machineLearning.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "523b18e0-045d-4d85-b2a3-b6efbea2b130",
   "metadata": {
    "tags": []
   },
   "source": [
    "### Aplicando Machine Learning sobre os dados\n",
    " - Carregando as bibliotecas para a nova etapa"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "d95dd979-0aad-4b04-9b2b-895ddea65ec2",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "1ba239c6-7bd9-4742-abb6-bd62e8e6cb48",
   "metadata": {},
   "outputs": [],
   "source": [
    "import warnings\n",
    "warnings.filterwarnings('ignore')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "ff451294-55c0-431a-9034-48ee0b56f654",
   "metadata": {},
   "outputs": [],
   "source": [
    "sp_houses_df = pd.read_csv(\"./dados/SaoPauloClean.csv\").drop(\"Unnamed: 0\", axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "85876b6a-0750-48e8-997e-313ba5727587",
   "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>area_metro_quadrado_int</th>\n",
       "      <th>quartos_int</th>\n",
       "      <th>banheiros_int</th>\n",
       "      <th>vagas_int</th>\n",
       "      <th>preco_anuncio_float</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1000</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "      <td>7000000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>469</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>3700000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>125</td>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>380000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>310</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>685000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>100</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>540000.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   area_metro_quadrado_int  quartos_int  banheiros_int  vagas_int  \\\n",
       "0                     1000            4              8          6   \n",
       "1                      469            3              3          4   \n",
       "2                      125            4              3          2   \n",
       "3                      310            3              2          4   \n",
       "4                      100            3              2          2   \n",
       "\n",
       "   preco_anuncio_float  \n",
       "0            7000000.0  \n",
       "1            3700000.0  \n",
       "2             380000.0  \n",
       "3             685000.0  \n",
       "4             540000.0  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sp_houses_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5f5f9928-8e45-455c-ab31-1281e43e04b8",
   "metadata": {},
   "source": [
    "## 5 - Produzindo o modelo de Regressão"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0c61f7b3-f127-4c5a-9a76-703b187e6d01",
   "metadata": {},
   "source": [
    "### 5.1 - Analisando a matriz de correlação"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "d6840b19-6372-4e59-ba59-8b53a947d2d7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import seaborn as sns; sns.set_theme()\n",
    "ax = sns.heatmap(sp_houses_df.corr(method=\"pearson\") ,cmap=\"YlGnBu\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fcc89fe4-229c-4f3f-8626-548c4f8b16e1",
   "metadata": {},
   "source": [
    "- A respeito da seleção de features para uma regressão linear, é importante que as variáveis explicativas sejam colunas com pouca correlação entre si e que as variáveis explicativas e que tenham uma correlação significativa com a variável explicativa.\n",
    "- Em se tratando do presente trabalho, as variáveis de área em m², numero de quartos, numero de banheiros e vagas de estacionamento serão escolhidas como explicativas.\n",
    "- A variável de controle será o preço anunciado do imóvel."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "49ed7575-3d24-4272-b3f7-00905522db2c",
   "metadata": {},
   "source": [
    "### 5.2 - Analisando o histograma da variável de controle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "1f6700e9-fafc-4c72-b8ff-41bdb506acfc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "sns.set()\n",
    "plt.figure(figsize=(18, 10))\n",
    "ax = sns.histplot(data = sp_houses_df, x = \"preco_anuncio_float\", kde = True)\n",
    "ax.set_title(\"Histograma Valor Imóvel\", fontsize = 22)\n",
    "ax.set_xlabel(\"Preço imovél (R$)\", fontsize=18)\n",
    "ax.set_ylabel(\"Total de imóveis\", fontsize=18)\n",
    "plt.xlim((-50, 10000000))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "960f4e23-c45c-4094-9802-c8bef412339c",
   "metadata": {},
   "source": [
    "- Devido a assimetria que concentra o preço dos imóveis nas primeiras barras do histograma, as variáveis receberão uma transformação logarítmica para tentar normalizar os resultados e fazer previsões com maior precisão. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "2c2b07d3-ec2e-489e-b13b-fb7e7026a36b",
   "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>area_metro_quadrado_int</th>\n",
       "      <th>quartos_int</th>\n",
       "      <th>banheiros_int</th>\n",
       "      <th>vagas_int</th>\n",
       "      <th>preco_anuncio_float</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>6.908755</td>\n",
       "      <td>1.609438</td>\n",
       "      <td>2.197225</td>\n",
       "      <td>1.945910</td>\n",
       "      <td>15.761421</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>6.152733</td>\n",
       "      <td>1.386294</td>\n",
       "      <td>1.386294</td>\n",
       "      <td>1.609438</td>\n",
       "      <td>15.123844</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4.836282</td>\n",
       "      <td>1.609438</td>\n",
       "      <td>1.386294</td>\n",
       "      <td>1.098612</td>\n",
       "      <td>12.847929</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>5.739793</td>\n",
       "      <td>1.386294</td>\n",
       "      <td>1.098612</td>\n",
       "      <td>1.609438</td>\n",
       "      <td>13.437176</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4.615121</td>\n",
       "      <td>1.386294</td>\n",
       "      <td>1.098612</td>\n",
       "      <td>1.098612</td>\n",
       "      <td>13.199326</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   area_metro_quadrado_int  quartos_int  banheiros_int  vagas_int  \\\n",
       "0                 6.908755     1.609438       2.197225   1.945910   \n",
       "1                 6.152733     1.386294       1.386294   1.609438   \n",
       "2                 4.836282     1.609438       1.386294   1.098612   \n",
       "3                 5.739793     1.386294       1.098612   1.609438   \n",
       "4                 4.615121     1.386294       1.098612   1.098612   \n",
       "\n",
       "   preco_anuncio_float  \n",
       "0            15.761421  \n",
       "1            15.123844  \n",
       "2            12.847929  \n",
       "3            13.437176  \n",
       "4            13.199326  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.preprocessing import FunctionTransformer\n",
    "\n",
    "transformer = FunctionTransformer(np. log1p, validate = True)\n",
    "sp_houses_df_transformado = transformer.transform(sp_houses_df)\n",
    "colunas_sp_house_df = sp_houses_df.columns\n",
    "sp_houses_df_transformado = pd.DataFrame(sp_houses_df_transformado, columns=colunas_sp_house_df)\n",
    "sp_houses_df_transformado.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "8e82130f-5630-4a5b-8322-4d0b425f94cb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set()\n",
    "plt.figure(figsize=(18, 10))\n",
    "ax = sns.histplot(data = sp_houses_df_transformado, x = \"preco_anuncio_float\", kde = True)\n",
    "ax.set_title(\"Histograma Valor Imóvel Transformado\", fontsize = 22)\n",
    "ax.set_xlabel(\"Preço imovél (R$)\", fontsize=18)\n",
    "ax.set_ylabel(\"Total de imóveis\", fontsize=18)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "814cafe2-b768-4752-b5f2-c7a8691bfbc1",
   "metadata": {},
   "source": [
    "- Após a transformação logarítmica, a distribuição está mais próxima de uma distribuição simétrica.\n",
    "- Com uma distribuição mais simétrica, as previsões serão feitas com maior precisão."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "45a133f0-608d-4dc4-8d48-37c9e4079630",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ax = sns.heatmap(sp_houses_df_transformado.corr(method=\"pearson\") ,cmap=\"YlGnBu\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ca11d836-a8bc-450b-98ea-120c41fc5bd2",
   "metadata": {},
   "source": [
    "- Com a normalização também fica evidente a alta correlação entre preço do imóvel e a sua quantidade de metros quadrados disponível."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "bfb430e9-185d-4ad0-a0d4-7068824c324d",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = sp_houses_df_transformado[['area_metro_quadrado_int','quartos_int', 'banheiros_int', 'vagas_int']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "aa98c586-4e67-4224-86c4-6fe4a51c0a00",
   "metadata": {},
   "outputs": [],
   "source": [
    "y = sp_houses_df_transformado['preco_anuncio_float']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "cc1fd482-cea4-4c31-a8cb-3c08e4b95078",
   "metadata": {},
   "outputs": [],
   "source": [
    "resultados_modelos_df = pd.DataFrame()\n",
    "resultados_modelos_dict = {\n",
    "    \"modelo\": [],\n",
    "    \"R2\": [],\n",
    "    \"MSE\": [],\n",
    "    \"MSA\": [],\n",
    "    \"media_cv\": [],\n",
    "    \"intervalo_cv\": []\n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "38e3b199-49de-4eca-98e8-f6ffaf9523a8",
   "metadata": {},
   "source": [
    "### 5.3 - Separação entre dados de treino e dados de teste\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "2991a427-1190-4b93-b83a-d5ca885dfeb7",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "9d3a14e7-8445-4eed-8772-ea1ade98c4fc",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "b1cd2306-9fe1-4ddd-80bb-fc6298826688",
   "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>area_metro_quadrado_int</th>\n",
       "      <th>quartos_int</th>\n",
       "      <th>banheiros_int</th>\n",
       "      <th>vagas_int</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>6304</th>\n",
       "      <td>5.564520</td>\n",
       "      <td>1.609438</td>\n",
       "      <td>1.386294</td>\n",
       "      <td>1.386294</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5459</th>\n",
       "      <td>5.602119</td>\n",
       "      <td>1.386294</td>\n",
       "      <td>1.791759</td>\n",
       "      <td>1.791759</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2300</th>\n",
       "      <td>6.115892</td>\n",
       "      <td>1.386294</td>\n",
       "      <td>1.609438</td>\n",
       "      <td>1.098612</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5851</th>\n",
       "      <td>4.753590</td>\n",
       "      <td>1.098612</td>\n",
       "      <td>1.386294</td>\n",
       "      <td>1.098612</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>254</th>\n",
       "      <td>4.624973</td>\n",
       "      <td>1.386294</td>\n",
       "      <td>1.609438</td>\n",
       "      <td>1.098612</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      area_metro_quadrado_int  quartos_int  banheiros_int  vagas_int\n",
       "6304                 5.564520     1.609438       1.386294   1.386294\n",
       "5459                 5.602119     1.386294       1.791759   1.791759\n",
       "2300                 6.115892     1.386294       1.609438   1.098612\n",
       "5851                 4.753590     1.098612       1.386294   1.098612\n",
       "254                  4.624973     1.386294       1.609438   1.098612"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3f0aefb9-5e19-462a-a082-53e6867dc1c1",
   "metadata": {},
   "source": [
    "### 5.4 - Criando os modelos de Regressão"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dc6b6896-a677-450a-9bba-3aa527005147",
   "metadata": {},
   "source": [
    "#### a) Dummy Regressor\n",
    "- A ideia desse modelo será funcionar como um baseline para as análises, e à medida que outro modelos forem feitos, o baseline será atualizado."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "09efed94-b6cd-45f8-97c9-082c49908b62",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import r2_score\n",
    "from sklearn.metrics import mean_squared_error, mean_absolute_error\n",
    "from sklearn.model_selection import cross_validate\n",
    "from sklearn.model_selection import KFold"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "70cde1e2-f6ef-47aa-a409-3926c2647d3e",
   "metadata": {},
   "outputs": [],
   "source": [
    "def cv_stats (resultado_cv):\n",
    "    media = resultado_cv['test_score'].mean()\n",
    "    desvio_padrao = resultado_cv['test_score'].std()\n",
    "    return media, [(media - 2 * desvio_padrao) * 100, (media + 2 * desvio_padrao) * 100]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "07fc9a08-adb0-4a9f-b5aa-362a38229367",
   "metadata": {},
   "outputs": [],
   "source": [
    "def imprime_resultado(r2, MSE, MSA, media_cv, intervalo_cv):\n",
    "    print(f\"R2: {r2_score_dummy:.4f}\")\n",
    "    print(f\"MSE: {MSE_dummy:.4f}\")\n",
    "    print(f\"MSA: {MSA_dummy:.4f}\")\n",
    "    print(f\"Média CV: {media_cv:4f}\")\n",
    "    print(f\"Intervalo CV: [{intervalo_cv[0]:.4f}, {intervalo_cv[1]:.4f}]\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "578e0916-c3eb-4db5-aa37-77c4d3be2d77",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.dummy import DummyRegressor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "fc012e4f-496d-4ca1-b75b-2db1850a94ad",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Dummy Regressor usando a estratégia de sempre chutar a mediana para a predição\n",
    "modelo_dummy = DummyRegressor(strategy=\"median\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "6f744114-7ed1-47bc-871f-3ca0dcca912e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DummyRegressor(strategy='median')"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Treinando o modelo com os dados de treino\n",
    "modelo_dummy.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "d62c8500-3621-4c19-8a2c-806f45e9f946",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Utilizando Cross Validate (cv)\n",
    "cv = KFold(n_splits = 10, shuffle=True, random_state=42)\n",
    "resultado_cv = cross_validate(modelo_dummy, X, y, cv = cv, return_train_score=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "4f809783-7c90-40cd-8b01-11b0eca67f69",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Realizando a predicao\n",
    "y_pred = modelo_dummy.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "2c99f518-4e1b-4bce-8ca3-84bb22348e2e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "R2: -0.0010\n",
      "MSE: 0.8861\n",
      "MSA: 0.7670\n",
      "Média CV: -0.002026\n",
      "Intervalo CV: [-0.4652, 0.0601]\n"
     ]
    }
   ],
   "source": [
    "# Métricas e resultados\n",
    "r2_score_dummy = r2_score(y_test, y_pred)\n",
    "MSE_dummy = mean_squared_error(y_test, y_pred)\n",
    "MSA_dummy = mean_absolute_error(y_test, y_pred)\n",
    "media_cv_dummy, intervalo_cv_dummy = cv_stats(resultado_cv)\n",
    "\n",
    "imprime_resultado(r2_score_dummy, MSE_dummy, MSA_dummy, media_cv_dummy, intervalo_cv_dummy)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6557d144-4f32-43b0-9b89-5ff65485be8c",
   "metadata": {},
   "source": [
    "- Observacoes importantes:\n",
    "    - Para a métrica R2 (grau de ajuste das previsões aos dados reais), o valor ideal é 1 (modelo perfeito), e os piores modelos tem valores com valor 0 ou menor que 0 para o R2.\n",
    "    - Para a métrica MSE (Erro Quadrático Médio), mede a média dos erros quadráticos sobre a diferença entre os valores estimados e os valores reais. Quanto mais próximo de 0.0, melhor é o modelo, e quanto mais distante, pior (maior erro)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "052ed8a6-fd6b-4b8e-b1e9-95abb22d7717",
   "metadata": {},
   "outputs": [],
   "source": [
    "resultados_modelos_dict[\"modelo\"].append(\"Dummy\")\n",
    "resultados_modelos_dict[\"R2\"].append(r2_score_dummy)\n",
    "resultados_modelos_dict[\"MSE\"].append(MSE_dummy)\n",
    "resultados_modelos_dict[\"MSA\"].append(MSA_dummy)\n",
    "resultados_modelos_dict[\"media_cv\"].append(media_cv_dummy)\n",
    "resultados_modelos_dict[\"intervalo_cv\"].append(f\"[{intervalo_cv_dummy[0]:.4f}, {intervalo_cv_dummy[1]:.4f}]\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "173be3c5-fcc6-4aae-af00-7753d7d602ed",
   "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>R2</th>\n",
       "      <th>MSE</th>\n",
       "      <th>MSA</th>\n",
       "      <th>media_cv</th>\n",
       "      <th>intervalo_cv</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>modelo</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>Dummy</th>\n",
       "      <td>-0.001048</td>\n",
       "      <td>0.886094</td>\n",
       "      <td>0.767022</td>\n",
       "      <td>-0.002026</td>\n",
       "      <td>[-0.4652, 0.0601]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              R2       MSE       MSA  media_cv       intervalo_cv\n",
       "modelo                                                           \n",
       "Dummy  -0.001048  0.886094  0.767022 -0.002026  [-0.4652, 0.0601]"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "resultados_modelos_df = pd.DataFrame(resultados_modelos_dict).set_index(\"modelo\")\n",
    "resultados_modelos_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8928f819-7d17-402a-bb58-455fd763fe86",
   "metadata": {},
   "source": [
    "#### b) Regressão Linear/ Polinomial\n",
    "- Para problemas de regressão, a solução mais clássica é o uso de regressão linear. Porém, nem sempre os dados apresentam comportamento linear.\n",
    "- Por conta disso, alguns modelos de regressão polinomal também serão avaliados (polinomial de 2ª a 4ª grau)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "452b009c-8b03-4a17-8522-b15ad15459d6",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Importando o modelo para regressão linear e polinomial\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.preprocessing import PolynomialFeatures"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "196da648-6c60-43a3-a5ff-24d3efba04a0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "R2: -0.0010\n",
      "MSE: 0.8861\n",
      "MSA: 0.7670\n",
      "Média CV: 0.667432\n",
      "Intervalo CV: [61.7446, 71.7419]\n",
      "R2: -0.0010\n",
      "MSE: 0.8861\n",
      "MSA: 0.7670\n",
      "Média CV: 0.667432\n",
      "Intervalo CV: [61.7446, 71.7419]\n",
      "R2: -0.0010\n",
      "MSE: 0.8861\n",
      "MSA: 0.7670\n",
      "Média CV: 0.667432\n",
      "Intervalo CV: [61.7446, 71.7419]\n",
      "R2: -0.0010\n",
      "MSE: 0.8861\n",
      "MSA: 0.7670\n",
      "Média CV: 0.667432\n",
      "Intervalo CV: [61.7446, 71.7419]\n"
     ]
    }
   ],
   "source": [
    "for i in range(1, 5):\n",
    "    model_LR = LinearRegression()\n",
    "    \n",
    "    # Itera os graus dos polinômios, de 1ª a 5ª\n",
    "    poly = PolynomialFeatures(i)\n",
    "    \n",
    "    # Treinando o modelo corrente com os dados de treino\n",
    "    X_train_poly = poly.fit_transform(X_train)\n",
    "    X_test_poly = poly.fit_transform(X_test)\n",
    "    model_LR.fit(X_train_poly, y_train)\n",
    "    y_pred = model_LR.predict(X_test_poly)\n",
    "    \n",
    "    # Utilizando Cross Validate (cv)\n",
    "    cv = KFold(n_splits = 10, shuffle=True, random_state=42)\n",
    "    resultado_cv = cross_validate(model_LR, X, y, cv = cv, return_train_score=False)\n",
    "    \n",
    "    # Métricas e resultados\n",
    "    r2_score_lr = r2_score(y_test, y_pred)\n",
    "    MSE_lr = mean_squared_error(y_test, y_pred)\n",
    "    MSA_lr = mean_absolute_error(y_test, y_pred)\n",
    "    media_cv_lr, intervalo_cv_lr = cv_stats(resultado_cv)\n",
    "    imprime_resultado(r2_score_lr, MSE_lr, MSA_lr, media_cv_lr, intervalo_cv_lr)\n",
    "    \n",
    "    # Adicionando resultados para análises posteriores\n",
    "    resultados_modelos_dict[\"modelo\"].append(f\"Regressao grau {i}\")\n",
    "    resultados_modelos_dict[\"R2\"].append(r2_score_lr)\n",
    "    resultados_modelos_dict[\"MSE\"].append(MSE_lr)\n",
    "    resultados_modelos_dict[\"MSA\"].append(MSA_lr)\n",
    "    resultados_modelos_dict[\"media_cv\"].append(media_cv_lr)\n",
    "    resultados_modelos_dict[\"intervalo_cv\"].append(f\"[{intervalo_cv_lr[0]:.4f}, {intervalo_cv_lr[1]:.4f}]\")\n",
    "    resultados_modelos_df = pd.DataFrame(resultados_modelos_dict).set_index(\"modelo\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "c55447f9-3f50-482e-b6ab-e2793a5856b9",
   "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>R2</th>\n",
       "      <th>MSE</th>\n",
       "      <th>MSA</th>\n",
       "      <th>media_cv</th>\n",
       "      <th>intervalo_cv</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>modelo</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>Dummy</th>\n",
       "      <td>-0.001048</td>\n",
       "      <td>0.886094</td>\n",
       "      <td>0.767022</td>\n",
       "      <td>-0.002026</td>\n",
       "      <td>[-0.4652, 0.0601]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Regressao grau 1</th>\n",
       "      <td>0.668827</td>\n",
       "      <td>0.293144</td>\n",
       "      <td>0.410012</td>\n",
       "      <td>0.667432</td>\n",
       "      <td>[61.7446, 71.7419]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Regressao grau 2</th>\n",
       "      <td>0.672947</td>\n",
       "      <td>0.289497</td>\n",
       "      <td>0.407328</td>\n",
       "      <td>0.667432</td>\n",
       "      <td>[61.7446, 71.7419]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Regressao grau 3</th>\n",
       "      <td>0.633021</td>\n",
       "      <td>0.324838</td>\n",
       "      <td>0.406381</td>\n",
       "      <td>0.667432</td>\n",
       "      <td>[61.7446, 71.7419]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Regressao grau 4</th>\n",
       "      <td>0.190064</td>\n",
       "      <td>0.716928</td>\n",
       "      <td>0.422995</td>\n",
       "      <td>0.667432</td>\n",
       "      <td>[61.7446, 71.7419]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                        R2       MSE       MSA  media_cv        intervalo_cv\n",
       "modelo                                                                      \n",
       "Dummy            -0.001048  0.886094  0.767022 -0.002026   [-0.4652, 0.0601]\n",
       "Regressao grau 1  0.668827  0.293144  0.410012  0.667432  [61.7446, 71.7419]\n",
       "Regressao grau 2  0.672947  0.289497  0.407328  0.667432  [61.7446, 71.7419]\n",
       "Regressao grau 3  0.633021  0.324838  0.406381  0.667432  [61.7446, 71.7419]\n",
       "Regressao grau 4  0.190064  0.716928  0.422995  0.667432  [61.7446, 71.7419]"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "resultados_modelos_df"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d6fb1777-a454-4844-9c70-5e53bd9f094a",
   "metadata": {},
   "source": [
    " - O modelo de regressao linear/ polinomial superou o baseline.\n",
    " - Outro fator importante é que a regressão polinomial de grau se destacou como a melhor taxa de ajuste aos dados (R²) e com o menor erro quadrático médio (MSE), embora com valores próximos ao da regressão linear (grau 1). "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "208b8c3b-5b54-446b-b378-e1162a048074",
   "metadata": {},
   "source": [
    "#### c) Decision Tree Regressor: Árvore de Decisao\n",
    "- Para o modelo de regressão usando árvores de decisão, serão usadas 5 e 10 níveis (depth) para avaliar o desempenho dessa técnica."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "535d75b5-5bb2-4135-8754-07b00a702855",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.tree import DecisionTreeRegressor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "a4ed6580-9790-46ee-a5e9-a4fc7f79cf57",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "R2: -0.0010\n",
      "MSE: 0.8861\n",
      "MSA: 0.7670\n",
      "Média CV: 0.676848\n",
      "Intervalo CV: [64.1467, 71.2230]\n",
      "R2: -0.0010\n",
      "MSE: 0.8861\n",
      "MSA: 0.7670\n",
      "Média CV: 0.684414\n",
      "Intervalo CV: [62.7645, 74.1184]\n"
     ]
    }
   ],
   "source": [
    "for i in range(1, 3):\n",
    "    \n",
    "    # Treinando o modelo corrente com os dados de treino\n",
    "    modelo_dtr = DecisionTreeRegressor(random_state=42, max_depth = 5 * i)\n",
    "    modelo_dtr.fit(X_train, y_train)\n",
    "    y_pred = modelo_dtr.predict(X_test)\n",
    "        \n",
    "    # Utilizando Cross Validate (cv)\n",
    "    cv = KFold(n_splits = 10, shuffle=True, random_state=42)\n",
    "    resultado_cv = cross_validate(modelo_dtr, X, y, cv = cv, return_train_score=False)\n",
    "    \n",
    "    # Métricas e resultados\n",
    "    r2_score_dtr = r2_score(y_test, y_pred)\n",
    "    MSE_dtr = mean_squared_error(y_test, y_pred)\n",
    "    MSA_dtr = mean_absolute_error(y_test, y_pred)\n",
    "    media_cv_dtr, intervalo_cv_dtr = cv_stats(resultado_cv)\n",
    "    imprime_resultado(r2_score_dtr, MSE_dtr, MSA_dtr, media_cv_dtr, intervalo_cv_dtr)\n",
    "    \n",
    "    # Adicionando resultados para análises posteriores\n",
    "    resultados_modelos_dict[\"modelo\"].append(f\"Decision Tree (depth={5 * i})\")\n",
    "    resultados_modelos_dict[\"R2\"].append(r2_score_dtr)\n",
    "    resultados_modelos_dict[\"MSE\"].append(MSE_dtr)\n",
    "    resultados_modelos_dict[\"MSA\"].append(MSA_dtr)\n",
    "    resultados_modelos_dict[\"media_cv\"].append(media_cv_dtr)\n",
    "    resultados_modelos_dict[\"intervalo_cv\"].append(f\"[{intervalo_cv_dtr[0]:.4f}, {intervalo_cv_dtr[1]:.4f}]\")\n",
    "    resultados_modelos_df = pd.DataFrame(resultados_modelos_dict).set_index(\"modelo\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "07e32f50-2636-4e04-929f-1f32d96e34c4",
   "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>R2</th>\n",
       "      <th>MSE</th>\n",
       "      <th>MSA</th>\n",
       "      <th>media_cv</th>\n",
       "      <th>intervalo_cv</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>modelo</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>Dummy</th>\n",
       "      <td>-0.001048</td>\n",
       "      <td>0.886094</td>\n",
       "      <td>0.767022</td>\n",
       "      <td>-0.002026</td>\n",
       "      <td>[-0.4652, 0.0601]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Regressao grau 1</th>\n",
       "      <td>0.668827</td>\n",
       "      <td>0.293144</td>\n",
       "      <td>0.410012</td>\n",
       "      <td>0.667432</td>\n",
       "      <td>[61.7446, 71.7419]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Regressao grau 2</th>\n",
       "      <td>0.672947</td>\n",
       "      <td>0.289497</td>\n",
       "      <td>0.407328</td>\n",
       "      <td>0.667432</td>\n",
       "      <td>[61.7446, 71.7419]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Regressao grau 3</th>\n",
       "      <td>0.633021</td>\n",
       "      <td>0.324838</td>\n",
       "      <td>0.406381</td>\n",
       "      <td>0.667432</td>\n",
       "      <td>[61.7446, 71.7419]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Regressao grau 4</th>\n",
       "      <td>0.190064</td>\n",
       "      <td>0.716928</td>\n",
       "      <td>0.422995</td>\n",
       "      <td>0.667432</td>\n",
       "      <td>[61.7446, 71.7419]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Decision Tree (depth=5)</th>\n",
       "      <td>0.688423</td>\n",
       "      <td>0.275797</td>\n",
       "      <td>0.409847</td>\n",
       "      <td>0.676848</td>\n",
       "      <td>[64.1467, 71.2230]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Decision Tree (depth=10)</th>\n",
       "      <td>0.684988</td>\n",
       "      <td>0.278838</td>\n",
       "      <td>0.399057</td>\n",
       "      <td>0.684414</td>\n",
       "      <td>[62.7645, 74.1184]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                R2       MSE       MSA  media_cv  \\\n",
       "modelo                                                             \n",
       "Dummy                    -0.001048  0.886094  0.767022 -0.002026   \n",
       "Regressao grau 1          0.668827  0.293144  0.410012  0.667432   \n",
       "Regressao grau 2          0.672947  0.289497  0.407328  0.667432   \n",
       "Regressao grau 3          0.633021  0.324838  0.406381  0.667432   \n",
       "Regressao grau 4          0.190064  0.716928  0.422995  0.667432   \n",
       "Decision Tree (depth=5)   0.688423  0.275797  0.409847  0.676848   \n",
       "Decision Tree (depth=10)  0.684988  0.278838  0.399057  0.684414   \n",
       "\n",
       "                                intervalo_cv  \n",
       "modelo                                        \n",
       "Dummy                      [-0.4652, 0.0601]  \n",
       "Regressao grau 1          [61.7446, 71.7419]  \n",
       "Regressao grau 2          [61.7446, 71.7419]  \n",
       "Regressao grau 3          [61.7446, 71.7419]  \n",
       "Regressao grau 4          [61.7446, 71.7419]  \n",
       "Decision Tree (depth=5)   [64.1467, 71.2230]  \n",
       "Decision Tree (depth=10)  [62.7645, 74.1184]  "
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "resultados_modelos_df"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "06103be0-4e94-41d9-b120-d8968341d8a8",
   "metadata": {},
   "source": [
    "#### d) Random Forest Regressor: Floresta Aleatória\n",
    "- Para o modelo de Regressão baseado em Floresta Aleatória, serão utilizados 100 estimadores por padrão e o nível máximo de cada árvore será 5, 10, 15 e 20."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "8901eb00-c1fb-414c-b9bf-f77bc50a0dbc",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import RandomForestRegressor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "f6efba29-d1e5-456e-bc2b-272894dde2b0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "R2: -0.0010\n",
      "MSE: 0.8861\n",
      "MSA: 0.7670\n",
      "Média CV: 0.689551\n",
      "Intervalo CV: [64.8861, 73.0241]\n",
      "R2: -0.0010\n",
      "MSE: 0.8861\n",
      "MSA: 0.7670\n",
      "Média CV: 0.729265\n",
      "Intervalo CV: [68.8328, 77.0202]\n",
      "R2: -0.0010\n",
      "MSE: 0.8861\n",
      "MSA: 0.7670\n",
      "Média CV: 0.754516\n",
      "Intervalo CV: [70.8857, 80.0175]\n",
      "R2: -0.0010\n",
      "MSE: 0.8861\n",
      "MSA: 0.7670\n",
      "Média CV: 0.753305\n",
      "Intervalo CV: [70.4168, 80.2442]\n"
     ]
    }
   ],
   "source": [
    "for i in range(1, 5):\n",
    "    \n",
    "    # Treinando o modelo corrente com os dados de treino\n",
    "    modelo_rfr = RandomForestRegressor(n_estimators=100, max_depth=5 * i, random_state=42)\n",
    "    modelo_rfr.fit(X_train, y_train)\n",
    "    y_pred = modelo_rfr.predict(X_test)\n",
    "    \n",
    "    # Utilizando Cross Validate (cv)\n",
    "    cv = KFold(n_splits = 10, shuffle=True, random_state=42)\n",
    "    resultado_cv = cross_validate(modelo_rfr, X, y, cv = cv, return_train_score=False)\n",
    "    \n",
    "    \n",
    "    # Métricas e resultados\n",
    "    r2_score_rfr = r2_score(y_test, y_pred)\n",
    "    MSE_rfr = mean_squared_error(y_test, y_pred)\n",
    "    MSA_rfr = mean_absolute_error(y_test, y_pred)\n",
    "    media_cv_rfr, intervalo_cv_rfr = cv_stats(resultado_cv)\n",
    "    imprime_resultado(r2_score_rfr, MSE_rfr, MSA_rfr, media_cv_rfr, intervalo_cv_rfr)\n",
    "    \n",
    "    # Adicionando resultados para análises posteriores\n",
    "    resultados_modelos_dict[\"modelo\"].append(f\"Random Forest (depth={5 * i}, n_estimators=100)\")\n",
    "    resultados_modelos_dict[\"R2\"].append(r2_score_rfr)\n",
    "    resultados_modelos_dict[\"MSE\"].append(MSE_rfr)\n",
    "    resultados_modelos_dict[\"MSA\"].append(MSA_rfr)\n",
    "    resultados_modelos_dict[\"media_cv\"].append(media_cv_rfr)\n",
    "    resultados_modelos_dict[\"intervalo_cv\"].append(f\"[{intervalo_cv_rfr[0]:.4f}, {intervalo_cv_rfr[1]:.4f}]\")\n",
    "    resultados_modelos_df = pd.DataFrame(resultados_modelos_dict).set_index(\"modelo\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "e41fbcf0-dc8c-4544-b3ac-b29a9241f4f3",
   "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>R2</th>\n",
       "      <th>MSE</th>\n",
       "      <th>MSA</th>\n",
       "      <th>media_cv</th>\n",
       "      <th>intervalo_cv</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>modelo</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>Dummy</th>\n",
       "      <td>-0.001048</td>\n",
       "      <td>0.886094</td>\n",
       "      <td>0.767022</td>\n",
       "      <td>-0.002026</td>\n",
       "      <td>[-0.4652, 0.0601]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Regressao grau 1</th>\n",
       "      <td>0.668827</td>\n",
       "      <td>0.293144</td>\n",
       "      <td>0.410012</td>\n",
       "      <td>0.667432</td>\n",
       "      <td>[61.7446, 71.7419]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Regressao grau 2</th>\n",
       "      <td>0.672947</td>\n",
       "      <td>0.289497</td>\n",
       "      <td>0.407328</td>\n",
       "      <td>0.667432</td>\n",
       "      <td>[61.7446, 71.7419]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Regressao grau 3</th>\n",
       "      <td>0.633021</td>\n",
       "      <td>0.324838</td>\n",
       "      <td>0.406381</td>\n",
       "      <td>0.667432</td>\n",
       "      <td>[61.7446, 71.7419]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Regressao grau 4</th>\n",
       "      <td>0.190064</td>\n",
       "      <td>0.716928</td>\n",
       "      <td>0.422995</td>\n",
       "      <td>0.667432</td>\n",
       "      <td>[61.7446, 71.7419]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Decision Tree (depth=5)</th>\n",
       "      <td>0.688423</td>\n",
       "      <td>0.275797</td>\n",
       "      <td>0.409847</td>\n",
       "      <td>0.676848</td>\n",
       "      <td>[64.1467, 71.2230]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Decision Tree (depth=10)</th>\n",
       "      <td>0.684988</td>\n",
       "      <td>0.278838</td>\n",
       "      <td>0.399057</td>\n",
       "      <td>0.684414</td>\n",
       "      <td>[62.7645, 74.1184]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Random Forest (depth=5, n_estimators=100)</th>\n",
       "      <td>0.705993</td>\n",
       "      <td>0.260245</td>\n",
       "      <td>0.399486</td>\n",
       "      <td>0.689551</td>\n",
       "      <td>[64.8861, 73.0241]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Random Forest (depth=10, n_estimators=100)</th>\n",
       "      <td>0.739192</td>\n",
       "      <td>0.230858</td>\n",
       "      <td>0.374197</td>\n",
       "      <td>0.729265</td>\n",
       "      <td>[68.8328, 77.0202]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Random Forest (depth=15, n_estimators=100)</th>\n",
       "      <td>0.751706</td>\n",
       "      <td>0.219781</td>\n",
       "      <td>0.352551</td>\n",
       "      <td>0.754516</td>\n",
       "      <td>[70.8857, 80.0175]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Random Forest (depth=20, n_estimators=100)</th>\n",
       "      <td>0.746360</td>\n",
       "      <td>0.224514</td>\n",
       "      <td>0.346492</td>\n",
       "      <td>0.753305</td>\n",
       "      <td>[70.4168, 80.2442]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  R2       MSE       MSA  \\\n",
       "modelo                                                                     \n",
       "Dummy                                      -0.001048  0.886094  0.767022   \n",
       "Regressao grau 1                            0.668827  0.293144  0.410012   \n",
       "Regressao grau 2                            0.672947  0.289497  0.407328   \n",
       "Regressao grau 3                            0.633021  0.324838  0.406381   \n",
       "Regressao grau 4                            0.190064  0.716928  0.422995   \n",
       "Decision Tree (depth=5)                     0.688423  0.275797  0.409847   \n",
       "Decision Tree (depth=10)                    0.684988  0.278838  0.399057   \n",
       "Random Forest (depth=5, n_estimators=100)   0.705993  0.260245  0.399486   \n",
       "Random Forest (depth=10, n_estimators=100)  0.739192  0.230858  0.374197   \n",
       "Random Forest (depth=15, n_estimators=100)  0.751706  0.219781  0.352551   \n",
       "Random Forest (depth=20, n_estimators=100)  0.746360  0.224514  0.346492   \n",
       "\n",
       "                                            media_cv        intervalo_cv  \n",
       "modelo                                                                    \n",
       "Dummy                                      -0.002026   [-0.4652, 0.0601]  \n",
       "Regressao grau 1                            0.667432  [61.7446, 71.7419]  \n",
       "Regressao grau 2                            0.667432  [61.7446, 71.7419]  \n",
       "Regressao grau 3                            0.667432  [61.7446, 71.7419]  \n",
       "Regressao grau 4                            0.667432  [61.7446, 71.7419]  \n",
       "Decision Tree (depth=5)                     0.676848  [64.1467, 71.2230]  \n",
       "Decision Tree (depth=10)                    0.684414  [62.7645, 74.1184]  \n",
       "Random Forest (depth=5, n_estimators=100)   0.689551  [64.8861, 73.0241]  \n",
       "Random Forest (depth=10, n_estimators=100)  0.729265  [68.8328, 77.0202]  \n",
       "Random Forest (depth=15, n_estimators=100)  0.754516  [70.8857, 80.0175]  \n",
       "Random Forest (depth=20, n_estimators=100)  0.753305  [70.4168, 80.2442]  "
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "resultados_modelos_df"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6d605ba7-2e2a-4553-9515-d4d7226e1fc0",
   "metadata": {},
   "source": [
    " - Pelo que pode-se notar da tabela com os resultados dos desempenhos de cada modelo, o modelo de Random Forest com 15 de profundidade e 100 estimadores é o mais preciso, uma vez que ela atingiu o maior R2, o menor MSE, o segundo menor MSA, a melhor e intervalo mais consistente de validação cruzada.\n",
    " - Vale ressaltar que o desempenho médio do modelo de floresta aleatória superou os demais modelos.\n",
    " - O modelo \"Random Forest\" será o escolhido para a otimização com hiperparâmetros! "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dae6e169-75f5-4512-bf9b-b9402a07beae",
   "metadata": {},
   "source": [
    "### 5.5 - Otimizando o modelo de Random Forest com hiperparâmetros"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "id": "029120c1-a31e-48d7-8fac-21701b95c388",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import KFold\n",
    "from sklearn.model_selection import cross_val_score \n",
    "from sklearn.model_selection import RandomizedSearchCV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "id": "90c09b70-9f0c-4b54-a60f-58a726071196",
   "metadata": {},
   "outputs": [],
   "source": [
    "#definindo o K - número de subconjuntos\n",
    "cv = KFold(n_splits = 5, shuffle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "id": "2808d736-1132-4f67-825c-92a1d827e8ca",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Desempenho médio de r2: 0.714\n"
     ]
    }
   ],
   "source": [
    "desempenhoRF_cross_val_r2 = cross_val_score(estimator = modelo_rfr,\n",
    "                                       X = X_train, \n",
    "                                       y = y_train,\n",
    "                                       scoring = 'r2',\n",
    "                                       cv = cv)\n",
    "print('Desempenho médio de r2: {}'.format(round(desempenhoRF_cross_val_r2.mean(), 3)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "id": "186827c6-af40-4889-83fd-7818d9e750c0",
   "metadata": {},
   "outputs": [],
   "source": [
    "param_random_rf = {\n",
    "    'criterion': [\"squared_error\", \"absolute_error\", \"poisson\"],\n",
    "    'n_estimators': [25, 50, 75, 100],\n",
    "    'max_depth': [15, 17, 19, 20],\n",
    "    'min_samples_split': [1, 3, 5, 7, 9, 11],\n",
    "    'min_samples_leaf': [1, 5, 10, 15, 20]\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "id": "7dacb851-4090-44b8-ac59-9a28c052bfa3",
   "metadata": {},
   "outputs": [],
   "source": [
    "random_cv_rf = RandomizedSearchCV(\n",
    "    modelo_rfr, \n",
    "    param_random_rf, \n",
    "    n_iter=100, \n",
    "    cv=cv, \n",
    "    scoring='r2',\n",
    "    n_jobs=-1,\n",
    "    random_state = 42\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "id": "2f7d1629-8b4b-44f3-9791-0aeaab697e20",
   "metadata": {},
   "outputs": [],
   "source": [
    "rf_randomcv = random_cv_rf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "id": "bb86d14f-022e-461e-9f0c-f2590a24421d",
   "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>mean_fit_time</th>\n",
       "      <th>std_fit_time</th>\n",
       "      <th>mean_score_time</th>\n",
       "      <th>std_score_time</th>\n",
       "      <th>param_n_estimators</th>\n",
       "      <th>param_min_samples_split</th>\n",
       "      <th>param_min_samples_leaf</th>\n",
       "      <th>param_max_depth</th>\n",
       "      <th>param_criterion</th>\n",
       "      <th>params</th>\n",
       "      <th>split0_test_score</th>\n",
       "      <th>split1_test_score</th>\n",
       "      <th>split2_test_score</th>\n",
       "      <th>split3_test_score</th>\n",
       "      <th>split4_test_score</th>\n",
       "      <th>mean_test_score</th>\n",
       "      <th>std_test_score</th>\n",
       "      <th>rank_test_score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.043493</td>\n",
       "      <td>0.002985</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>25</td>\n",
       "      <td>1</td>\n",
       "      <td>10</td>\n",
       "      <td>17</td>\n",
       "      <td>squared_error</td>\n",
       "      <td>{'n_estimators': 25, 'min_samples_split': 1, '...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>38.411541</td>\n",
       "      <td>0.372252</td>\n",
       "      <td>0.054944</td>\n",
       "      <td>0.007893</td>\n",
       "      <td>50</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>17</td>\n",
       "      <td>absolute_error</td>\n",
       "      <td>{'n_estimators': 50, 'min_samples_split': 3, '...</td>\n",
       "      <td>0.667410</td>\n",
       "      <td>0.734960</td>\n",
       "      <td>0.725862</td>\n",
       "      <td>0.708837</td>\n",
       "      <td>0.734589</td>\n",
       "      <td>0.714332</td>\n",
       "      <td>0.025301</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>16.707454</td>\n",
       "      <td>0.281759</td>\n",
       "      <td>0.021905</td>\n",
       "      <td>0.002810</td>\n",
       "      <td>25</td>\n",
       "      <td>11</td>\n",
       "      <td>10</td>\n",
       "      <td>15</td>\n",
       "      <td>absolute_error</td>\n",
       "      <td>{'n_estimators': 25, 'min_samples_split': 11, ...</td>\n",
       "      <td>0.642046</td>\n",
       "      <td>0.709425</td>\n",
       "      <td>0.696773</td>\n",
       "      <td>0.675052</td>\n",
       "      <td>0.704227</td>\n",
       "      <td>0.685505</td>\n",
       "      <td>0.024688</td>\n",
       "      <td>50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.808387</td>\n",
       "      <td>0.007672</td>\n",
       "      <td>0.038603</td>\n",
       "      <td>0.001897</td>\n",
       "      <td>50</td>\n",
       "      <td>9</td>\n",
       "      <td>10</td>\n",
       "      <td>15</td>\n",
       "      <td>squared_error</td>\n",
       "      <td>{'n_estimators': 50, 'min_samples_split': 9, '...</td>\n",
       "      <td>0.652380</td>\n",
       "      <td>0.712975</td>\n",
       "      <td>0.706173</td>\n",
       "      <td>0.688720</td>\n",
       "      <td>0.712884</td>\n",
       "      <td>0.694626</td>\n",
       "      <td>0.022904</td>\n",
       "      <td>24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>17.278913</td>\n",
       "      <td>0.235417</td>\n",
       "      <td>0.022864</td>\n",
       "      <td>0.004378</td>\n",
       "      <td>25</td>\n",
       "      <td>3</td>\n",
       "      <td>5</td>\n",
       "      <td>17</td>\n",
       "      <td>absolute_error</td>\n",
       "      <td>{'n_estimators': 25, 'min_samples_split': 3, '...</td>\n",
       "      <td>0.648930</td>\n",
       "      <td>0.714027</td>\n",
       "      <td>0.706326</td>\n",
       "      <td>0.683911</td>\n",
       "      <td>0.706579</td>\n",
       "      <td>0.691955</td>\n",
       "      <td>0.023763</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>95</th>\n",
       "      <td>0.114557</td>\n",
       "      <td>0.010070</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>100</td>\n",
       "      <td>1</td>\n",
       "      <td>10</td>\n",
       "      <td>20</td>\n",
       "      <td>squared_error</td>\n",
       "      <td>{'n_estimators': 100, 'min_samples_split': 1, ...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>86</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>96</th>\n",
       "      <td>0.719981</td>\n",
       "      <td>0.042447</td>\n",
       "      <td>0.036128</td>\n",
       "      <td>0.003474</td>\n",
       "      <td>50</td>\n",
       "      <td>3</td>\n",
       "      <td>10</td>\n",
       "      <td>20</td>\n",
       "      <td>squared_error</td>\n",
       "      <td>{'n_estimators': 50, 'min_samples_split': 3, '...</td>\n",
       "      <td>0.652378</td>\n",
       "      <td>0.712936</td>\n",
       "      <td>0.706189</td>\n",
       "      <td>0.688704</td>\n",
       "      <td>0.712847</td>\n",
       "      <td>0.694611</td>\n",
       "      <td>0.022895</td>\n",
       "      <td>26</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>97</th>\n",
       "      <td>1.521897</td>\n",
       "      <td>0.086100</td>\n",
       "      <td>0.032090</td>\n",
       "      <td>0.007564</td>\n",
       "      <td>50</td>\n",
       "      <td>9</td>\n",
       "      <td>1</td>\n",
       "      <td>19</td>\n",
       "      <td>poisson</td>\n",
       "      <td>{'n_estimators': 50, 'min_samples_split': 9, '...</td>\n",
       "      <td>0.328706</td>\n",
       "      <td>0.419161</td>\n",
       "      <td>0.456549</td>\n",
       "      <td>0.401927</td>\n",
       "      <td>0.331177</td>\n",
       "      <td>0.387504</td>\n",
       "      <td>0.050214</td>\n",
       "      <td>80</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>98</th>\n",
       "      <td>0.033764</td>\n",
       "      <td>0.005513</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>25</td>\n",
       "      <td>1</td>\n",
       "      <td>15</td>\n",
       "      <td>15</td>\n",
       "      <td>poisson</td>\n",
       "      <td>{'n_estimators': 25, 'min_samples_split': 1, '...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>95</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99</th>\n",
       "      <td>1.423578</td>\n",
       "      <td>0.125746</td>\n",
       "      <td>0.074476</td>\n",
       "      <td>0.002964</td>\n",
       "      <td>100</td>\n",
       "      <td>3</td>\n",
       "      <td>10</td>\n",
       "      <td>17</td>\n",
       "      <td>squared_error</td>\n",
       "      <td>{'n_estimators': 100, 'min_samples_split': 3, ...</td>\n",
       "      <td>0.654022</td>\n",
       "      <td>0.714555</td>\n",
       "      <td>0.705414</td>\n",
       "      <td>0.687976</td>\n",
       "      <td>0.713572</td>\n",
       "      <td>0.695108</td>\n",
       "      <td>0.022646</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>100 rows × 18 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    mean_fit_time  std_fit_time  mean_score_time  std_score_time  \\\n",
       "0        0.043493      0.002985         0.000000        0.000000   \n",
       "1       38.411541      0.372252         0.054944        0.007893   \n",
       "2       16.707454      0.281759         0.021905        0.002810   \n",
       "3        0.808387      0.007672         0.038603        0.001897   \n",
       "4       17.278913      0.235417         0.022864        0.004378   \n",
       "..            ...           ...              ...             ...   \n",
       "95       0.114557      0.010070         0.000000        0.000000   \n",
       "96       0.719981      0.042447         0.036128        0.003474   \n",
       "97       1.521897      0.086100         0.032090        0.007564   \n",
       "98       0.033764      0.005513         0.000000        0.000000   \n",
       "99       1.423578      0.125746         0.074476        0.002964   \n",
       "\n",
       "   param_n_estimators param_min_samples_split param_min_samples_leaf  \\\n",
       "0                  25                       1                     10   \n",
       "1                  50                       3                      1   \n",
       "2                  25                      11                     10   \n",
       "3                  50                       9                     10   \n",
       "4                  25                       3                      5   \n",
       "..                ...                     ...                    ...   \n",
       "95                100                       1                     10   \n",
       "96                 50                       3                     10   \n",
       "97                 50                       9                      1   \n",
       "98                 25                       1                     15   \n",
       "99                100                       3                     10   \n",
       "\n",
       "   param_max_depth param_criterion  \\\n",
       "0               17   squared_error   \n",
       "1               17  absolute_error   \n",
       "2               15  absolute_error   \n",
       "3               15   squared_error   \n",
       "4               17  absolute_error   \n",
       "..             ...             ...   \n",
       "95              20   squared_error   \n",
       "96              20   squared_error   \n",
       "97              19         poisson   \n",
       "98              15         poisson   \n",
       "99              17   squared_error   \n",
       "\n",
       "                                               params  split0_test_score  \\\n",
       "0   {'n_estimators': 25, 'min_samples_split': 1, '...                NaN   \n",
       "1   {'n_estimators': 50, 'min_samples_split': 3, '...           0.667410   \n",
       "2   {'n_estimators': 25, 'min_samples_split': 11, ...           0.642046   \n",
       "3   {'n_estimators': 50, 'min_samples_split': 9, '...           0.652380   \n",
       "4   {'n_estimators': 25, 'min_samples_split': 3, '...           0.648930   \n",
       "..                                                ...                ...   \n",
       "95  {'n_estimators': 100, 'min_samples_split': 1, ...                NaN   \n",
       "96  {'n_estimators': 50, 'min_samples_split': 3, '...           0.652378   \n",
       "97  {'n_estimators': 50, 'min_samples_split': 9, '...           0.328706   \n",
       "98  {'n_estimators': 25, 'min_samples_split': 1, '...                NaN   \n",
       "99  {'n_estimators': 100, 'min_samples_split': 3, ...           0.654022   \n",
       "\n",
       "    split1_test_score  split2_test_score  split3_test_score  \\\n",
       "0                 NaN                NaN                NaN   \n",
       "1            0.734960           0.725862           0.708837   \n",
       "2            0.709425           0.696773           0.675052   \n",
       "3            0.712975           0.706173           0.688720   \n",
       "4            0.714027           0.706326           0.683911   \n",
       "..                ...                ...                ...   \n",
       "95                NaN                NaN                NaN   \n",
       "96           0.712936           0.706189           0.688704   \n",
       "97           0.419161           0.456549           0.401927   \n",
       "98                NaN                NaN                NaN   \n",
       "99           0.714555           0.705414           0.687976   \n",
       "\n",
       "    split4_test_score  mean_test_score  std_test_score  rank_test_score  \n",
       "0                 NaN              NaN             NaN              100  \n",
       "1            0.734589         0.714332        0.025301                2  \n",
       "2            0.704227         0.685505        0.024688               50  \n",
       "3            0.712884         0.694626        0.022904               24  \n",
       "4            0.706579         0.691955        0.023763               30  \n",
       "..                ...              ...             ...              ...  \n",
       "95                NaN              NaN             NaN               86  \n",
       "96           0.712847         0.694611        0.022895               26  \n",
       "97           0.331177         0.387504        0.050214               80  \n",
       "98                NaN              NaN             NaN               95  \n",
       "99           0.713572         0.695108        0.022646               15  \n",
       "\n",
       "[100 rows x 18 columns]"
      ]
     },
     "execution_count": 130,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame(rf_randomcv.cv_results_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "id": "50261bd9-021b-447a-9b52-97b2353e4b79",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'n_estimators': 100,\n",
       " 'min_samples_split': 3,\n",
       " 'min_samples_leaf': 1,\n",
       " 'max_depth': 20,\n",
       " 'criterion': 'absolute_error'}"
      ]
     },
     "execution_count": 131,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf_randomcv.best_params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "id": "2c4adde8-3acd-430e-9013-b2aa6c859ed7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7143355139948764"
      ]
     },
     "execution_count": 132,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf_randomcv.best_score_"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "581acc68-94ef-4270-a6fb-818ff43f04b9",
   "metadata": {},
   "source": [
    "#### Modelo com os parâmetros otimizados"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "id": "019636e9-997f-4b84-a54c-f6cefcc70a47",
   "metadata": {},
   "outputs": [],
   "source": [
    "rf_otimizado2 = RandomForestRegressor(random_state=42,\n",
    "                                      n_estimators=100,\n",
    "                                     min_samples_split= 3,\n",
    "                                     min_samples_leaf=1,\n",
    "                                     max_depth = 20,\n",
    "                                     criterion = 'absolute_error')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "id": "c14b188a-1b70-46d7-9188-9d69bb5a7eec",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestRegressor(criterion='absolute_error', max_depth=20,\n",
       "                      min_samples_split=3, random_state=42)"
      ]
     },
     "execution_count": 134,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf_otimizado2.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "id": "2108454a-03c0-49f7-8ed3-df0a90bd5a23",
   "metadata": {},
   "outputs": [],
   "source": [
    "predicao_rf_otimizado2 = rf_otimizado2.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "id": "debf204d-1436-456c-bca0-b8e32932607d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7488487288292761"
      ]
     },
     "execution_count": 136,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "r2_score(y_test, predicao_rf_otimizado2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "af4828fe-1667-4d48-a237-53063d3ad8a2",
   "metadata": {},
   "outputs": [],
   "source": [
    "# best"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "id": "274bffe2-1229-4d88-ab4c-2b4a85de37f0",
   "metadata": {},
   "outputs": [],
   "source": [
    "rf_otimizado_best = RandomForestRegressor(random_state=42,\n",
    "                                      n_estimators=100,\n",
    "                                     # min_samples_split= 2,\n",
    "                                     # min_samples_leaf=1,\n",
    "                                     max_depth = 15,\n",
    "                                     criterion = 'squared_error')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "id": "6d7623af-b2cc-4c54-9509-f5255c19375e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestRegressor(max_depth=15, random_state=42)"
      ]
     },
     "execution_count": 108,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf_otimizado_best.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "id": "f804abf7-7ae6-4406-baf0-ca21a1f8f9d9",
   "metadata": {},
   "outputs": [],
   "source": [
    "predicao_rf_otimizado_best = rf_otimizado_best.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "id": "2fe6cbe6-e2fa-41d2-bba1-ada91af31d2d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.751706332183583"
      ]
     },
     "execution_count": 110,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "r2_score(y_test, predicao_rf_otimizado_best)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}