1 - Hiperparâmetros, overfitting e otimização.ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h1>Conceito</h1>\n",
    "\n",
    "<h2>Hiperparâmetros</h2>\n",
    "\n",
    "Hiperparâmetros são parâmetros que não são diretamente 'aprendidos' pelos estimadores. Na biblioteca *SKLearn* por exemplo, eles são passados como argumentos para o construtor da classe de estimador.\n",
    "\n",
    "Exemplos de hiperparâmetros são:\n",
    "- C, kernell e gamma para classificadores de vetores de suporte\n",
    "- Profundidade máxima para árvores de decisão\n",
    "- Número de *epochs*\n",
    "\n",
    "<h2><i>Overfitting</i></h2>\n",
    "\n",
    "O *overfitting* acontece quando um modelo se ajusta muito bem ao conjunto de dados anteriormente observados, mas é ineficaz para prever novos resultados. Isso pode acontecer por diferentes motivos, desde uma base de dados de treino 'enviesada' até valores de hiperparâmetros.\n",
    "\n",
    "Um exemplo disso é o caso em que se aumenta indefinidamente a profundidade máxima de uma árvore de decisão: intuitivamente, imagina-se que o modelo irá se tornar mais eficaz. Porém, a partir de uma certa profundidade, ele irá se tornar mais eficaz apenas em prever os dados de treino, mas não em prever novos dados."
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h1>Aplicação</h1>"
   ]
  },
  {
   "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 cross_validate\n",
    "from sklearn.model_selection import KFold\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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>preco</th>\n",
       "      <th>vendido</th>\n",
       "      <th>idade_do_modelo</th>\n",
       "      <th>km_por_ano</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>30941.02</td>\n",
       "      <td>1</td>\n",
       "      <td>18</td>\n",
       "      <td>35085.22134</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>40557.96</td>\n",
       "      <td>1</td>\n",
       "      <td>20</td>\n",
       "      <td>12622.05362</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>89627.50</td>\n",
       "      <td>0</td>\n",
       "      <td>12</td>\n",
       "      <td>11440.79806</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>95276.14</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>43167.32682</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>117384.68</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>12770.11290</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       preco  vendido  idade_do_modelo   km_por_ano\n",
       "0   30941.02        1               18  35085.22134\n",
       "1   40557.96        1               20  12622.05362\n",
       "2   89627.50        0               12  11440.79806\n",
       "3   95276.14        0                3  43167.32682\n",
       "4  117384.68        1                4  12770.11290"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Importando a base de dados\n",
    "dados = pd.read_csv(r'Dados\\base.csv')\n",
    "dados.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Dividindo inputs e outputs\n",
    "x = dados[['preco', 'idade_do_modelo', 'km_por_ano']].values\n",
    "y = dados['vendido'].values.ravel()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Criando função para treinar n árvores de decisão com diferentes max_depths e retornar seus resultados em um dataframe\n",
    "\n",
    "def dtc_maxdepth_cv(maxdepth, x, y, SEED=20):\n",
    "\n",
    "    columns = [\n",
    "        ['', 'Teste', 'Teste', 'Teste', 'Treino', 'Treino', 'Treino'],\n",
    "        ['Max depth', 'Acurácia mínima', 'Acurácia média', 'Acurácia máxima', 'Acurácia mínima', 'Acurácia média', 'Acurácia máxima']\n",
    "    ]\n",
    "\n",
    "    df_resultados = pd.DataFrame(columns=columns)\n",
    "\n",
    "    for i in range(maxdepth):\n",
    "\n",
    "        np.random.seed(SEED)\n",
    "\n",
    "        resultados = cross_validate(\n",
    "            DecisionTreeClassifier(max_depth=i+1),\n",
    "            x,\n",
    "            y,\n",
    "            cv = KFold(n_splits=10, shuffle=True),\n",
    "            return_train_score=True\n",
    "        )\n",
    "\n",
    "        acc_min_test = (resultados['test_score'].mean() - resultados['test_score'].std()) * 100\n",
    "        acc_mean_test = resultados['test_score'].mean() * 100\n",
    "        acc_max_test = (resultados['test_score'].mean() + resultados['test_score'].std()) * 100\n",
    "        acc_min_train = (resultados['train_score'].mean() - resultados['train_score'].std()) * 100\n",
    "        acc_mean_train = resultados['train_score'].mean() * 100\n",
    "        acc_max_train = (resultados['train_score'].mean() + resultados['train_score'].std()) * 100\n",
    "\n",
    "        df_resultados = pd.concat([df_resultados,\n",
    "                                   pd.DataFrame(data=[[i+1, acc_min_test, acc_mean_test, acc_max_test, acc_min_train, acc_mean_train, acc_max_train]], columns=columns)],\n",
    "                                   axis = 0,\n",
    "                                   ignore_index=True)\n",
    "    \n",
    "    return df_resultados"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "teste_maxdepth = dtc_maxdepth_cv(30, x, y, SEED=20)"
   ]
  },
  {
   "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 tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th colspan=\"3\" halign=\"left\">Teste</th>\n",
       "      <th colspan=\"3\" halign=\"left\">Treino</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>Max depth</th>\n",
       "      <th>Acurácia mínima</th>\n",
       "      <th>Acurácia média</th>\n",
       "      <th>Acurácia máxima</th>\n",
       "      <th>Acurácia mínima</th>\n",
       "      <th>Acurácia média</th>\n",
       "      <th>Acurácia máxima</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>74.568555</td>\n",
       "      <td>75.78</td>\n",
       "      <td>76.991445</td>\n",
       "      <td>75.656506</td>\n",
       "      <td>75.791111</td>\n",
       "      <td>75.925716</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>74.568555</td>\n",
       "      <td>75.78</td>\n",
       "      <td>76.991445</td>\n",
       "      <td>75.656506</td>\n",
       "      <td>75.791111</td>\n",
       "      <td>75.925716</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>77.578739</td>\n",
       "      <td>78.69</td>\n",
       "      <td>79.801261</td>\n",
       "      <td>78.629579</td>\n",
       "      <td>78.754444</td>\n",
       "      <td>78.879310</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>77.483972</td>\n",
       "      <td>78.64</td>\n",
       "      <td>79.796028</td>\n",
       "      <td>78.663736</td>\n",
       "      <td>78.800000</td>\n",
       "      <td>78.936264</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>77.455584</td>\n",
       "      <td>78.59</td>\n",
       "      <td>79.724416</td>\n",
       "      <td>78.759007</td>\n",
       "      <td>78.911111</td>\n",
       "      <td>79.063216</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>6</td>\n",
       "      <td>77.257348</td>\n",
       "      <td>78.39</td>\n",
       "      <td>79.522652</td>\n",
       "      <td>79.031479</td>\n",
       "      <td>79.172222</td>\n",
       "      <td>79.312965</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>7</td>\n",
       "      <td>76.817146</td>\n",
       "      <td>78.15</td>\n",
       "      <td>79.482854</td>\n",
       "      <td>79.223534</td>\n",
       "      <td>79.411111</td>\n",
       "      <td>79.598688</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>8</td>\n",
       "      <td>76.621790</td>\n",
       "      <td>77.85</td>\n",
       "      <td>79.078210</td>\n",
       "      <td>79.575908</td>\n",
       "      <td>79.876667</td>\n",
       "      <td>80.177425</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>9</td>\n",
       "      <td>76.650703</td>\n",
       "      <td>77.86</td>\n",
       "      <td>79.069297</td>\n",
       "      <td>80.015702</td>\n",
       "      <td>80.353333</td>\n",
       "      <td>80.690965</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>10</td>\n",
       "      <td>76.445928</td>\n",
       "      <td>77.59</td>\n",
       "      <td>78.734072</td>\n",
       "      <td>80.534383</td>\n",
       "      <td>80.987778</td>\n",
       "      <td>81.441172</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>11</td>\n",
       "      <td>75.884778</td>\n",
       "      <td>77.18</td>\n",
       "      <td>78.475222</td>\n",
       "      <td>81.218543</td>\n",
       "      <td>81.798889</td>\n",
       "      <td>82.379235</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>12</td>\n",
       "      <td>75.205467</td>\n",
       "      <td>76.41</td>\n",
       "      <td>77.614533</td>\n",
       "      <td>82.023740</td>\n",
       "      <td>82.662222</td>\n",
       "      <td>83.300704</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>13</td>\n",
       "      <td>75.159840</td>\n",
       "      <td>76.41</td>\n",
       "      <td>77.660160</td>\n",
       "      <td>82.847400</td>\n",
       "      <td>83.604444</td>\n",
       "      <td>84.361489</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>14</td>\n",
       "      <td>74.116032</td>\n",
       "      <td>75.63</td>\n",
       "      <td>77.143968</td>\n",
       "      <td>83.744518</td>\n",
       "      <td>84.684444</td>\n",
       "      <td>85.624371</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>15</td>\n",
       "      <td>73.837180</td>\n",
       "      <td>75.47</td>\n",
       "      <td>77.102820</td>\n",
       "      <td>84.775828</td>\n",
       "      <td>85.798889</td>\n",
       "      <td>86.821950</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>16</td>\n",
       "      <td>73.529497</td>\n",
       "      <td>75.02</td>\n",
       "      <td>76.510503</td>\n",
       "      <td>85.741778</td>\n",
       "      <td>86.903333</td>\n",
       "      <td>88.064889</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>17</td>\n",
       "      <td>73.467652</td>\n",
       "      <td>74.64</td>\n",
       "      <td>75.812348</td>\n",
       "      <td>86.718605</td>\n",
       "      <td>88.032222</td>\n",
       "      <td>89.345840</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>18</td>\n",
       "      <td>73.179187</td>\n",
       "      <td>74.41</td>\n",
       "      <td>75.640813</td>\n",
       "      <td>87.774018</td>\n",
       "      <td>89.071111</td>\n",
       "      <td>90.368204</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>19</td>\n",
       "      <td>72.634752</td>\n",
       "      <td>74.20</td>\n",
       "      <td>75.765248</td>\n",
       "      <td>88.733943</td>\n",
       "      <td>90.058889</td>\n",
       "      <td>91.383835</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>20</td>\n",
       "      <td>72.614739</td>\n",
       "      <td>74.10</td>\n",
       "      <td>75.585261</td>\n",
       "      <td>89.689715</td>\n",
       "      <td>91.043333</td>\n",
       "      <td>92.396952</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>21</td>\n",
       "      <td>72.521074</td>\n",
       "      <td>74.01</td>\n",
       "      <td>75.498926</td>\n",
       "      <td>90.413043</td>\n",
       "      <td>91.807778</td>\n",
       "      <td>93.202513</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>22</td>\n",
       "      <td>72.055519</td>\n",
       "      <td>73.62</td>\n",
       "      <td>75.184481</td>\n",
       "      <td>91.266240</td>\n",
       "      <td>92.702222</td>\n",
       "      <td>94.138204</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>23</td>\n",
       "      <td>72.052130</td>\n",
       "      <td>73.53</td>\n",
       "      <td>75.007870</td>\n",
       "      <td>92.039281</td>\n",
       "      <td>93.446667</td>\n",
       "      <td>94.854053</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>24</td>\n",
       "      <td>72.094302</td>\n",
       "      <td>73.08</td>\n",
       "      <td>74.065698</td>\n",
       "      <td>92.703975</td>\n",
       "      <td>94.151111</td>\n",
       "      <td>95.598247</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>25</td>\n",
       "      <td>71.736408</td>\n",
       "      <td>73.13</td>\n",
       "      <td>74.523592</td>\n",
       "      <td>93.472444</td>\n",
       "      <td>94.850000</td>\n",
       "      <td>96.227556</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>26</td>\n",
       "      <td>71.555360</td>\n",
       "      <td>72.98</td>\n",
       "      <td>74.404640</td>\n",
       "      <td>94.086597</td>\n",
       "      <td>95.472222</td>\n",
       "      <td>96.857847</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>27</td>\n",
       "      <td>71.844848</td>\n",
       "      <td>73.08</td>\n",
       "      <td>74.315152</td>\n",
       "      <td>94.744356</td>\n",
       "      <td>96.060000</td>\n",
       "      <td>97.375644</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>28</td>\n",
       "      <td>71.468852</td>\n",
       "      <td>72.82</td>\n",
       "      <td>74.171148</td>\n",
       "      <td>95.256641</td>\n",
       "      <td>96.557778</td>\n",
       "      <td>97.858915</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>29</td>\n",
       "      <td>71.329522</td>\n",
       "      <td>72.48</td>\n",
       "      <td>73.630478</td>\n",
       "      <td>95.772329</td>\n",
       "      <td>97.002222</td>\n",
       "      <td>98.232115</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>30</td>\n",
       "      <td>70.966397</td>\n",
       "      <td>72.37</td>\n",
       "      <td>73.773603</td>\n",
       "      <td>96.271418</td>\n",
       "      <td>97.452222</td>\n",
       "      <td>98.633026</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                       Teste                                         Treino   \n",
       "   Max depth Acurácia mínima Acurácia média Acurácia máxima Acurácia mínima   \n",
       "0          1       74.568555          75.78       76.991445       75.656506  \\\n",
       "1          2       74.568555          75.78       76.991445       75.656506   \n",
       "2          3       77.578739          78.69       79.801261       78.629579   \n",
       "3          4       77.483972          78.64       79.796028       78.663736   \n",
       "4          5       77.455584          78.59       79.724416       78.759007   \n",
       "5          6       77.257348          78.39       79.522652       79.031479   \n",
       "6          7       76.817146          78.15       79.482854       79.223534   \n",
       "7          8       76.621790          77.85       79.078210       79.575908   \n",
       "8          9       76.650703          77.86       79.069297       80.015702   \n",
       "9         10       76.445928          77.59       78.734072       80.534383   \n",
       "10        11       75.884778          77.18       78.475222       81.218543   \n",
       "11        12       75.205467          76.41       77.614533       82.023740   \n",
       "12        13       75.159840          76.41       77.660160       82.847400   \n",
       "13        14       74.116032          75.63       77.143968       83.744518   \n",
       "14        15       73.837180          75.47       77.102820       84.775828   \n",
       "15        16       73.529497          75.02       76.510503       85.741778   \n",
       "16        17       73.467652          74.64       75.812348       86.718605   \n",
       "17        18       73.179187          74.41       75.640813       87.774018   \n",
       "18        19       72.634752          74.20       75.765248       88.733943   \n",
       "19        20       72.614739          74.10       75.585261       89.689715   \n",
       "20        21       72.521074          74.01       75.498926       90.413043   \n",
       "21        22       72.055519          73.62       75.184481       91.266240   \n",
       "22        23       72.052130          73.53       75.007870       92.039281   \n",
       "23        24       72.094302          73.08       74.065698       92.703975   \n",
       "24        25       71.736408          73.13       74.523592       93.472444   \n",
       "25        26       71.555360          72.98       74.404640       94.086597   \n",
       "26        27       71.844848          73.08       74.315152       94.744356   \n",
       "27        28       71.468852          72.82       74.171148       95.256641   \n",
       "28        29       71.329522          72.48       73.630478       95.772329   \n",
       "29        30       70.966397          72.37       73.773603       96.271418   \n",
       "\n",
       "                                   \n",
       "   Acurácia média Acurácia máxima  \n",
       "0       75.791111       75.925716  \n",
       "1       75.791111       75.925716  \n",
       "2       78.754444       78.879310  \n",
       "3       78.800000       78.936264  \n",
       "4       78.911111       79.063216  \n",
       "5       79.172222       79.312965  \n",
       "6       79.411111       79.598688  \n",
       "7       79.876667       80.177425  \n",
       "8       80.353333       80.690965  \n",
       "9       80.987778       81.441172  \n",
       "10      81.798889       82.379235  \n",
       "11      82.662222       83.300704  \n",
       "12      83.604444       84.361489  \n",
       "13      84.684444       85.624371  \n",
       "14      85.798889       86.821950  \n",
       "15      86.903333       88.064889  \n",
       "16      88.032222       89.345840  \n",
       "17      89.071111       90.368204  \n",
       "18      90.058889       91.383835  \n",
       "19      91.043333       92.396952  \n",
       "20      91.807778       93.202513  \n",
       "21      92.702222       94.138204  \n",
       "22      93.446667       94.854053  \n",
       "23      94.151111       95.598247  \n",
       "24      94.850000       96.227556  \n",
       "25      95.472222       96.857847  \n",
       "26      96.060000       97.375644  \n",
       "27      96.557778       97.858915  \n",
       "28      97.002222       98.232115  \n",
       "29      97.452222       98.633026  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "teste_maxdepth"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.lineplot(x = ('', 'Max depth'), y = ('Treino', 'Acurácia média'), data = teste_maxdepth, label = 'Treino').set(xlabel =\"Max depth\", ylabel = \"Acurácia média\")\n",
    "sns.lineplot(x = ('', 'Max depth'), y = ('Teste', 'Acurácia média'), data = teste_maxdepth, label = 'Teste').set(xlabel =\"Max depth\", ylabel = \"Acurácia média\")\n",
    "\n",
    "plt.legend()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Como foi possível ver no exemplo acima, **a partir de uma certa profundidade o modelo aumenta apenas sua acurácia de treino e acaba até perdendo acurácia de teste**"
   ]
  }
 ],
 "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": "3.11.0"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}