Dimensionality_reduction_modeling_tuning.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Dimensionality_reduction_modeling_tuning.ipynb",
      "provenance": [],
      "authorship_tag": "ABX9TyN9vf6SiqYv6zMOu7kjx/yl",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/salmankhaliq22/10-Time-Series-Analysis/blob/main/Dimensionality_reduction_modeling_tuning.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# RFE for Regression\n",
        "\n",
        "- Here we are using data from sklearn.datasets and would be using diabetes dataset for regression models, you can check out the details in the [link](https://scikit-learn.org/stable/modules/generated/sklearn.datasets.load_diabetes.html#sklearn.datasets.load_diabetes)"
      ],
      "metadata": {
        "id": "zkibyhO4nIBu"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 21,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "jd8n7laZnGX4",
        "outputId": "3e3e5f7c-5eef-4775-f102-7d533a1e0158"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Index(['Open', 'Close', 'Volume', 'Prev_Open', 'Prev_Close', 'Prev_Vol',\n",
              "       'daily_returns', 'Weekday', '5SMA', '10SMA', '20SMA', '50SMA', '100SMA',\n",
              "       '200SMA', 'MACD', 'MACD_signal', 'RSI', 'RSI_Volume', 'Upper_band',\n",
              "       'Lower_band', '1d_Prev_Close', '1d_Prev_Vol', '1d_5SMA', '1d_10SMA',\n",
              "       '1d_20SMA', '1d_50SMA', '1d_100SMA', '1d_200SMA', '1d_MACD',\n",
              "       '1d_MACD_signal', '1d_RSI', '1d_RSI_Volume', '1d_Upper_band',\n",
              "       '1d_Lower_band', '2d_Prev_Close', '2d_Prev_Vol', '2d_5SMA', '2d_10SMA',\n",
              "       '2d_20SMA', '2d_50SMA', '2d_100SMA', '2d_200SMA', '2d_MACD',\n",
              "       '2d_MACD_signal', '2d_RSI', '2d_RSI_Volume', '2d_Upper_band',\n",
              "       '2d_Lower_band', '5d_Prev_Close', '5d_Prev_Vol', '5d_5SMA', '5d_10SMA',\n",
              "       '5d_20SMA', '5d_50SMA', '5d_100SMA', '5d_200SMA', '5d_MACD',\n",
              "       '5d_MACD_signal', '5d_RSI', '5d_RSI_Volume', '5d_Upper_band',\n",
              "       '5d_Lower_band', '10d_Prev_Close', '10d_Prev_Vol', '10d_5SMA',\n",
              "       '10d_10SMA', '10d_20SMA', '10d_50SMA', '10d_100SMA', '10d_200SMA',\n",
              "       '10d_MACD', '10d_MACD_signal', '10d_RSI', '10d_RSI_Volume',\n",
              "       '10d_Upper_band', '10d_Lower_band', 'DEXJPUS', 'DEXUSEU',\n",
              "       'DCOILBRENTEU', 'GVZCLS'],\n",
              "      dtype='object')"
            ]
          },
          "metadata": {},
          "execution_count": 21
        }
      ],
      "source": [
        "import pandas as pd\n",
        "df = pd.read_csv('AAPL.csv', index_col = 'DATE', parse_dates = True)\n",
        "df.columns"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "X = df[['Prev_Open', 'Prev_Close', '5SMA', 'MACD', '2d_Prev_Vol', 'daily_returns', 'Upper_band'\n",
        "        , '10SMA', '1d_Prev_Vol', 'Prev_Vol', 'GVZCLS', 'DEXUSEU', 'Lower_band', 'MACD_signal'\n",
        "        , '2d_RSI', '50SMA', '20SMA', '5d_Prev_Vol', '10d_RSI', '1d_RSI', 'RSI', '2d_Prev_Close'\n",
        "        , 'DEXJPUS', '200SMA', '10d_Prev_Vol', '2d_5SMA', '1d_Upper_band', 'DCOILBRENTEU']].copy()\n",
        "y = df['Close'].copy()\n"
      ],
      "metadata": {
        "id": "V2x3XqCavZEC"
      },
      "execution_count": 23,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Lets look at the head of the X variable , which contains the features\n",
        "X.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 418
        },
        "id": "CpKZB1zbnvf3",
        "outputId": "2969b442-c58c-4cdc-d0b4-f01647c51579"
      },
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "            Prev_Open  Prev_Close      5SMA      MACD  2d_Prev_Vol  \\\n",
              "DATE                                                                 \n",
              "2008-06-04   6.673571    6.620357  6.670857  0.166305     0.229989   \n",
              "2008-06-05   6.572143    6.613929  6.657857  0.154645     0.069345   \n",
              "2008-06-06   6.655000    6.765357  6.677429  0.155827     0.006562   \n",
              "2008-06-09   6.714286    6.630000  6.655214  0.144179     0.329360   \n",
              "2008-06-10   6.599643    6.486071  6.623143  0.121929     1.499707   \n",
              "\n",
              "            daily_returns  Upper_band     10SMA  1d_Prev_Vol      Prev_Vol  \\\n",
              "DATE                                                                         \n",
              "2008-06-04      -0.003923    6.863812  6.580929     0.103966  7.505204e+08   \n",
              "2008-06-05      -0.000971    6.860104  6.578393    -0.031361  7.269836e+08   \n",
              "2008-06-06       0.022895    6.876954  6.618536     0.039151  7.554456e+08   \n",
              "2008-06-09      -0.020007    6.877852  6.649214     0.279275  9.664228e+08   \n",
              "2008-06-10      -0.021709    6.880449  6.650786     0.954003  1.888393e+09   \n",
              "\n",
              "            ...   10d_RSI    1d_RSI        RSI  2d_Prev_Close  DEXJPUS  \\\n",
              "DATE        ...                                                          \n",
              "2008-06-04  ... -0.054854 -0.018504  57.484047      -0.017907   105.29   \n",
              "2008-06-05  ... -0.090112 -0.004889  57.202994      -0.004890   104.99   \n",
              "2008-06-06  ...  0.170872  0.082557  61.925510       0.021902   106.00   \n",
              "2008-06-09  ...  0.085196 -0.096023  55.979232       0.002430   105.33   \n",
              "2008-06-10  ... -0.096550 -0.099065  50.433633      -0.041282   106.06   \n",
              "\n",
              "              200SMA  10d_Prev_Vol   2d_5SMA  1d_Upper_band  DCOILBRENTEU  \n",
              "DATE                                                                       \n",
              "2008-06-04  5.676195     -0.206487  0.004161       0.000144        126.28  \n",
              "2008-06-05  5.687468     -0.250416 -0.003080      -0.000540        121.72  \n",
              "2008-06-06  5.699470     -0.347436  0.000985       0.002456        122.36  \n",
              "2008-06-09  5.709839     -0.199143 -0.000397       0.000131        132.81  \n",
              "2008-06-10  5.718607      1.082211 -0.008130       0.000377        134.43  \n",
              "\n",
              "[5 rows x 28 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-a0644956-6c21-4413-9cc1-29f0678c063a\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <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>Prev_Open</th>\n",
              "      <th>Prev_Close</th>\n",
              "      <th>5SMA</th>\n",
              "      <th>MACD</th>\n",
              "      <th>2d_Prev_Vol</th>\n",
              "      <th>daily_returns</th>\n",
              "      <th>Upper_band</th>\n",
              "      <th>10SMA</th>\n",
              "      <th>1d_Prev_Vol</th>\n",
              "      <th>Prev_Vol</th>\n",
              "      <th>...</th>\n",
              "      <th>10d_RSI</th>\n",
              "      <th>1d_RSI</th>\n",
              "      <th>RSI</th>\n",
              "      <th>2d_Prev_Close</th>\n",
              "      <th>DEXJPUS</th>\n",
              "      <th>200SMA</th>\n",
              "      <th>10d_Prev_Vol</th>\n",
              "      <th>2d_5SMA</th>\n",
              "      <th>1d_Upper_band</th>\n",
              "      <th>DCOILBRENTEU</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>DATE</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>2008-06-04</th>\n",
              "      <td>6.673571</td>\n",
              "      <td>6.620357</td>\n",
              "      <td>6.670857</td>\n",
              "      <td>0.166305</td>\n",
              "      <td>0.229989</td>\n",
              "      <td>-0.003923</td>\n",
              "      <td>6.863812</td>\n",
              "      <td>6.580929</td>\n",
              "      <td>0.103966</td>\n",
              "      <td>7.505204e+08</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.054854</td>\n",
              "      <td>-0.018504</td>\n",
              "      <td>57.484047</td>\n",
              "      <td>-0.017907</td>\n",
              "      <td>105.29</td>\n",
              "      <td>5.676195</td>\n",
              "      <td>-0.206487</td>\n",
              "      <td>0.004161</td>\n",
              "      <td>0.000144</td>\n",
              "      <td>126.28</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2008-06-05</th>\n",
              "      <td>6.572143</td>\n",
              "      <td>6.613929</td>\n",
              "      <td>6.657857</td>\n",
              "      <td>0.154645</td>\n",
              "      <td>0.069345</td>\n",
              "      <td>-0.000971</td>\n",
              "      <td>6.860104</td>\n",
              "      <td>6.578393</td>\n",
              "      <td>-0.031361</td>\n",
              "      <td>7.269836e+08</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.090112</td>\n",
              "      <td>-0.004889</td>\n",
              "      <td>57.202994</td>\n",
              "      <td>-0.004890</td>\n",
              "      <td>104.99</td>\n",
              "      <td>5.687468</td>\n",
              "      <td>-0.250416</td>\n",
              "      <td>-0.003080</td>\n",
              "      <td>-0.000540</td>\n",
              "      <td>121.72</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2008-06-06</th>\n",
              "      <td>6.655000</td>\n",
              "      <td>6.765357</td>\n",
              "      <td>6.677429</td>\n",
              "      <td>0.155827</td>\n",
              "      <td>0.006562</td>\n",
              "      <td>0.022895</td>\n",
              "      <td>6.876954</td>\n",
              "      <td>6.618536</td>\n",
              "      <td>0.039151</td>\n",
              "      <td>7.554456e+08</td>\n",
              "      <td>...</td>\n",
              "      <td>0.170872</td>\n",
              "      <td>0.082557</td>\n",
              "      <td>61.925510</td>\n",
              "      <td>0.021902</td>\n",
              "      <td>106.00</td>\n",
              "      <td>5.699470</td>\n",
              "      <td>-0.347436</td>\n",
              "      <td>0.000985</td>\n",
              "      <td>0.002456</td>\n",
              "      <td>122.36</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2008-06-09</th>\n",
              "      <td>6.714286</td>\n",
              "      <td>6.630000</td>\n",
              "      <td>6.655214</td>\n",
              "      <td>0.144179</td>\n",
              "      <td>0.329360</td>\n",
              "      <td>-0.020007</td>\n",
              "      <td>6.877852</td>\n",
              "      <td>6.649214</td>\n",
              "      <td>0.279275</td>\n",
              "      <td>9.664228e+08</td>\n",
              "      <td>...</td>\n",
              "      <td>0.085196</td>\n",
              "      <td>-0.096023</td>\n",
              "      <td>55.979232</td>\n",
              "      <td>0.002430</td>\n",
              "      <td>105.33</td>\n",
              "      <td>5.709839</td>\n",
              "      <td>-0.199143</td>\n",
              "      <td>-0.000397</td>\n",
              "      <td>0.000131</td>\n",
              "      <td>132.81</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2008-06-10</th>\n",
              "      <td>6.599643</td>\n",
              "      <td>6.486071</td>\n",
              "      <td>6.623143</td>\n",
              "      <td>0.121929</td>\n",
              "      <td>1.499707</td>\n",
              "      <td>-0.021709</td>\n",
              "      <td>6.880449</td>\n",
              "      <td>6.650786</td>\n",
              "      <td>0.954003</td>\n",
              "      <td>1.888393e+09</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.096550</td>\n",
              "      <td>-0.099065</td>\n",
              "      <td>50.433633</td>\n",
              "      <td>-0.041282</td>\n",
              "      <td>106.06</td>\n",
              "      <td>5.718607</td>\n",
              "      <td>1.082211</td>\n",
              "      <td>-0.008130</td>\n",
              "      <td>0.000377</td>\n",
              "      <td>134.43</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>5 rows × 28 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-a0644956-6c21-4413-9cc1-29f0678c063a')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-a0644956-6c21-4413-9cc1-29f0678c063a button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-a0644956-6c21-4413-9cc1-29f0678c063a');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 24
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Lets do the same for the target variable\n",
        "y.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "exZlMedjn07v",
        "outputId": "6bcacf80-a115-494d-cb57-ed15eb47fc54"
      },
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "DATE\n",
              "2008-06-04    6.613929\n",
              "2008-06-05    6.765357\n",
              "2008-06-06    6.630000\n",
              "2008-06-09    6.486071\n",
              "2008-06-10    6.630000\n",
              "Name: Close, dtype: float64"
            ]
          },
          "metadata": {},
          "execution_count": 25
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# evaluate RFE for Regression\n",
        "import numpy as np\n",
        "from sklearn.model_selection import cross_val_score, KFold\n",
        "from sklearn.feature_selection import RFE\n",
        "from sklearn.tree import DecisionTreeRegressor\n",
        "from sklearn.pipeline import make_pipeline\n",
        "# this ignores the depreciation warnings etc\n",
        "import warnings\n",
        "warnings.filterwarnings(\"ignore\")\n",
        "\n",
        "# create pipeline\n",
        "rfe = RFE(estimator = DecisionTreeRegressor(), n_features_to_select = 4)\n",
        "model = DecisionTreeRegressor()\n",
        "pipeline = make_pipeline(rfe, model)\n",
        "\n",
        "# evaluate model\n",
        "cv = KFold(n_splits = 4)\n",
        "n_scores = cross_val_score(pipeline, X, y, scoring = 'neg_mean_absolute_error', cv=cv, n_jobs=-1)\n",
        "\n",
        "# report performance\n",
        "print('MAE: %.3f (%.3f)' % (np.mean(n_scores), np.std(n_scores)))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sAb8DoRZn3Jf",
        "outputId": "91bc3e84-8b5a-43b4-8737-e0b13322b8e4"
      },
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "MAE: -15.071 (21.195)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# evaluate RFE for Regression\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "from sklearn.model_selection import cross_val_score, KFold\n",
        "from sklearn.feature_selection import RFECV\n",
        "from sklearn.tree import DecisionTreeRegressor\n",
        "from sklearn.pipeline import make_pipeline\n",
        "# this ignores the depreciation warnings etc\n",
        "import warnings\n",
        "warnings.filterwarnings(\"ignore\")\n",
        "\n",
        "# Create the RFE object and compute a cross-validated score.\n",
        "min_features_to_select = 1  # Minimum number of features to consider\n",
        "cv = KFold(n_splits = 6)\n",
        "rfecv = RFECV(estimator = DecisionTreeRegressor(), step = 1, cv = cv, scoring = 'neg_mean_absolute_error', min_features_to_select = min_features_to_select)\n",
        "rfecv.fit(X, y)\n",
        "model = DecisionTreeRegressor()\n",
        "pipeline = make_pipeline(rfecv, model)\n",
        "\n",
        "# evaluate model\n",
        "cv = KFold(n_splits = 6)\n",
        "n_scores = cross_val_score(pipeline, X, y, scoring = 'neg_mean_absolute_error', cv=cv, n_jobs=-1)\n",
        "\n",
        "# report performance\n",
        "print(\"Optimal number of features : %d\" % rfecv.n_features_)\n",
        "# print('MAE: %.3f (%.3f)' % (np.mean(n_scores), np.std(n_scores)))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wYiAlKCfsYTR",
        "outputId": "dd50f911-fcd6-4dc4-bfe3-2dfbc72e55c6"
      },
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Optimal number of features : 25\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# summarize the selection of the attributes\n",
        "print(rfecv.support_)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "PD8Tc5zjw0Wy",
        "outputId": "c398b8b1-449e-4142-b5d8-9bfe8fa6ae3f"
      },
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[ True  True  True  True  True False  True  True  True  True  True  True\n",
            "  True  True  True  True  True  True  True  True False  True False  True\n",
            "  True  True  True  True]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# summarize the ranking of the attributes\n",
        "fea_rank_ = pd.DataFrame({'cols':X.columns, 'fea_rank':rfecv.ranking_})\n",
        "fea_rank_.loc[fea_rank_.fea_rank > 0].sort_values(by=['fea_rank'], ascending = True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 927
        },
        "id": "pBkRrsK7xfR5",
        "outputId": "6b8cd845-44da-41a3-da2b-ff1bb6cc2ba7"
      },
      "execution_count": 29,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "             cols  fea_rank\n",
              "0       Prev_Open         1\n",
              "25        2d_5SMA         1\n",
              "24   10d_Prev_Vol         1\n",
              "23         200SMA         1\n",
              "21  2d_Prev_Close         1\n",
              "19         1d_RSI         1\n",
              "18        10d_RSI         1\n",
              "17    5d_Prev_Vol         1\n",
              "16          20SMA         1\n",
              "15          50SMA         1\n",
              "14         2d_RSI         1\n",
              "26  1d_Upper_band         1\n",
              "13    MACD_signal         1\n",
              "11        DEXUSEU         1\n",
              "10         GVZCLS         1\n",
              "9        Prev_Vol         1\n",
              "8     1d_Prev_Vol         1\n",
              "7           10SMA         1\n",
              "6      Upper_band         1\n",
              "4     2d_Prev_Vol         1\n",
              "3            MACD         1\n",
              "2            5SMA         1\n",
              "1      Prev_Close         1\n",
              "12     Lower_band         1\n",
              "27   DCOILBRENTEU         1\n",
              "22        DEXJPUS         2\n",
              "20            RSI         3\n",
              "5   daily_returns         4"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-c2331902-24e9-4407-bc1c-18ccc4bdec28\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <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>cols</th>\n",
              "      <th>fea_rank</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Prev_Open</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25</th>\n",
              "      <td>2d_5SMA</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>24</th>\n",
              "      <td>10d_Prev_Vol</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>23</th>\n",
              "      <td>200SMA</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>21</th>\n",
              "      <td>2d_Prev_Close</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>19</th>\n",
              "      <td>1d_RSI</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18</th>\n",
              "      <td>10d_RSI</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17</th>\n",
              "      <td>5d_Prev_Vol</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>16</th>\n",
              "      <td>20SMA</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15</th>\n",
              "      <td>50SMA</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>14</th>\n",
              "      <td>2d_RSI</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>26</th>\n",
              "      <td>1d_Upper_band</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>MACD_signal</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>DEXUSEU</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>GVZCLS</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>Prev_Vol</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>1d_Prev_Vol</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>10SMA</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>Upper_band</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>2d_Prev_Vol</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>MACD</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>5SMA</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Prev_Close</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>Lower_band</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>27</th>\n",
              "      <td>DCOILBRENTEU</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>22</th>\n",
              "      <td>DEXJPUS</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>20</th>\n",
              "      <td>RSI</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>daily_returns</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-c2331902-24e9-4407-bc1c-18ccc4bdec28')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-c2331902-24e9-4407-bc1c-18ccc4bdec28 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-c2331902-24e9-4407-bc1c-18ccc4bdec28');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 29
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Plot number of features VS. cross-validation scores\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "plt.figure()\n",
        "plt.xlabel(\"Number of features selected\")\n",
        "plt.ylabel(\"Cross validation score (MAE)\")\n",
        "plt.plot(\n",
        "    range(min_features_to_select, len(rfecv.grid_scores_) + min_features_to_select),\n",
        "    rfecv.grid_scores_,\n",
        ")\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 361
        },
        "id": "rVCfIzD7ypGu",
        "outputId": "4e02c327-6304-49b9-edfd-158177128338"
      },
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 576x396 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "[Reference](https://www.scikit-yb.org/en/latest/api/model_selection/rfecv.html)"
      ],
      "metadata": {
        "id": "KD_eP9O54Ff6"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from yellowbrick.model_selection import RFECV\n",
        "\n",
        "# Instantiate RFECV visualizer with a DecisionTreeRegressor\n",
        "min_features_to_select = 10\n",
        "cv = KFold(n_splits = 6)\n",
        "rfecv = RFECV(estimator = DecisionTreeRegressor(), step = 1, cv = cv, scoring = 'neg_mean_absolute_error', min_features_to_select = min_features_to_select)\n",
        "visualizer = rfecv\n",
        "\n",
        "# print(\"Optimal number of features : %d\" % rfecv.n_features_)\n",
        "\n",
        "visualizer.fit(X, y)        # Fit the data to the visualizer\n",
        "visualizer.show()           # Finalize and render the figure"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 393
        },
        "id": "rDZa2WHl1GeX",
        "outputId": "a2a15ca6-2082-4c66-a0d1-8fed2084cb1a"
      },
      "execution_count": 31,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 576x396 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f0e40317dd0>"
            ]
          },
          "metadata": {},
          "execution_count": 31
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Dimensionality Reduction"
      ],
      "metadata": {
        "id": "dindWUrcz9jp"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import pandas as pd\n",
        "df = pd.read_csv('AAPL.csv', index_col = 'DATE', parse_dates = True)\n",
        "X = df.drop(['Open', 'Close', 'Volume'], axis=1)\n",
        "y = df['Close']\n",
        "\n",
        "from sklearn.model_selection import train_test_split\n",
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.15, random_state = 22)\n",
        "X_train, X_valid, y_train, y_valid = train_test_split(X_train, y_train, test_size = 0.15, random_state = 22)\n",
        "\n",
        "from sklearn.preprocessing import StandardScaler\n",
        "scaler = StandardScaler()\n",
        "X_train_std = scaler.fit_transform(X_train)"
      ],
      "metadata": {
        "id": "vlpI-_Cq0KTJ"
      },
      "execution_count": 32,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.decomposition import PCA\n",
        "pca = PCA(n_components=25)\n",
        "X_train_reduced = pca.fit_transform(X_train_std)\n",
        "X_train_reduced"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KlQb3snn3VjI",
        "outputId": "a78bcde3-bae0-44af-b7e8-d3297173ca9c"
      },
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[ 0.77777112, -1.1072998 , -1.52495257, ..., -0.82278681,\n",
              "        -0.40313656, -0.21804445],\n",
              "       [-5.26410882, -3.15325151, -1.20345282, ...,  0.50073853,\n",
              "         0.66177511,  1.39249964],\n",
              "       [ 0.38916325, -2.34298324,  0.9162748 , ..., -0.3987408 ,\n",
              "        -1.39974383,  0.68861864],\n",
              "       ...,\n",
              "       [ 2.71442573, -3.51152246,  6.89113161, ..., -0.57376631,\n",
              "         0.10010326, -0.88141325],\n",
              "       [ 1.95578428, -0.52723828, -2.69088013, ..., -0.96994425,\n",
              "        -0.38624072, -0.08616727],\n",
              "       [15.72059623, -4.33589229,  4.61701145, ...,  1.13075354,\n",
              "        -0.36835776, -1.48386445]])"
            ]
          },
          "metadata": {},
          "execution_count": 41
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(pca.explained_variance_ratio_.cumsum())"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fo0jKyJ40MNa",
        "outputId": "940dd0e5-efdc-41f0-cb19-94c085764bd9"
      },
      "execution_count": 42,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[0.24913235 0.39768042 0.51621825 0.58631949 0.63562541 0.67615272\n",
            " 0.70824766 0.7338576  0.75589352 0.7743485  0.78972709 0.80402339\n",
            " 0.81771686 0.83113549 0.84420745 0.85713685 0.87005323 0.88279596\n",
            " 0.89532453 0.90743207 0.91876101 0.92983203 0.93947965 0.94883396\n",
            " 0.95647369]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "var = pca.explained_variance_ratio_\n",
        "plt.figure(figsize = (10,8))\n",
        "plt.plot(var)\n",
        "plt.xlabel('Principal Component Index')\n",
        "plt.ylabel('Explained Variance Ratio')\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 497
        },
        "id": "F5lpkrKx0PJK",
        "outputId": "77873ada-7987-4b4a-fca0-a1acbd5cba07"
      },
      "execution_count": 43,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x576 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install xgb"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Ivr-7Al9-YUp",
        "outputId": "5656155c-cbce-405d-cb50-ec39e6f605b3"
      },
      "execution_count": 69,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
            "\u001b[31mERROR: Could not find a version that satisfies the requirement xgb (from versions: none)\u001b[0m\n",
            "\u001b[31mERROR: No matching distribution found for xgb\u001b[0m\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# import multiple models \n",
        "from time import time\n",
        "from sklearn.model_selection import RepeatedKFold\n",
        "from sklearn.model_selection import GridSearchCV\n",
        "from sklearn.linear_model import LinearRegression, Ridge, Lasso, ElasticNet\n",
        "from sklearn.neighbors import KNeighborsRegressor\n",
        "from sklearn.ensemble import GradientBoostingRegressor\n",
        "from sklearn.ensemble import ExtraTreesRegressor, RandomForestRegressor\n",
        "from sklearn.tree import DecisionTreeRegressor\n",
        "from lightgbm import LGBMRegressor\n",
        "from xgboost import XGBRegressor\n",
        "\n",
        "from sklearn.metrics import explained_variance_score, mean_absolute_error, r2_score, mean_squared_error\n",
        "import seaborn as sns\n",
        "\n",
        "# create a list of these models\n",
        "regressors = [\n",
        "    KNeighborsRegressor(),\n",
        "    GradientBoostingRegressor(),\n",
        "    ExtraTreesRegressor(),\n",
        "    RandomForestRegressor(),\n",
        "    DecisionTreeRegressor(),\n",
        "    LGBMRegressor(),\n",
        "    XGBRegressor(objective='reg:squarederror', n_estimators=10, seed = 22),\n",
        "    LinearRegression(),\n",
        "    Lasso(),\n",
        "    Ridge(),\n",
        "    ElasticNet()\n",
        "    ]"
      ],
      "metadata": {
        "id": "xA0RJyR_0xWi"
      },
      "execution_count": 95,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "X_valid_std = scaler.transform(X_valid)\n",
        "X_valid_reduced = pca.transform(X_valid)"
      ],
      "metadata": {
        "id": "SNtMQX3L1Ppi"
      },
      "execution_count": 96,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# loops through models and record their performance \n",
        "head = 12\n",
        "for model in regressors[:head]:\n",
        "    start = time()\n",
        "    model.fit(X_train_reduced, y_train)\n",
        "    train_time = time() - start\n",
        "    start = time()\n",
        "    y_pred = model.predict(X_valid_reduced)\n",
        "    predict_time = time()-start    \n",
        "    print(model)\n",
        "    print(\"\\tTraining time: %0.3fs\" % train_time)\n",
        "    print(\"\\tPrediction time: %0.3fs\" % predict_time)\n",
        "    print(\"\\tExplained variance:\", explained_variance_score(y_valid, y_pred))\n",
        "    print('\\tRMSE: %f' % np.sqrt(mean_squared_error(y_valid, y_pred)))\n",
        "    print(\"\\tMean absolute error:\", mean_absolute_error(y_valid, y_pred))\n",
        "    print(\"\\tR2 score:\", r2_score(y_valid, y_pred))\n",
        "    print()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ltdHoFQk088L",
        "outputId": "41c02f6d-8a75-4f85-a792-b303d1ce6c34"
      },
      "execution_count": 97,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "KNeighborsRegressor()\n",
            "\tTraining time: 0.001s\n",
            "\tPrediction time: 0.025s\n",
            "\tExplained variance: 2.220446049250313e-16\n",
            "\tRMSE: 59.680497\n",
            "\tMean absolute error: 40.1005854807637\n",
            "\tR2 score: -0.8223953128043051\n",
            "\n",
            "GradientBoostingRegressor()\n",
            "\tTraining time: 2.040s\n",
            "\tPrediction time: 0.001s\n",
            "\tExplained variance: 0.0\n",
            "\tRMSE: 70.208360\n",
            "\tMean absolute error: 54.541489603922386\n",
            "\tR2 score: -1.5220598143778474\n",
            "\n",
            "ExtraTreesRegressor()\n",
            "\tTraining time: 1.423s\n",
            "\tPrediction time: 0.010s\n",
            "\tExplained variance: 0.0\n",
            "\tRMSE: 57.462413\n",
            "\tMean absolute error: 37.07253351845283\n",
            "\tR2 score: -0.6894504354309159\n",
            "\n",
            "RandomForestRegressor()\n",
            "\tTraining time: 4.081s\n",
            "\tPrediction time: 0.011s\n",
            "\tExplained variance: 0.0\n",
            "\tRMSE: 58.637518\n",
            "\tMean absolute error: 38.6240406637002\n",
            "\tR2 score: -0.7592553898214698\n",
            "\n",
            "DecisionTreeRegressor()\n",
            "\tTraining time: 0.070s\n",
            "\tPrediction time: 0.001s\n",
            "\tExplained variance: 0.0\n",
            "\tRMSE: 59.545993\n",
            "\tMean absolute error: 39.90572017640643\n",
            "\tR2 score: -0.8141901700018894\n",
            "\n",
            "LGBMRegressor()\n",
            "\tTraining time: 0.636s\n",
            "\tPrediction time: 0.001s\n",
            "\tExplained variance: 2.220446049250313e-16\n",
            "\tRMSE: 63.849508\n",
            "\tMean absolute error: 46.0686447443756\n",
            "\tR2 score: -1.0858969229138125\n",
            "\n",
            "XGBRegressor(n_estimators=10, objective='reg:squarederror', seed=22)\n",
            "\tTraining time: 0.055s\n",
            "\tPrediction time: 0.000s\n",
            "\tExplained variance: 0.0\n",
            "\tRMSE: 56.955650\n",
            "\tMean absolute error: 36.45041960017184\n",
            "\tR2 score: -0.6597831396979463\n",
            "\n",
            "LinearRegression()\n",
            "\tTraining time: 0.011s\n",
            "\tPrediction time: 0.004s\n",
            "\tExplained variance: -4880702662801.098\n",
            "\tRMSE: 145545987.024962\n",
            "\tMean absolute error: 107910126.5653402\n",
            "\tR2 score: -10838725090993.53\n",
            "\n",
            "Lasso()\n",
            "\tTraining time: 0.005s\n",
            "\tPrediction time: 0.001s\n",
            "\tExplained variance: -24675399306617.613\n",
            "\tRMSE: 327261353.985748\n",
            "\tMean absolute error: 242638443.82758328\n",
            "\tR2 score: -54798311356838.6\n",
            "\n",
            "Ridge()\n",
            "\tTraining time: 0.005s\n",
            "\tPrediction time: 0.001s\n",
            "\tExplained variance: -4889494004867.59\n",
            "\tRMSE: 145677012.052666\n",
            "\tMean absolute error: 108007271.83455148\n",
            "\tR2 score: -10858248590244.707\n",
            "\n",
            "ElasticNet()\n",
            "\tTraining time: 0.006s\n",
            "\tPrediction time: 0.001s\n",
            "\tExplained variance: -27279798256088.56\n",
            "\tRMSE: 344099021.370310\n",
            "\tMean absolute error: 255122355.90665746\n",
            "\tR2 score: -60582139245099.13\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Hyperparameter Tuning\n",
        "\n",
        "Fitting 10 folds for each of 84 candidates, totalling 840 fits\n",
        "Best parameters found: {'learning_rate': 0.1, 'n_estimators': 350, 'subsample': 0.5}\n",
        "Lowest RMSE found: 3.656886711120419"
      ],
      "metadata": {
        "id": "Y8l6Kff9PAwr"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# grid search\n",
        "model = XGBRegressor(objective='reg:squarederror', seed = 22)\n",
        "# n_estimators = range(50, 400, 50)\n",
        "param_grid = {'learning_rate': [0.01, 0.1, 0.5, 0.9],\n",
        "              'n_estimators': range(50, 400, 50),\n",
        "              'subsample':[0.3, 0.5, 0.9]}\n",
        "              # Best learning rate turns out to be 0.5 out of 'learning_rate': [0.01, 0.1, 0.5, 0.9]\n",
        "              # best n_estimator turns out to be 350 out of 'n_estimators': range(50, 400, 50)\n",
        "              # best subsample turns out to be \n",
        "# dict(n_estimators=n_estimators)\n",
        "kfold = KFold(n_splits=10)\n",
        "grid_search = GridSearchCV(model, param_grid, scoring=\"neg_mean_squared_error\", n_jobs=-1, cv=kfold, verbose=1)\n",
        "grid_result = grid_search.fit(X_train_reduced, y_train)\n",
        "# summarize results\n",
        "# print(\"Best: %f using %s\" % (grid_result.best_score_, grid_result.best_params_))\n",
        "print('Best parameters found:', grid_result.best_params_)\n",
        "print('Lowest RMSE found:', np.sqrt(np.abs(grid_result.best_score_)))\n",
        "means = grid_result.cv_results_['mean_test_score']\n",
        "stds = grid_result.cv_results_['std_test_score']\n",
        "params = grid_result.cv_results_['params']\n",
        "# for mean, stdev, param in zip(means, stds, params):\n",
        "# \tprint(\"%f (%f) with: %r\" % (mean, stdev, param))\n",
        "# plot\n",
        "# plt.errorbar(n_estimators, means, yerr=stds)\n",
        "# plt.title(\"XGBoost n_estimators vs RMSE\")\n",
        "# plt.xlabel('n_estimators')\n",
        "# plt.ylabel('RMSE')\n",
        "# plt.savefig('n_estimators.png')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "iUmVhyJ74JVM",
        "outputId": "4240e06d-c68d-48b0-cb27-e104b07e9e25"
      },
      "execution_count": 91,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Fitting 10 folds for each of 84 candidates, totalling 840 fits\n",
            "Best parameters found: {'learning_rate': 0.1, 'n_estimators': 350, 'subsample': 0.5}\n",
            "Lowest RMSE found: 3.656886711120419\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "for mean, stdev, param in zip(means, stds, params):\n",
        "  print('%.2f   (%.2f)    with %r' % (mean, stdev, param))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "oh0wXENS7fub",
        "outputId": "ca6316b7-49e0-4240-b7cf-07e452141cb2"
      },
      "execution_count": 94,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "-1414.33   (271.11)    with {'learning_rate': 0.01, 'n_estimators': 50, 'subsample': 0.3}\n",
            "-1405.25   (268.90)    with {'learning_rate': 0.01, 'n_estimators': 50, 'subsample': 0.5}\n",
            "-1399.67   (269.69)    with {'learning_rate': 0.01, 'n_estimators': 50, 'subsample': 0.9}\n",
            "-609.67   (116.56)    with {'learning_rate': 0.01, 'n_estimators': 100, 'subsample': 0.3}\n",
            "-602.04   (115.47)    with {'learning_rate': 0.01, 'n_estimators': 100, 'subsample': 0.5}\n",
            "-598.90   (116.07)    with {'learning_rate': 0.01, 'n_estimators': 100, 'subsample': 0.9}\n",
            "-289.33   (54.84)    with {'learning_rate': 0.01, 'n_estimators': 150, 'subsample': 0.3}\n",
            "-285.31   (53.71)    with {'learning_rate': 0.01, 'n_estimators': 150, 'subsample': 0.5}\n",
            "-285.56   (53.75)    with {'learning_rate': 0.01, 'n_estimators': 150, 'subsample': 0.9}\n",
            "-156.27   (28.98)    with {'learning_rate': 0.01, 'n_estimators': 200, 'subsample': 0.3}\n",
            "-154.59   (28.52)    with {'learning_rate': 0.01, 'n_estimators': 200, 'subsample': 0.5}\n",
            "-156.36   (28.90)    with {'learning_rate': 0.01, 'n_estimators': 200, 'subsample': 0.9}\n",
            "-97.66   (17.84)    with {'learning_rate': 0.01, 'n_estimators': 250, 'subsample': 0.3}\n",
            "-97.70   (18.23)    with {'learning_rate': 0.01, 'n_estimators': 250, 'subsample': 0.5}\n",
            "-100.16   (18.67)    with {'learning_rate': 0.01, 'n_estimators': 250, 'subsample': 0.9}\n",
            "-69.05   (12.65)    with {'learning_rate': 0.01, 'n_estimators': 300, 'subsample': 0.3}\n",
            "-69.62   (12.78)    with {'learning_rate': 0.01, 'n_estimators': 300, 'subsample': 0.5}\n",
            "-72.96   (13.23)    with {'learning_rate': 0.01, 'n_estimators': 300, 'subsample': 0.9}\n",
            "-53.32   (9.55)    with {'learning_rate': 0.01, 'n_estimators': 350, 'subsample': 0.3}\n",
            "-54.14   (9.78)    with {'learning_rate': 0.01, 'n_estimators': 350, 'subsample': 0.5}\n",
            "-57.48   (9.99)    with {'learning_rate': 0.01, 'n_estimators': 350, 'subsample': 0.9}\n",
            "-37.48   (6.25)    with {'learning_rate': 0.1, 'n_estimators': 50, 'subsample': 0.3}\n",
            "-36.74   (6.04)    with {'learning_rate': 0.1, 'n_estimators': 50, 'subsample': 0.5}\n",
            "-38.41   (5.92)    with {'learning_rate': 0.1, 'n_estimators': 50, 'subsample': 0.9}\n",
            "-23.57   (3.50)    with {'learning_rate': 0.1, 'n_estimators': 100, 'subsample': 0.3}\n",
            "-22.85   (3.85)    with {'learning_rate': 0.1, 'n_estimators': 100, 'subsample': 0.5}\n",
            "-24.70   (3.90)    with {'learning_rate': 0.1, 'n_estimators': 100, 'subsample': 0.9}\n",
            "-19.34   (2.93)    with {'learning_rate': 0.1, 'n_estimators': 150, 'subsample': 0.3}\n",
            "-18.27   (3.21)    with {'learning_rate': 0.1, 'n_estimators': 150, 'subsample': 0.5}\n",
            "-19.77   (3.24)    with {'learning_rate': 0.1, 'n_estimators': 150, 'subsample': 0.9}\n",
            "-17.17   (2.65)    with {'learning_rate': 0.1, 'n_estimators': 200, 'subsample': 0.3}\n",
            "-16.11   (3.05)    with {'learning_rate': 0.1, 'n_estimators': 200, 'subsample': 0.5}\n",
            "-17.29   (2.94)    with {'learning_rate': 0.1, 'n_estimators': 200, 'subsample': 0.9}\n",
            "-16.22   (2.46)    with {'learning_rate': 0.1, 'n_estimators': 250, 'subsample': 0.3}\n",
            "-14.83   (2.80)    with {'learning_rate': 0.1, 'n_estimators': 250, 'subsample': 0.5}\n",
            "-15.89   (2.74)    with {'learning_rate': 0.1, 'n_estimators': 250, 'subsample': 0.9}\n",
            "-15.36   (2.49)    with {'learning_rate': 0.1, 'n_estimators': 300, 'subsample': 0.3}\n",
            "-13.92   (2.72)    with {'learning_rate': 0.1, 'n_estimators': 300, 'subsample': 0.5}\n",
            "-14.96   (2.63)    with {'learning_rate': 0.1, 'n_estimators': 300, 'subsample': 0.9}\n",
            "-14.71   (2.36)    with {'learning_rate': 0.1, 'n_estimators': 350, 'subsample': 0.3}\n",
            "-13.37   (2.61)    with {'learning_rate': 0.1, 'n_estimators': 350, 'subsample': 0.5}\n",
            "-14.29   (2.55)    with {'learning_rate': 0.1, 'n_estimators': 350, 'subsample': 0.9}\n",
            "-53.21   (10.09)    with {'learning_rate': 0.5, 'n_estimators': 50, 'subsample': 0.3}\n",
            "-37.34   (6.27)    with {'learning_rate': 0.5, 'n_estimators': 50, 'subsample': 0.5}\n",
            "-30.05   (3.98)    with {'learning_rate': 0.5, 'n_estimators': 50, 'subsample': 0.9}\n",
            "-58.50   (12.05)    with {'learning_rate': 0.5, 'n_estimators': 100, 'subsample': 0.3}\n",
            "-35.48   (4.84)    with {'learning_rate': 0.5, 'n_estimators': 100, 'subsample': 0.5}\n",
            "-27.23   (3.46)    with {'learning_rate': 0.5, 'n_estimators': 100, 'subsample': 0.9}\n",
            "-61.24   (10.86)    with {'learning_rate': 0.5, 'n_estimators': 150, 'subsample': 0.3}\n",
            "-35.41   (4.91)    with {'learning_rate': 0.5, 'n_estimators': 150, 'subsample': 0.5}\n",
            "-26.18   (3.16)    with {'learning_rate': 0.5, 'n_estimators': 150, 'subsample': 0.9}\n",
            "-65.40   (9.03)    with {'learning_rate': 0.5, 'n_estimators': 200, 'subsample': 0.3}\n",
            "-35.25   (4.80)    with {'learning_rate': 0.5, 'n_estimators': 200, 'subsample': 0.5}\n",
            "-25.78   (3.19)    with {'learning_rate': 0.5, 'n_estimators': 200, 'subsample': 0.9}\n",
            "-68.29   (9.86)    with {'learning_rate': 0.5, 'n_estimators': 250, 'subsample': 0.3}\n",
            "-35.54   (5.05)    with {'learning_rate': 0.5, 'n_estimators': 250, 'subsample': 0.5}\n",
            "-25.65   (3.12)    with {'learning_rate': 0.5, 'n_estimators': 250, 'subsample': 0.9}\n",
            "-70.24   (12.14)    with {'learning_rate': 0.5, 'n_estimators': 300, 'subsample': 0.3}\n",
            "-35.58   (5.00)    with {'learning_rate': 0.5, 'n_estimators': 300, 'subsample': 0.5}\n",
            "-25.58   (3.15)    with {'learning_rate': 0.5, 'n_estimators': 300, 'subsample': 0.9}\n",
            "-71.48   (10.82)    with {'learning_rate': 0.5, 'n_estimators': 350, 'subsample': 0.3}\n",
            "-35.47   (4.98)    with {'learning_rate': 0.5, 'n_estimators': 350, 'subsample': 0.5}\n",
            "-25.50   (3.10)    with {'learning_rate': 0.5, 'n_estimators': 350, 'subsample': 0.9}\n",
            "-227.35   (60.93)    with {'learning_rate': 0.9, 'n_estimators': 50, 'subsample': 0.3}\n",
            "-101.54   (28.12)    with {'learning_rate': 0.9, 'n_estimators': 50, 'subsample': 0.5}\n",
            "-51.52   (4.61)    with {'learning_rate': 0.9, 'n_estimators': 50, 'subsample': 0.9}\n",
            "-362.45   (80.05)    with {'learning_rate': 0.9, 'n_estimators': 100, 'subsample': 0.3}\n",
            "-120.65   (35.30)    with {'learning_rate': 0.9, 'n_estimators': 100, 'subsample': 0.5}\n",
            "-52.82   (5.39)    with {'learning_rate': 0.9, 'n_estimators': 100, 'subsample': 0.9}\n",
            "-525.64   (141.25)    with {'learning_rate': 0.9, 'n_estimators': 150, 'subsample': 0.3}\n",
            "-130.97   (40.01)    with {'learning_rate': 0.9, 'n_estimators': 150, 'subsample': 0.5}\n",
            "-53.98   (5.55)    with {'learning_rate': 0.9, 'n_estimators': 150, 'subsample': 0.9}\n",
            "-711.25   (135.62)    with {'learning_rate': 0.9, 'n_estimators': 200, 'subsample': 0.3}\n",
            "-133.34   (39.79)    with {'learning_rate': 0.9, 'n_estimators': 200, 'subsample': 0.5}\n",
            "-54.42   (5.67)    with {'learning_rate': 0.9, 'n_estimators': 200, 'subsample': 0.9}\n",
            "-958.66   (178.30)    with {'learning_rate': 0.9, 'n_estimators': 250, 'subsample': 0.3}\n",
            "-138.07   (39.44)    with {'learning_rate': 0.9, 'n_estimators': 250, 'subsample': 0.5}\n",
            "-54.65   (5.76)    with {'learning_rate': 0.9, 'n_estimators': 250, 'subsample': 0.9}\n",
            "-1296.58   (290.68)    with {'learning_rate': 0.9, 'n_estimators': 300, 'subsample': 0.3}\n",
            "-141.07   (37.73)    with {'learning_rate': 0.9, 'n_estimators': 300, 'subsample': 0.5}\n",
            "-54.71   (5.98)    with {'learning_rate': 0.9, 'n_estimators': 300, 'subsample': 0.9}\n",
            "-1590.53   (341.26)    with {'learning_rate': 0.9, 'n_estimators': 350, 'subsample': 0.3}\n",
            "-143.92   (38.79)    with {'learning_rate': 0.9, 'n_estimators': 350, 'subsample': 0.5}\n",
            "-54.69   (5.97)    with {'learning_rate': 0.9, 'n_estimators': 350, 'subsample': 0.9}\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "LrWu9lvFNz-e"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}