "      <td>400000</td>\n",
       "      <td>vanutilitário</td>\n",
       "      <td>voyage</td>\n",
       "      <td>semiautomático</td>\n",
       "      <td>72000000</td>\n",
       "      <td>2.02.9</td>\n",
       "      <td>flex</td>\n",
       "      <td>mecnica</td>\n",
       "      <td>vermelho</td>\n",
       "      <td>sim</td>\n",
       "      <td>voyage trendline 1.6 t.flex 8v 4p</td>\n",
       "      <td>4 portas</td>\n",
       "      <td>ipva pago</td>\n",
       "      <td>vidro elétrico, trava elétrica, ar condicionad...</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018</th>\n",
       "      <td>vwvolkswagen</td>\n",
       "      <td>154900</td>\n",
       "      <td>vanutilitário</td>\n",
       "      <td>voyage</td>\n",
       "      <td>semiautomático</td>\n",
       "      <td>22000000</td>\n",
       "      <td>2.02.9</td>\n",
       "      <td>flex</td>\n",
       "      <td>mecnica</td>\n",
       "      <td>vermelho</td>\n",
       "      <td>sim</td>\n",
       "      <td>voyage trendline 1.6 t.flex 8v 4p</td>\n",
       "      <td>4 portas</td>\n",
       "      <td>ipva pago</td>\n",
       "      <td>vidro elétrico, trava elétrica, ar condicionad...</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                brand     price        cartype   model         gearbox  \\\n",
       "regdate                                                                  \n",
       "0                   0         0              0       0               0   \n",
       "2013     vwvolkswagen     75000  vanutilitário  voyage  semiautomático   \n",
       "2014     vwvolkswagen    110000  vanutilitário  voyage  semiautomático   \n",
       "2015     vwvolkswagen  29900000  vanutilitário  voyage  semiautomático   \n",
       "2016     vwvolkswagen    102999  vanutilitário  voyage  semiautomático   \n",
       "2017     vwvolkswagen    400000  vanutilitário  voyage  semiautomático   \n",
       "2018     vwvolkswagen    154900  vanutilitário  voyage  semiautomático   \n",
       "\n",
       "            mileage motorpower  fuel car_steering  carcolor exchange  \\\n",
       "regdate                                                                \n",
       "0                 0          0     0            0         0        0   \n",
       "2013       51000000     3.03.9  flex      mecnica  vermelho      sim   \n",
       "2014     9999999999  4.0oumais  flex      mecnica  vermelho      sim   \n",
       "2015       76800000     2.02.9  flex      mecnica  vermelho      sim   \n",
       "2016       66000000     2.02.9  flex      mecnica  vermelho      sim   \n",
       "2017       72000000     2.02.9  flex      mecnica  vermelho      sim   \n",
       "2018       22000000     2.02.9  flex      mecnica  vermelho      sim   \n",
       "\n",
       "                                   version     doors  financial  \\\n",
       "regdate                                                           \n",
       "0                                        0         0          0   \n",
       "2013     voyage trendline 1.6 t.flex 8v 4p  4 portas  ipva pago   \n",
       "2014     voyage trendline 1.6 t.flex 8v 4p  4 portas  ipva pago   \n",
       "2015     voyage trendline 1.6 t.flex 8v 4p  4 portas  ipva pago   \n",
       "2016     voyage trendline 1.6 t.flex 8v 4p  4 portas  ipva pago   \n",
       "2017     voyage trendline 1.6 t.flex 8v 4p  4 portas  ipva pago   \n",
       "2018     voyage trendline 1.6 t.flex 8v 4p  4 portas  ipva pago   \n",
       "\n",
       "                                                     extra    Y  \n",
       "regdate                                                          \n",
       "0                                                        0  0.0  \n",
       "2013     vidro elétrico, trava elétrica, direção hidráu...  1.0  \n",
       "2014     vidro elétrico, trava elétrica, direção hidráu...  1.0  \n",
       "2015     vidro elétrico, trava elétrica, ar condicionad...  1.0  \n",
       "2016     vidro elétrico, trava elétrica, ar condicionad...  1.0  \n",
       "2017     vidro elétrico, trava elétrica, ar condicionad...  1.0  \n",
       "2018     vidro elétrico, trava elétrica, ar condicionad...  1.0  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.groupby('regdate').max()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "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>brand</th>\n",
       "      <th>price</th>\n",
       "      <th>cartype</th>\n",
       "      <th>model</th>\n",
       "      <th>gearbox</th>\n",
       "      <th>mileage</th>\n",
       "      <th>motorpower</th>\n",
       "      <th>fuel</th>\n",
       "      <th>car_steering</th>\n",
       "      <th>carcolor</th>\n",
       "      <th>exchange</th>\n",
       "      <th>version</th>\n",
       "      <th>doors</th>\n",
       "      <th>financial</th>\n",
       "      <th>extra</th>\n",
       "      <th>Y</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>regdate</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2013</th>\n",
       "      <td>ford</td>\n",
       "      <td>5000</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>flex</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2014</th>\n",
       "      <td>ford</td>\n",
       "      <td>100</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>flex</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015</th>\n",
       "      <td>ford</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>flex</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016</th>\n",
       "      <td>ford</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>flex</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017</th>\n",
       "      <td>ford</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>flex</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018</th>\n",
       "      <td>ford</td>\n",
       "      <td>7000</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>flex</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        brand  price cartype model gearbox  mileage motorpower  fuel  \\\n",
       "regdate                                                                \n",
       "0           0      0       0     0       0        0          0     0   \n",
       "2013     ford   5000       0     0       0        0          0  flex   \n",
       "2014     ford    100       0     0       0        0          0  flex   \n",
       "2015     ford      0       0     0       0        0          0  flex   \n",
       "2016     ford      0       0     0       0        0          0  flex   \n",
       "2017     ford      0       0     0       0        0          0  flex   \n",
       "2018     ford   7000       0     0       0        0          0  flex   \n",
       "\n",
       "        car_steering carcolor exchange version doors financial extra    Y  \n",
       "regdate                                                                    \n",
       "0                  0        0        0       0     0         0     0  0.0  \n",
       "2013               0        0        0       0     0         0     0  0.0  \n",
       "2014               0        0        0       0     0         0     0  0.0  \n",
       "2015               0        0        0       0     0         0     0  0.0  \n",
       "2016               0        0        0       0     0         0     0  0.0  \n",
       "2017               0        0        0       0     0         0     0  0.0  \n",
       "2018               0        0        0       0     0         0     0  0.0  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.groupby('regdate').min()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Número de linhas: 59\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>brand</th>\n",
       "      <th>price</th>\n",
       "      <th>cartype</th>\n",
       "      <th>model</th>\n",
       "      <th>gearbox</th>\n",
       "      <th>regdate</th>\n",
       "      <th>mileage</th>\n",
       "      <th>motorpower</th>\n",
       "      <th>fuel</th>\n",
       "      <th>car_steering</th>\n",
       "      <th>carcolor</th>\n",
       "      <th>exchange</th>\n",
       "      <th>version</th>\n",
       "      <th>doors</th>\n",
       "      <th>financial</th>\n",
       "      <th>extra</th>\n",
       "      <th>Y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>142</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>321</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>349</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    brand  price cartype model gearbox  regdate  mileage motorpower fuel  \\\n",
       "142     0      0       0     0       0        0        0          0    0   \n",
       "321     0      0       0     0       0        0        0          0    0   \n",
       "349     0      0       0     0       0        0        0          0    0   \n",
       "\n",
       "    car_steering carcolor exchange version doors financial extra    Y  \n",
       "142            0        0        0       0     0         0     0  0.0  \n",
       "321            0        0        0       0     0         0     0  0.0  \n",
       "349            0        0        0       0     0         0     0  0.0  "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('Número de linhas: {}'.format(df[df.price <= 5000].shape[0]))\n",
    "df[df.price <= 5000].head(3)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Número de linhas: 143\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>brand</th>\n",
       "      <th>price</th>\n",
       "      <th>cartype</th>\n",
       "      <th>model</th>\n",
       "      <th>gearbox</th>\n",
       "      <th>regdate</th>\n",
       "      <th>mileage</th>\n",
       "      <th>motorpower</th>\n",
       "      <th>fuel</th>\n",
       "      <th>car_steering</th>\n",
       "      <th>carcolor</th>\n",
       "      <th>exchange</th>\n",
       "      <th>version</th>\n",
       "      <th>doors</th>\n",
       "      <th>financial</th>\n",
       "      <th>extra</th>\n",
       "      <th>Y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>ford</td>\n",
       "      <td>75000</td>\n",
       "      <td>0</td>\n",
       "      <td>focus</td>\n",
       "      <td>automático</td>\n",
       "      <td>2018</td>\n",
       "      <td>25000</td>\n",
       "      <td>2.02.9</td>\n",
       "      <td>flex</td>\n",
       "      <td>elétrica</td>\n",
       "      <td>prata</td>\n",
       "      <td>sim</td>\n",
       "      <td>focus tita tita plus 2.0  flex 5p aut.</td>\n",
       "      <td>4 portas</td>\n",
       "      <td>0</td>\n",
       "      <td>vidro elétrico, air bag, trava elétrica, ar co...</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>47</th>\n",
       "      <td>ford</td>\n",
       "      <td>77900</td>\n",
       "      <td>passeio</td>\n",
       "      <td>fusion</td>\n",
       "      <td>automático</td>\n",
       "      <td>2016</td>\n",
       "      <td>46231</td>\n",
       "      <td>2.02.9</td>\n",
       "      <td>flex</td>\n",
       "      <td>hidráulica</td>\n",
       "      <td>preto</td>\n",
       "      <td>sim</td>\n",
       "      <td>0</td>\n",
       "      <td>4 portas</td>\n",
       "      <td>0</td>\n",
       "      <td>vidro elétrico, air bag, trava elétrica, ar co...</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   brand  price  cartype   model     gearbox  regdate  mileage motorpower  \\\n",
       "6   ford  75000        0   focus  automático     2018    25000     2.02.9   \n",
       "47  ford  77900  passeio  fusion  automático     2016    46231     2.02.9   \n",
       "\n",
       "    fuel car_steering carcolor exchange  \\\n",
       "6   flex     elétrica    prata      sim   \n",
       "47  flex   hidráulica    preto      sim   \n",
       "\n",
       "                                   version     doors financial  \\\n",
       "6   focus tita tita plus 2.0  flex 5p aut.  4 portas         0   \n",
       "47                                       0  4 portas         0   \n",
       "\n",
       "                                                extra    Y  \n",
       "6   vidro elétrico, air bag, trava elétrica, ar co...  0.0  \n",
       "47  vidro elétrico, air bag, trava elétrica, ar co...  0.0  "
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('Número de linhas: {}'.format(df[df.price >= 70000].shape[0]))\n",
    "df[df.price >= 70000].head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Int64Index([   6,   47,   72,   85,  155,  192,  260,  327,  343,  352,\n",
       "            ...\n",
       "            6229, 6251, 6362, 6365, 6383, 6529, 6564, 6611, 6712, 6737],\n",
       "           dtype='int64', length=143)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# seleciona index\n",
    "index_menor_5k = df[df.price <= 5000].index\n",
    "index_maior_70k = df[df.price >= 70000].index\n",
    "index_maior_70k"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "def drop_rows(df, index):\n",
    "    df.drop(index, inplace =True)  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "drop_rows(df,index_menor_5k)\n",
    "drop_rows(df,index_maior_70k)\n",
    "df = df.reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "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>brand</th>\n",
       "      <th>price</th>\n",
       "      <th>cartype</th>\n",
       "      <th>model</th>\n",
       "      <th>gearbox</th>\n",
       "      <th>regdate</th>\n",
       "      <th>mileage</th>\n",
       "      <th>motorpower</th>\n",
       "      <th>fuel</th>\n",
       "      <th>car_steering</th>\n",
       "      <th>carcolor</th>\n",
       "      <th>exchange</th>\n",
       "      <th>version</th>\n",
       "      <th>doors</th>\n",
       "      <th>financial</th>\n",
       "      <th>extra</th>\n",
       "      <th>Y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ford</td>\n",
       "      <td>31900</td>\n",
       "      <td>hatch</td>\n",
       "      <td>ka</td>\n",
       "      <td>manual</td>\n",
       "      <td>2015</td>\n",
       "      <td>39869</td>\n",
       "      <td>1</td>\n",
       "      <td>flex</td>\n",
       "      <td>elétrica</td>\n",
       "      <td>preto</td>\n",
       "      <td>sim</td>\n",
       "      <td>ka 1.0 se se plus tivct flex 5p</td>\n",
       "      <td>4 portas</td>\n",
       "      <td>ipva pago</td>\n",
       "      <td>vidro elétrico, air bag, trava elétrica, ar co...</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>vwvolkswagen</td>\n",
       "      <td>26000</td>\n",
       "      <td>passeio</td>\n",
       "      <td>gol</td>\n",
       "      <td>manual</td>\n",
       "      <td>2013</td>\n",
       "      <td>109000</td>\n",
       "      <td>1.6</td>\n",
       "      <td>flex</td>\n",
       "      <td>hidráulica</td>\n",
       "      <td>prata</td>\n",
       "      <td>sim</td>\n",
       "      <td>gol 1.6 mi plus total flex 8v 4p</td>\n",
       "      <td>4 portas</td>\n",
       "      <td>0</td>\n",
       "      <td>vidro elétrico, trava elétrica, ar condicionad...</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>vwvolkswagen</td>\n",
       "      <td>43990</td>\n",
       "      <td>hatch</td>\n",
       "      <td>fox</td>\n",
       "      <td>manual</td>\n",
       "      <td>2017</td>\n",
       "      <td>53000</td>\n",
       "      <td>1.6</td>\n",
       "      <td>flex</td>\n",
       "      <td>elétrica</td>\n",
       "      <td>branco</td>\n",
       "      <td>sim</td>\n",
       "      <td>fox pepper 1.6 flex 16v 5p</td>\n",
       "      <td>4 portas</td>\n",
       "      <td>0</td>\n",
       "      <td>vidro elétrico, air bag, trava elétrica, ar co...</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          brand  price  cartype model gearbox  regdate  mileage motorpower  \\\n",
       "0          ford  31900    hatch    ka  manual     2015    39869          1   \n",
       "1  vwvolkswagen  26000  passeio   gol  manual     2013   109000        1.6   \n",
       "2  vwvolkswagen  43990    hatch   fox  manual     2017    53000        1.6   \n",
       "\n",
       "   fuel car_steering carcolor exchange                           version  \\\n",
       "0  flex     elétrica    preto      sim   ka 1.0 se se plus tivct flex 5p   \n",
       "1  flex   hidráulica    prata      sim  gol 1.6 mi plus total flex 8v 4p   \n",
       "2  flex     elétrica   branco      sim        fox pepper 1.6 flex 16v 5p   \n",
       "\n",
       "      doors  financial                                              extra    Y  \n",
       "0  4 portas  ipva pago  vidro elétrico, air bag, trava elétrica, ar co...  1.0  \n",
       "1  4 portas          0  vidro elétrico, trava elétrica, ar condicionad...  1.0  \n",
       "2  4 portas          0  vidro elétrico, air bag, trava elétrica, ar co...  0.0  "
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mediana do preço: 35900.0\n",
      "Média do preço: 36866.89613970588\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df.price.hist()\n",
    "print('Mediana do preço:', df.price.median())\n",
    "print('Média do preço:', df.price.mean())\n",
    "print()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Selecionar colunas para modelo 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "cols = ['price','regdate', 'Y']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "#df = df.reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "# shuffle dataframe\n",
    "from sklearn.utils import shuffle\n",
    "df = shuffle(df).reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "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>price</th>\n",
       "      <th>regdate</th>\n",
       "      <th>Y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>58900</td>\n",
       "      <td>2018</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>59901</td>\n",
       "      <td>2017</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   price  regdate    Y\n",
       "0  58900     2018  0.0\n",
       "1  59901     2017  0.0"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "simple_df = df[cols]\n",
    "simple_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Separando dados features e target\n",
    "Y = simple_df['Y']\n",
    "X = simple_df.loc[:, 'price':'regdate']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((2176, 2), (2176, 2), (2176,), (2176,))"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Separa dados de treino e validação\n",
    "Xtrain, Xval, y_train, y_val = train_test_split(X, Y, test_size=0.5, random_state=42)\n",
    "Xtrain.shape, Xval.shape, y_train.shape, y_val.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.0    1854\n",
       "1.0     322\n",
       "Name: Y, dtype: int64"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DecisionTreeClassifier(random_state=42)"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Modelo\n",
    "model = DecisionTreeClassifier(random_state= 42)\n",
    "model.fit(Xtrain,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Previsão\n",
    "pred = model.predict(Xval)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Importa métricas\n",
    "from sklearn.metrics import roc_auc_score, average_precision_score"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- **roc_auc_score** : Nos informa o quanto o nosso modelo está conseguindo distinguir as classes, ou seja, classificar 1 como 1 e 0 como 0.\n",
    "- **average_precision_score** : A average precision indica se nosso modelo pode identificar corretamente todos os exemplos positivos sem marcar acidentalmente muitos exemplos negativos como positivos. Portanto, a average precision é alta quando seu modelo pode classificar corretamente os positivos. A average precision é calculada como a área sob uma curva que mede a compensação entre precisão e recall em diferentes limites de decisão"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "average_precision_score : 0.2972\n",
      "roc_auc_score : 0.6562\n"
     ]
    }
   ],
   "source": [
    "print('average_precision_score :', np.round(average_precision_score(y_val, pred),4))\n",
    "print('roc_auc_score :', np.round(roc_auc_score(y_val, pred),4))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Populating the interactive namespace from numpy and matplotlib\n"
     ]
    }
   ],
   "source": [
    "from sklearn.tree import plot_tree\n",
    "%matplotlib inline\n",
    "%pylab inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x1080 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig,ax = pylab.subplots(1,1, figsize = (15,15))\n",
    "plot_tree(model , ax = ax, feature_names = Xtrain.columns);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Estratégia 1 : Adicionar features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "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>brand</th>\n",
       "      <th>price</th>\n",
       "      <th>cartype</th>\n",
       "      <th>model</th>\n",
       "      <th>gearbox</th>\n",
       "      <th>regdate</th>\n",
       "      <th>mileage</th>\n",
       "      <th>motorpower</th>\n",
       "      <th>fuel</th>\n",
       "      <th>car_steering</th>\n",
       "      <th>carcolor</th>\n",
       "      <th>exchange</th>\n",
       "      <th>version</th>\n",
       "      <th>doors</th>\n",
       "      <th>financial</th>\n",
       "      <th>extra</th>\n",
       "      <th>Y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ford</td>\n",
       "      <td>31900</td>\n",
       "      <td>hatch</td>\n",
       "      <td>ka</td>\n",
       "      <td>manual</td>\n",
       "      <td>2015</td>\n",
       "      <td>39869</td>\n",
       "      <td>1</td>\n",
       "      <td>flex</td>\n",
       "      <td>elétrica</td>\n",
       "      <td>preto</td>\n",
       "      <td>sim</td>\n",
       "      <td>ka 1.0 se se plus tivct flex 5p</td>\n",
       "      <td>4 portas</td>\n",
       "      <td>ipva pago</td>\n",
       "      <td>vidro elétrico, air bag, trava elétrica, ar co...</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>vwvolkswagen</td>\n",
       "      <td>26000</td>\n",
       "      <td>passeio</td>\n",
       "      <td>gol</td>\n",
       "      <td>manual</td>\n",
       "      <td>2013</td>\n",
       "      <td>109000</td>\n",
       "      <td>1.6</td>\n",
       "      <td>flex</td>\n",
       "      <td>hidráulica</td>\n",
       "      <td>prata</td>\n",
       "      <td>sim</td>\n",
       "      <td>gol 1.6 mi plus total flex 8v 4p</td>\n",
       "      <td>4 portas</td>\n",
       "      <td>0</td>\n",
       "      <td>vidro elétrico, trava elétrica, ar condicionad...</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>vwvolkswagen</td>\n",
       "      <td>43990</td>\n",
       "      <td>hatch</td>\n",
       "      <td>fox</td>\n",
       "      <td>manual</td>\n",
       "      <td>2017</td>\n",
       "      <td>53000</td>\n",
       "      <td>1.6</td>\n",
       "      <td>flex</td>\n",
       "      <td>elétrica</td>\n",
       "      <td>branco</td>\n",
       "      <td>sim</td>\n",
       "      <td>fox pepper 1.6 flex 16v 5p</td>\n",
       "      <td>4 portas</td>\n",
       "      <td>0</td>\n",
       "      <td>vidro elétrico, air bag, trava elétrica, ar co...</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ford</td>\n",
       "      <td>23900</td>\n",
       "      <td>0</td>\n",
       "      <td>fiesta</td>\n",
       "      <td>manual</td>\n",
       "      <td>2013</td>\n",
       "      <td>70000</td>\n",
       "      <td>1</td>\n",
       "      <td>flex</td>\n",
       "      <td>hidráulica</td>\n",
       "      <td>preto</td>\n",
       "      <td>sim</td>\n",
       "      <td>fiesta class 1.0 4p</td>\n",
       "      <td>4 portas</td>\n",
       "      <td>ipva pago</td>\n",
       "      <td>0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          brand  price  cartype   model gearbox  regdate  mileage motorpower  \\\n",
       "0          ford  31900    hatch      ka  manual     2015    39869          1   \n",
       "1  vwvolkswagen  26000  passeio     gol  manual     2013   109000        1.6   \n",
       "0  vwvolkswagen  43990    hatch     fox  manual     2017    53000        1.6   \n",
       "3          ford  23900        0  fiesta  manual     2013    70000          1   \n",
       "\n",
       "   fuel car_steering carcolor exchange                           version  \\\n",
       "0  flex     elétrica    preto      sim   ka 1.0 se se plus tivct flex 5p   \n",
       "1  flex   hidráulica    prata      sim  gol 1.6 mi plus total flex 8v 4p   \n",
       "0  flex     elétrica   branco      sim        fox pepper 1.6 flex 16v 5p   \n",
       "3  flex   hidráulica    preto      sim               fiesta class 1.0 4p   \n",
       "\n",
       "      doors  financial                                              extra    Y  \n",
       "0  4 portas  ipva pago  vidro elétrico, air bag, trava elétrica, ar co...  1.0  \n",
       "1  4 portas          0  vidro elétrico, trava elétrica, ar condicionad...  1.0  \n",
       "0  4 portas          0  vidro elétrico, air bag, trava elétrica, ar co...  0.0  \n",
       "3  4 portas  ipva pago                                                  0  1.0  "
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "file_path = 'car_information.csv'\n",
    "df = read_df(file_path)\n",
    "df.head(4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "cols = ['price','regdate','mileage', 'Y']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.utils import shuffle\n",
    "df = shuffle(df).reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "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>price</th>\n",
       "      <th>regdate</th>\n",
       "      <th>mileage</th>\n",
       "      <th>Y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>38900</td>\n",
       "      <td>2016</td>\n",
       "      <td>46158</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>39920</td>\n",
       "      <td>2015</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   price  regdate  mileage    Y\n",
       "0  38900     2016    46158  0.0\n",
       "1  39920     2015        0  0.0"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "simple_df = df[cols]\n",
    "simple_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "# Separando dados features e target\n",
    "Y = simple_df['Y']\n",
    "X = simple_df.loc[:, 'price':'mileage']\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((2277, 3), (2277, 3), (2277,), (2277,))"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Separa dados de treino e validação\n",
    "Xtrain, Xval, y_train, y_val = train_test_split(X, Y, test_size=0.5, random_state=42)\n",
    "Xtrain.shape, Xval.shape, y_train.shape, y_val.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DecisionTreeClassifier(random_state=42)"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Modelo\n",
    "model = DecisionTreeClassifier(random_state= 42)\n",
    "model.fit(Xtrain,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Previsão\n",
    "pred = model.predict(Xval)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "average_precision_score : 0.3296\n",
      "roc_auc_score : 0.7158\n"
     ]
    }
   ],
   "source": [
    "# Importa métricas\n",
    "from sklearn.metrics import roc_auc_score, average_precision_score\n",
    "print('average_precision_score :', np.round(average_precision_score(y_val, pred),4))\n",
    "print('roc_auc_score :', np.round(roc_auc_score(y_val, pred),4))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Possível estratégia"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Engenharia de features\n",
    "- Codificar variáveis categóricas\n",
    "- Adicionar mais features ao modelo"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "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>brand</th>\n",
       "      <th>price</th>\n",
       "      <th>cartype</th>\n",
       "      <th>model</th>\n",
       "      <th>gearbox</th>\n",
       "      <th>regdate</th>\n",
       "      <th>mileage</th>\n",
       "      <th>motorpower</th>\n",
       "      <th>fuel</th>\n",
       "      <th>car_steering</th>\n",
       "      <th>carcolor</th>\n",
       "      <th>exchange</th>\n",
       "      <th>version</th>\n",
       "      <th>doors</th>\n",
       "      <th>financial</th>\n",
       "      <th>extra</th>\n",
       "      <th>Y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>vwvolkswagen</td>\n",
       "      <td>38900</td>\n",
       "      <td>hatch</td>\n",
       "      <td>fox</td>\n",
       "      <td>manual</td>\n",
       "      <td>2016</td>\n",
       "      <td>46158</td>\n",
       "      <td>1.6</td>\n",
       "      <td>flex</td>\n",
       "      <td>hidráulica</td>\n",
       "      <td>branco</td>\n",
       "      <td>sim</td>\n",
       "      <td>fox rock in rio 1.6 mi total flex 8v 5p</td>\n",
       "      <td>4 portas</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ford</td>\n",
       "      <td>39920</td>\n",
       "      <td>passeio</td>\n",
       "      <td>ecosport</td>\n",
       "      <td>automático</td>\n",
       "      <td>2015</td>\n",
       "      <td>0</td>\n",
       "      <td>2.02.9</td>\n",
       "      <td>flex</td>\n",
       "      <td>elétrica</td>\n",
       "      <td>branco</td>\n",
       "      <td>sim</td>\n",
       "      <td>0</td>\n",
       "      <td>4 portas</td>\n",
       "      <td>0</td>\n",
       "      <td>vidro elétrico, air bag, trava elétrica, ar co...</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>vwvolkswagen</td>\n",
       "      <td>30000</td>\n",
       "      <td>hatch</td>\n",
       "      <td>gol</td>\n",
       "      <td>manual</td>\n",
       "      <td>2015</td>\n",
       "      <td>41000</td>\n",
       "      <td>1</td>\n",
       "      <td>flex</td>\n",
       "      <td>hidráulica</td>\n",
       "      <td>branco</td>\n",
       "      <td>sim</td>\n",
       "      <td>gol comfortline 1.0 t. flex 8v 5p</td>\n",
       "      <td>4 portas</td>\n",
       "      <td>ipva pago</td>\n",
       "      <td>vidro elétrico, air bag, trava elétrica, ar co...</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          brand  price  cartype     model     gearbox  regdate  mileage  \\\n",
       "0  vwvolkswagen  38900    hatch       fox      manual     2016    46158   \n",
       "1          ford  39920  passeio  ecosport  automático     2015        0   \n",
       "2  vwvolkswagen  30000    hatch       gol      manual     2015    41000   \n",
       "\n",
       "  motorpower  fuel car_steering carcolor exchange  \\\n",
       "0        1.6  flex   hidráulica   branco      sim   \n",
       "1     2.02.9  flex     elétrica   branco      sim   \n",
       "2          1  flex   hidráulica   branco      sim   \n",
       "\n",
       "                                   version     doors  financial  \\\n",
       "0  fox rock in rio 1.6 mi total flex 8v 5p  4 portas          0   \n",
       "1                                        0  4 portas          0   \n",
       "2        gol comfortline 1.0 t. flex 8v 5p  4 portas  ipva pago   \n",
       "\n",
       "                                               extra    Y  \n",
       "0                                                  0  0.0  \n",
       "1  vidro elétrico, air bag, trava elétrica, ar co...  0.0  \n",
       "2  vidro elétrico, air bag, trava elétrica, ar co...  0.0  "
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Estratégia 2 : Tuning"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Parametros :\n",
    "    \n",
    "- max_depth \n",
    "    - Indica o quão profunda a árvore de decisão será\n",
    "    - Quanto mais profunda a árvore, mais complexo será o modelo (maior número de divisões)\n",
    "    - Pode impactar significativamente o demolo em termos de overfitting\n",
    "- min_samples_split\n",
    "    - Especifica o valor mínimo da amostra necessárias para dividir um node interno \n",
    "    - Pode fornecer um número para especificar o valor correto ou um float para expecificar uma fração dos dados\n",
    "- min_samples_leaf:\n",
    "    - A folha é um node sem nenhum filho - ou sem divisão\n",
    "    - min_sample_leaf representa o número mínimo de amostras necessárias em cada folha\n",
    "    - É similar ao min_sample_split porém, especifica o valor mínimo em cada folha (no final da árvore)\n",
    "    - Documentação do scikit-learn recomenda começar com min_samples_leaf=5 \n",
    "    - Documentação do scikit-learn recomenda começar com min_samples_leaf=1 em problemas de classificação com poucas classes\n",
    "- max_features\n",
    "    - Representa o número de features serão utilizadas para realizar a divisão dos dados\n",
    "    - Pode ser passado valor inteiro, float, \"sqrt\" ( sqrt(features), \"log2\" e None\n",
    "    \n",
    "- creterion\n",
    "    - A função que irá determinar a qualidade das divisões\n",
    "    - gini ou entropy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Bayesian Optimization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "from skopt import forest_minimize"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Iteration No: 1 started. Evaluating function at random point.\n",
      "[21, 19, 8, 1, 'entropy']\n",
      "\n",
      "roc_auc_score : 0.833662\n",
      "average_precision_score : 0.385035\n",
      "\n",
      "Iteration No: 1 ended. Evaluation done at random point.\n",
      "Time taken: 0.0410\n",
      "Function value obtained: -0.3850\n",
      "Current minimum: -0.3850\n",
      "Iteration No: 2 started. Evaluating function at random point.\n",
      "[27, 23, 7, 3, 'gini']\n",
      "\n",
      "roc_auc_score : 0.832411\n",
      "average_precision_score : 0.395641\n",
      "\n",
      "Iteration No: 2 ended. Evaluation done at random point.\n",
      "Time taken: 0.0353\n",
      "Function value obtained: -0.3956\n",
      "Current minimum: -0.3956\n",
      "Iteration No: 3 started. Evaluating function at random point.\n",
      "[22, 8, 8, 3, 'gini']\n",
      "\n",
      "roc_auc_score : 0.823962\n",
      "average_precision_score : 0.390791\n",
      "\n",
      "Iteration No: 3 ended. Evaluation done at random point.\n",
      "Time taken: 0.0567\n",
      "Function value obtained: -0.3908\n",
      "Current minimum: -0.3956\n",
      "Iteration No: 4 started. Evaluating function at random point.\n",
      "[3, 28, 6, 2, 'gini']\n",
      "\n",
      "roc_auc_score : 0.864847\n",
      "average_precision_score : 0.420949\n",
      "\n",
      "Iteration No: 4 ended. Evaluation done at random point.\n",
      "Time taken: 0.0439\n",
      "Function value obtained: -0.4209\n",
      "Current minimum: -0.4209\n",
      "Iteration No: 5 started. Evaluating function at random point.\n",
      "[22, 5, 10, 2, 'gini']\n",
      "\n",
      "roc_auc_score : 0.833808\n",
      "average_precision_score : 0.391806\n",
      "\n",
      "Iteration No: 5 ended. Evaluation done at random point.\n",
      "Time taken: 0.0474\n",
      "Function value obtained: -0.3918\n",
      "Current minimum: -0.4209\n",
      "Iteration No: 6 started. Evaluating function at random point.\n",
      "[26, 21, 10, 3, 'gini']\n",
      "\n",
      "roc_auc_score : 0.839281\n",
      "average_precision_score : 0.395319\n",
      "\n",
      "Iteration No: 6 ended. Evaluation done at random point.\n",
      "Time taken: 0.0440\n",
      "Function value obtained: -0.3953\n",
      "Current minimum: -0.4209\n",
      "Iteration No: 7 started. Evaluating function at random point.\n",
      "[16, 23, 7, 1, 'gini']\n",
      "\n",
      "roc_auc_score : 0.824515\n",
      "average_precision_score : 0.381719\n",
      "\n",
      "Iteration No: 7 ended. Evaluation done at random point.\n",
      "Time taken: 0.0633\n",
      "Function value obtained: -0.3817\n",
      "Current minimum: -0.4209\n",
      "Iteration No: 8 started. Evaluating function at random point.\n",
      "[20, 11, 5, 1, 'gini']\n",
      "\n",
      "roc_auc_score : 0.795054\n",
      "average_precision_score : 0.358044\n",
      "\n",
      "Iteration No: 8 ended. Evaluation done at random point.\n",
      "Time taken: 0.0394\n",
      "Function value obtained: -0.3580\n",
      "Current minimum: -0.4209\n",
      "Iteration No: 9 started. Evaluating function at random point.\n",
      "[5, 29, 2, 2, 'gini']\n",
      "\n",
      "roc_auc_score : 0.8629\n",
      "average_precision_score : 0.421951\n",
      "\n",
      "Iteration No: 9 ended. Evaluation done at random point.\n",
      "Time taken: 0.0318\n",
      "Function value obtained: -0.4220\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 10 started. Evaluating function at random point.\n",
      "[22, 6, 4, 3, 'entropy']\n",
      "\n",
      "roc_auc_score : 0.810737\n",
      "average_precision_score : 0.380455\n",
      "\n",
      "Iteration No: 10 ended. Evaluation done at random point.\n",
      "Time taken: 0.0330\n",
      "Function value obtained: -0.3805\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 11 started. Evaluating function at random point.\n",
      "[13, 12, 3, 2, 'gini']\n",
      "\n",
      "roc_auc_score : 0.824255\n",
      "average_precision_score : 0.393317\n",
      "\n",
      "Iteration No: 11 ended. Evaluation done at random point.\n",
      "Time taken: 0.0353\n",
      "Function value obtained: -0.3933\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 12 started. Evaluating function at random point.\n",
      "[19, 12, 4, 2, 'gini']\n",
      "\n",
      "roc_auc_score : 0.810041\n",
      "average_precision_score : 0.370413\n",
      "\n",
      "Iteration No: 12 ended. Evaluation done at random point.\n",
      "Time taken: 0.0303\n",
      "Function value obtained: -0.3704\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 13 started. Evaluating function at random point.\n",
      "[23, 14, 4, 2, 'gini']\n",
      "\n",
      "roc_auc_score : 0.815403\n",
      "average_precision_score : 0.380346\n",
      "\n",
      "Iteration No: 13 ended. Evaluation done at random point.\n",
      "Time taken: 0.0301\n",
      "Function value obtained: -0.3803\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 14 started. Evaluating function at random point.\n",
      "[19, 30, 2, 2, 'entropy']\n",
      "\n",
      "roc_auc_score : 0.835644\n",
      "average_precision_score : 0.397611\n",
      "\n",
      "Iteration No: 14 ended. Evaluation done at random point.\n",
      "Time taken: 0.0304\n",
      "Function value obtained: -0.3976\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 15 started. Evaluating function at random point.\n",
      "[15, 20, 8, 2, 'gini']\n",
      "\n",
      "roc_auc_score : 0.839095\n",
      "average_precision_score : 0.411439\n",
      "\n",
      "Iteration No: 15 ended. Evaluation done at random point.\n",
      "Time taken: 0.0277\n",
      "Function value obtained: -0.4114\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 16 started. Evaluating function at random point.\n",
      "[26, 12, 5, 1, 'gini']\n",
      "\n",
      "roc_auc_score : 0.833\n",
      "average_precision_score : 0.38841\n",
      "\n",
      "Iteration No: 16 ended. Evaluation done at random point.\n",
      "Time taken: 0.0282\n",
      "Function value obtained: -0.3884\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 17 started. Evaluating function at random point.\n",
      "[22, 28, 10, 1, 'gini']\n",
      "\n",
      "roc_auc_score : 0.839209\n",
      "average_precision_score : 0.391883\n",
      "\n",
      "Iteration No: 17 ended. Evaluation done at random point.\n",
      "Time taken: 0.0345\n",
      "Function value obtained: -0.3919\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 18 started. Evaluating function at random point.\n",
      "[29, 17, 9, 1, 'entropy']\n",
      "\n",
      "roc_auc_score : 0.840899\n",
      "average_precision_score : 0.390673\n",
      "\n",
      "Iteration No: 18 ended. Evaluation done at random point.\n",
      "Time taken: 0.0387\n",
      "Function value obtained: -0.3907\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 19 started. Evaluating function at random point.\n",
      "[2, 29, 7, 1, 'entropy']\n",
      "\n",
      "roc_auc_score : 0.698805\n",
      "average_precision_score : 0.230174\n",
      "\n",
      "Iteration No: 19 ended. Evaluation done at random point.\n",
      "Time taken: 0.0290\n",
      "Function value obtained: -0.2302\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 20 started. Evaluating function at random point.\n",
      "[13, 12, 8, 3, 'entropy']\n",
      "\n",
      "roc_auc_score : 0.829362\n",
      "average_precision_score : 0.390563\n",
      "\n",
      "Iteration No: 20 ended. Evaluation done at random point.\n",
      "Time taken: 0.5888\n",
      "Function value obtained: -0.3906\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 21 started. Searching for the next optimal point.\n",
      "[27, 30, 9, 1, 'entropy']\n",
      "\n",
      "roc_auc_score : 0.809309\n",
      "average_precision_score : 0.359988\n",
      "\n",
      "Iteration No: 21 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.5893\n",
      "Function value obtained: -0.3600\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 22 started. Searching for the next optimal point.\n",
      "[9, 30, 6, 2, 'entropy']\n",
      "\n",
      "roc_auc_score : 0.835716\n",
      "average_precision_score : 0.401412\n",
      "\n",
      "Iteration No: 22 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.5876\n",
      "Function value obtained: -0.4014\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 23 started. Searching for the next optimal point.\n",
      "[9, 30, 6, 2, 'entropy']\n",
      "\n",
      "roc_auc_score : 0.835716\n",
      "average_precision_score : 0.401412\n",
      "\n",
      "Iteration No: 23 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.9617\n",
      "Function value obtained: -0.4014\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 24 started. Searching for the next optimal point.\n",
      "[6, 29, 2, 1, 'gini']\n",
      "\n",
      "roc_auc_score : 0.849348\n",
      "average_precision_score : 0.395715\n",
      "\n",
      "Iteration No: 24 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.6006\n",
      "Function value obtained: -0.3957\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 25 started. Searching for the next optimal point.\n",
      "[7, 30, 5, 1, 'entropy']\n",
      "\n",
      "roc_auc_score : 0.842039\n",
      "average_precision_score : 0.386934\n",
      "\n",
      "Iteration No: 25 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.5786\n",
      "Function value obtained: -0.3869\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 26 started. Searching for the next optimal point.\n",
      "[4, 30, 2, 1, 'gini']\n",
      "\n",
      "roc_auc_score : 0.857429\n",
      "average_precision_score : 0.406295\n",
      "\n",
      "Iteration No: 26 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.7444\n",
      "Function value obtained: -0.4063\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 27 started. Searching for the next optimal point.\n",
      "[7, 24, 6, 1, 'entropy']\n",
      "\n",
      "roc_auc_score : 0.843251\n",
      "average_precision_score : 0.386424\n",
      "\n",
      "Iteration No: 27 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.6008\n",
      "Function value obtained: -0.3864\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 28 started. Searching for the next optimal point.\n",
      "[5, 25, 6, 1, 'entropy']\n",
      "\n",
      "roc_auc_score : 0.841895\n",
      "average_precision_score : 0.372023\n",
      "\n",
      "Iteration No: 28 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.6880\n",
      "Function value obtained: -0.3720\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 29 started. Searching for the next optimal point.\n",
      "[9, 30, 7, 1, 'entropy']\n",
      "\n",
      "roc_auc_score : 0.843288\n",
      "average_precision_score : 0.387955\n",
      "\n",
      "Iteration No: 29 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.6954\n",
      "Function value obtained: -0.3880\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 30 started. Searching for the next optimal point.\n",
      "[6, 29, 7, 1, 'entropy']\n",
      "\n",
      "roc_auc_score : 0.814714\n",
      "average_precision_score : 0.344314\n",
      "\n",
      "Iteration No: 30 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.6214\n",
      "Function value obtained: -0.3443\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 31 started. Searching for the next optimal point.\n",
      "[2, 29, 7, 2, 'entropy']\n",
      "\n",
      "roc_auc_score : 0.845206\n",
      "average_precision_score : 0.361099\n",
      "\n",
      "Iteration No: 31 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.8015\n",
      "Function value obtained: -0.3611\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 32 started. Searching for the next optimal point.\n",
      "[2, 18, 7, 1, 'entropy']\n",
      "\n",
      "roc_auc_score : 0.698805\n",
      "average_precision_score : 0.230174\n",
      "\n",
      "Iteration No: 32 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.8140\n",
      "Function value obtained: -0.2302\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 33 started. Searching for the next optimal point.\n",
      "[2, 28, 6, 2, 'gini']\n",
      "\n",
      "roc_auc_score : 0.866872\n",
      "average_precision_score : 0.40124\n",
      "\n",
      "Iteration No: 33 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.7003\n",
      "Function value obtained: -0.4012\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 34 started. Searching for the next optimal point.\n",
      "[2, 30, 7, 1, 'gini']\n",
      "\n",
      "roc_auc_score : 0.720472\n",
      "average_precision_score : 0.247776\n",
      "\n",
      "Iteration No: 34 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.7309\n",
      "Function value obtained: -0.2478\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 35 started. Searching for the next optimal point.\n",
      "[2, 29, 8, 2, 'gini']\n",
      "\n",
      "roc_auc_score : 0.866872\n",
      "average_precision_score : 0.40124\n",
      "\n",
      "Iteration No: 35 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.7722\n",
      "Function value obtained: -0.4012\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 36 started. Searching for the next optimal point.\n",
      "[3, 30, 2, 1, 'gini']\n",
      "\n",
      "roc_auc_score : 0.796276\n",
      "average_precision_score : 0.324743\n",
      "\n",
      "Iteration No: 36 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.6308\n",
      "Function value obtained: -0.3247\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 37 started. Searching for the next optimal point.\n",
      "[6, 30, 7, 3, 'gini']\n",
      "\n",
      "roc_auc_score : 0.845378\n",
      "average_precision_score : 0.405388\n",
      "\n",
      "Iteration No: 37 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.7665\n",
      "Function value obtained: -0.4054\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 38 started. Searching for the next optimal point.\n",
      "[5, 26, 9, 1, 'gini']\n",
      "\n",
      "roc_auc_score : 0.853645\n",
      "average_precision_score : 0.40151\n",
      "\n",
      "Iteration No: 38 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.7184\n",
      "Function value obtained: -0.4015\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 39 started. Searching for the next optimal point.\n",
      "[4, 28, 8, 1, 'gini']\n",
      "\n",
      "roc_auc_score : 0.857429\n",
      "average_precision_score : 0.406295\n",
      "\n",
      "Iteration No: 39 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.7346\n",
      "Function value obtained: -0.4063\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 40 started. Searching for the next optimal point.\n",
      "[3, 27, 9, 1, 'gini']\n",
      "\n",
      "roc_auc_score : 0.797231\n",
      "average_precision_score : 0.324368\n",
      "\n",
      "Iteration No: 40 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.7540\n",
      "Function value obtained: -0.3244\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 41 started. Searching for the next optimal point.\n",
      "[4, 29, 7, 1, 'gini']\n",
      "\n",
      "roc_auc_score : 0.857429\n",
      "average_precision_score : 0.406295\n",
      "\n",
      "Iteration No: 41 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.6512\n",
      "Function value obtained: -0.4063\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 42 started. Searching for the next optimal point.\n",
      "[4, 30, 7, 1, 'gini']\n",
      "\n",
      "roc_auc_score : 0.857429\n",
      "average_precision_score : 0.406295\n",
      "\n",
      "Iteration No: 42 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.7042\n",
      "Function value obtained: -0.4063\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 43 started. Searching for the next optimal point.\n",
      "[4, 29, 7, 1, 'gini']\n",
      "\n",
      "roc_auc_score : 0.857429\n",
      "average_precision_score : 0.406295\n",
      "\n",
      "Iteration No: 43 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.6275\n",
      "Function value obtained: -0.4063\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 44 started. Searching for the next optimal point.\n",
      "[4, 30, 7, 1, 'gini']\n",
      "\n",
      "roc_auc_score : 0.857429\n",
      "average_precision_score : 0.406295\n",
      "\n",
      "Iteration No: 44 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.7085\n",
      "Function value obtained: -0.4063\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 45 started. Searching for the next optimal point.\n",
      "[2, 29, 7, 1, 'gini']\n",
      "\n",
      "roc_auc_score : 0.720472\n",
      "average_precision_score : 0.247776\n",
      "\n",
      "Iteration No: 45 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.6308\n",
      "Function value obtained: -0.2478\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 46 started. Searching for the next optimal point.\n",
      "[6, 30, 7, 1, 'gini']\n",
      "\n",
      "roc_auc_score : 0.858127\n",
      "average_precision_score : 0.404557\n",
      "\n",
      "Iteration No: 46 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.6345\n",
      "Function value obtained: -0.4046\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 47 started. Searching for the next optimal point.\n",
      "[6, 29, 7, 1, 'gini']\n",
      "\n",
      "roc_auc_score : 0.858127\n",
      "average_precision_score : 0.404557\n",
      "\n",
      "Iteration No: 47 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.7213\n",
      "Function value obtained: -0.4046\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 48 started. Searching for the next optimal point.\n",
      "[2, 22, 7, 2, 'entropy']\n",
      "\n",
      "roc_auc_score : 0.845206\n",
      "average_precision_score : 0.361099\n",
      "\n",
      "Iteration No: 48 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.6475\n",
      "Function value obtained: -0.3611\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 49 started. Searching for the next optimal point.\n",
      "[4, 27, 7, 1, 'gini']\n",
      "\n",
      "roc_auc_score : 0.857429\n",
      "average_precision_score : 0.406295\n",
      "\n",
      "Iteration No: 49 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.6862\n",
      "Function value obtained: -0.4063\n",
      "Current minimum: -0.4220\n",
      "Iteration No: 50 started. Searching for the next optimal point.\n",
      "[4, 29, 4, 1, 'gini']\n",
      "\n",
      "roc_auc_score : 0.857429\n",
      "average_precision_score : 0.406295\n",
      "\n",
      "Iteration No: 50 ended. Search finished for the next optimal point.\n",
      "Time taken: 0.6030\n",
      "Function value obtained: -0.4063\n",
      "Current minimum: -0.4220\n"
     ]
    }
   ],
   "source": [
    "def tune_dtree(params):\n",
    "    print(params)\n",
    "    max_depth = params[0]\n",
    "    min_samples_split = params[1]\n",
    "    min_samples_leaf = params[2]\n",
    "    max_features = params[3]\n",
    "    criterion = params[4]\n",
    "    \n",
    "    \n",
    "    model = DecisionTreeClassifier(max_depth = max_depth,\n",
    "                                  min_samples_split = min_samples_split,\n",
    "                                  min_samples_leaf = min_samples_leaf,\n",
    "                                  max_features = max_features,\n",
    "                                  criterion = criterion,\n",
    "                                  random_state = 42,\n",
    "                                  class_weight = 'balanced')\n",
    "    \n",
    "    \n",
    "    model.fit(Xtrain,y_train)\n",
    "    \n",
    "    pred = model.predict(Xval)\n",
    "    \n",
    "    print()\n",
    "    print('roc_auc_score :', np.round(roc_auc_score(y_val, pred),6))\n",
    "    print('average_precision_score :', np.round(average_precision_score(y_val, pred),6))\n",
    "    print()\n",
    "    \n",
    "    \n",
    "    return - average_precision_score(y_val, pred)\n",
    "\n",
    "space = [(2,30), #max_depth\n",
    "         (5,30), # min_samples_split\n",
    "         (1,10), # min_sample_leaf\n",
    "         (1,3), # max_features\n",
    "         ('gini','entropy')] # criterion \n",
    "         \n",
    "\n",
    "res = forest_minimize(tune_dtree,\n",
    "                     space,\n",
    "                     random_state = 42,\n",
    "                     n_random_starts = 20,\n",
    "                     n_calls = 50,\n",
    "                     verbose = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[5, 29, 2, 2, 'gini']"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "res.x\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "roc_auc_score : 0.8629\n",
      "average_precision_score : 0.421951\n",
      "\n"
     ]
    }
   ],
   "source": [
    "params = res.x\n",
    "max_depth = params[0]\n",
    "min_samples_split = params[1]\n",
    "min_samples_leaf = params[2]\n",
    "max_features = params[3]\n",
    "criterion = params[4]\n",
    "\n",
    "\n",
    "model_dtree = DecisionTreeClassifier(max_depth = max_depth,\n",
    "                              min_samples_split = min_samples_split,\n",
    "                              min_samples_leaf = min_samples_leaf,\n",
    "                              max_features = max_features,\n",
    "                              criterion = criterion,\n",
    "                              random_state = 42,\n",
    "                              class_weight = 'balanced')\n",
    "\n",
    "model_dtree.fit(Xtrain,y_train)\n",
    "\n",
    "pred_dtree = model_dtree.predict(Xval)\n",
    "\n",
    "print()\n",
    "print('roc_auc_score :', np.round(roc_auc_score(y_val, pred_dtree),6))\n",
    "print('average_precision_score :', np.round(average_precision_score(y_val, pred_dtree),6))\n",
    "print()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}