jamm1985/lab_16_intro_to_ml_decision_trees.ipynb

## lab_16_intro_to_ml_decision_trees.ipynb
{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Lab_16_intro_to_ML_decision_trees.ipynb",
      "provenance": [],
      "authorship_tag": "ABX9TyPV4wWPSiWyipiwnXU3RbL0",
      "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/gist/jamm1985/3f92508c378022803fb94cc95a22d3e9/lab_16_intro_to_ml_decision_trees.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Видео лабораторной: https://youtu.be/3N0Uz33_wFI\n",
        "\n",
        "TG: https://t.me/data_science_news\n",
        "\n",
        "\n",
        "\n",
        "---"
      ],
      "metadata": {
        "id": "HZsb7hnxGXX3"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "AQtJeVqMHMjq"
      },
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "\n",
        "import matplotlib.pylab as plt\n",
        "\n",
        "from sklearn import tree\n",
        "from sklearn.ensemble import BaggingRegressor\n",
        "from sklearn.ensemble import BaggingClassifier\n",
        "from sklearn.ensemble import RandomForestRegressor\n",
        "from sklearn.ensemble import RandomForestClassifier\n",
        "from sklearn.ensemble import GradientBoostingRegressor\n",
        "from sklearn.ensemble import GradientBoostingClassifier\n",
        "\n",
        "from sklearn.model_selection import cross_val_score\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "from sklearn.metrics import mean_squared_error\n",
        "from sklearn.metrics import r2_score"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Набор данных для регрессии\n",
        "\n",
        "[Лабораторная №13](https://youtu.be/A3LE-ZmtVGs)\n",
        "\n",
        "Model Name      | parameters | $r^2$        | Mean Squared Error|\n",
        "----------------|------------|--------------|-------------------|\n",
        "LR              | $\\bf4$     |$0.89\\pm0.04$ |$3.07\\pm1.28$      |\n",
        "LR poly 2       | $10$       |$0.98\\pm0.01$ |$0.44\\pm0.39$      |\n",
        "LR poly 3       | $20$       |$\\bf0.99\\pm0.01$ |$\\bf0.31\\pm0.24$      |\n",
        "NN              | $185$      |$0.91\\pm1.61$ |$1.86\\pm1.49$      |\n"
      ],
      "metadata": {
        "id": "9UcoBihHRodL"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!wget https://raw.githubusercontent.com/nguyen-toan/ISLR/master/dataset/Advertising.csv\n",
        "!head Advertising.csv"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hsVgffgRRbkR",
        "outputId": "0c0c54fc-892e-4a9d-b2fa-4ad04998471c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "--2022-04-20 05:24:38--  https://raw.githubusercontent.com/nguyen-toan/ISLR/master/dataset/Advertising.csv\n",
            "Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 185.199.108.133, 185.199.109.133, 185.199.110.133, ...\n",
            "Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|185.199.108.133|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 5166 (5.0K) [text/plain]\n",
            "Saving to: ‘Advertising.csv’\n",
            "\n",
            "\rAdvertising.csv       0%[                    ]       0  --.-KB/s               \rAdvertising.csv     100%[===================>]   5.04K  --.-KB/s    in 0s      \n",
            "\n",
            "2022-04-20 05:24:38 (45.8 MB/s) - ‘Advertising.csv’ saved [5166/5166]\n",
            "\n",
            "\"\",\"TV\",\"Radio\",\"Newspaper\",\"Sales\"\n",
            "\"1\",230.1,37.8,69.2,22.1\n",
            "\"2\",44.5,39.3,45.1,10.4\n",
            "\"3\",17.2,45.9,69.3,9.3\n",
            "\"4\",151.5,41.3,58.5,18.5\n",
            "\"5\",180.8,10.8,58.4,12.9\n",
            "\"6\",8.7,48.9,75,7.2\n",
            "\"7\",57.5,32.8,23.5,11.8\n",
            "\"8\",120.2,19.6,11.6,13.2\n",
            "\"9\",8.6,2.1,1,4.8\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "REGRESSION = pd.read_csv('Advertising.csv')\n",
        "REGRESSION = REGRESSION.drop(columns=['Unnamed: 0'])\n",
        "REGRESSION"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 423
        },
        "id": "9V6a71yRR4Bf",
        "outputId": "33798014-aeaf-4390-dfc4-42b7dc6d1008"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "        TV  Radio  Newspaper  Sales\n",
              "0    230.1   37.8       69.2   22.1\n",
              "1     44.5   39.3       45.1   10.4\n",
              "2     17.2   45.9       69.3    9.3\n",
              "3    151.5   41.3       58.5   18.5\n",
              "4    180.8   10.8       58.4   12.9\n",
              "..     ...    ...        ...    ...\n",
              "195   38.2    3.7       13.8    7.6\n",
              "196   94.2    4.9        8.1    9.7\n",
              "197  177.0    9.3        6.4   12.8\n",
              "198  283.6   42.0       66.2   25.5\n",
              "199  232.1    8.6        8.7   13.4\n",
              "\n",
              "[200 rows x 4 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-44e8b0a4-2e25-42f6-95bd-4fa83d4e4e51\">\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>TV</th>\n",
              "      <th>Radio</th>\n",
              "      <th>Newspaper</th>\n",
              "      <th>Sales</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>230.1</td>\n",
              "      <td>37.8</td>\n",
              "      <td>69.2</td>\n",
              "      <td>22.1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>44.5</td>\n",
              "      <td>39.3</td>\n",
              "      <td>45.1</td>\n",
              "      <td>10.4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>17.2</td>\n",
              "      <td>45.9</td>\n",
              "      <td>69.3</td>\n",
              "      <td>9.3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>151.5</td>\n",
              "      <td>41.3</td>\n",
              "      <td>58.5</td>\n",
              "      <td>18.5</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>180.8</td>\n",
              "      <td>10.8</td>\n",
              "      <td>58.4</td>\n",
              "      <td>12.9</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>195</th>\n",
              "      <td>38.2</td>\n",
              "      <td>3.7</td>\n",
              "      <td>13.8</td>\n",
              "      <td>7.6</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>196</th>\n",
              "      <td>94.2</td>\n",
              "      <td>4.9</td>\n",
              "      <td>8.1</td>\n",
              "      <td>9.7</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>197</th>\n",
              "      <td>177.0</td>\n",
              "      <td>9.3</td>\n",
              "      <td>6.4</td>\n",
              "      <td>12.8</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>198</th>\n",
              "      <td>283.6</td>\n",
              "      <td>42.0</td>\n",
              "      <td>66.2</td>\n",
              "      <td>25.5</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>199</th>\n",
              "      <td>232.1</td>\n",
              "      <td>8.6</td>\n",
              "      <td>8.7</td>\n",
              "      <td>13.4</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>200 rows × 4 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-44e8b0a4-2e25-42f6-95bd-4fa83d4e4e51')\"\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-44e8b0a4-2e25-42f6-95bd-4fa83d4e4e51 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-44e8b0a4-2e25-42f6-95bd-4fa83d4e4e51');\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": 3
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "X_reg = REGRESSION.loc[:, REGRESSION.columns != 'Sales'].to_numpy()\n",
        "y_reg = REGRESSION['Sales'].to_numpy()\n",
        "print(X_reg.shape)\n",
        "print(y_reg.shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4ruwQzKfT4wq",
        "outputId": "dca54b25-9367-42c4-c213-a4867d5eda63"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(200, 3)\n",
            "(200,)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# split regression data\n",
        "X_reg_train, X_reg_test, y_reg_train, y_reg_test = train_test_split(\n",
        "    X_reg, y_reg, test_size=0.33, random_state=0, shuffle=True)\n",
        "X_reg_train.shape, X_reg_test.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "NOFjExfQUEaH",
        "outputId": "df658d59-ecbd-446a-ef9e-0a827df9b5fc"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((134, 3), (66, 3))"
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Набор данных для классификации\n",
        "\n",
        "[Лабораторная №15](https://youtu.be/lkiFy6LQnSk)\n",
        "\n",
        "```\n",
        "Logit has 0.9961 OvR AUC with a standard deviation of 0.01\n",
        "LDA has 0.9937 OvR AUC with a standard deviation of 0.01\n",
        "QDA has 0.9888 OvR AUC with a standard deviation of 0.02\n",
        "```"
      ],
      "metadata": {
        "id": "cpCrQE49SCQD"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# download dataset\n",
        "# https://archive.ics.uci.edu/ml/datasets/seeds\n",
        "!wget https://archive.ics.uci.edu/ml/machine-learning-databases/00236/seeds_dataset.txt\n",
        "!head seeds_dataset.txt"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "85CzXy3bSAlA",
        "outputId": "63f4f619-42da-4973-fba3-19feb47988ec"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "--2022-04-20 05:26:13--  https://archive.ics.uci.edu/ml/machine-learning-databases/00236/seeds_dataset.txt\n",
            "Resolving archive.ics.uci.edu (archive.ics.uci.edu)... 128.195.10.252\n",
            "Connecting to archive.ics.uci.edu (archive.ics.uci.edu)|128.195.10.252|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 9300 (9.1K) [application/x-httpd-php]\n",
            "Saving to: ‘seeds_dataset.txt’\n",
            "\n",
            "seeds_dataset.txt   100%[===================>]   9.08K  --.-KB/s    in 0s      \n",
            "\n",
            "2022-04-20 05:26:13 (68.0 MB/s) - ‘seeds_dataset.txt’ saved [9300/9300]\n",
            "\n",
            "15.26\t14.84\t0.871\t5.763\t3.312\t2.221\t5.22\t1\n",
            "14.88\t14.57\t0.8811\t5.554\t3.333\t1.018\t4.956\t1\n",
            "14.29\t14.09\t0.905\t5.291\t3.337\t2.699\t4.825\t1\n",
            "13.84\t13.94\t0.8955\t5.324\t3.379\t2.259\t4.805\t1\n",
            "16.14\t14.99\t0.9034\t5.658\t3.562\t1.355\t5.175\t1\n",
            "14.38\t14.21\t0.8951\t5.386\t3.312\t2.462\t4.956\t1\n",
            "14.69\t14.49\t0.8799\t5.563\t3.259\t3.586\t5.219\t1\n",
            "14.11\t14.1\t0.8911\t5.42\t3.302\t2.7\t\t5\t\t1\n",
            "16.63\t15.46\t0.8747\t6.053\t3.465\t2.04\t5.877\t1\n",
            "16.44\t15.25\t0.888\t5.884\t3.505\t1.969\t5.533\t1\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# load the data\n",
        "CLASSIFICATION = pd.read_csv(\n",
        "    \"https://archive.ics.uci.edu/ml/machine-learning-databases/00236/seeds_dataset.txt\",\n",
        "    sep='\\t',\n",
        "    header=None,\n",
        "    names=[\"area\", \"perimeter\", \"compactness\", \"length\", \"width\", \"asymmetry\", \"length_groove\", \"class\"],\n",
        "    on_bad_lines='skip')\n",
        "CLASSIFICATION = CLASSIFICATION.dropna()\n",
        "CLASSIFICATION[\"class\"] = CLASSIFICATION[\"class\"].astype('int32')\n",
        "CLASSIFICATION"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 423
        },
        "id": "-PLzz6BESR8e",
        "outputId": "c0e20253-8bcd-42e8-e7b0-7db1b2ed6d9c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "      area  perimeter  compactness  length  width  asymmetry  length_groove  \\\n",
              "0    15.26      14.84       0.8710   5.763  3.312      2.221          5.220   \n",
              "1    14.88      14.57       0.8811   5.554  3.333      1.018          4.956   \n",
              "2    14.29      14.09       0.9050   5.291  3.337      2.699          4.825   \n",
              "3    13.84      13.94       0.8955   5.324  3.379      2.259          4.805   \n",
              "4    16.14      14.99       0.9034   5.658  3.562      1.355          5.175   \n",
              "..     ...        ...          ...     ...    ...        ...            ...   \n",
              "205  12.19      13.20       0.8783   5.137  2.981      3.631          4.870   \n",
              "206  11.23      12.88       0.8511   5.140  2.795      4.325          5.003   \n",
              "207  13.20      13.66       0.8883   5.236  3.232      8.315          5.056   \n",
              "208  11.84      13.21       0.8521   5.175  2.836      3.598          5.044   \n",
              "209  12.30      13.34       0.8684   5.243  2.974      5.637          5.063   \n",
              "\n",
              "     class  \n",
              "0        1  \n",
              "1        1  \n",
              "2        1  \n",
              "3        1  \n",
              "4        1  \n",
              "..     ...  \n",
              "205      3  \n",
              "206      3  \n",
              "207      3  \n",
              "208      3  \n",
              "209      3  \n",
              "\n",
              "[199 rows x 8 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-f7eb1d44-9a51-4936-bb6a-e7cff54a8561\">\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>area</th>\n",
              "      <th>perimeter</th>\n",
              "      <th>compactness</th>\n",
              "      <th>length</th>\n",
              "      <th>width</th>\n",
              "      <th>asymmetry</th>\n",
              "      <th>length_groove</th>\n",
              "      <th>class</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>15.26</td>\n",
              "      <td>14.84</td>\n",
              "      <td>0.8710</td>\n",
              "      <td>5.763</td>\n",
              "      <td>3.312</td>\n",
              "      <td>2.221</td>\n",
              "      <td>5.220</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>14.88</td>\n",
              "      <td>14.57</td>\n",
              "      <td>0.8811</td>\n",
              "      <td>5.554</td>\n",
              "      <td>3.333</td>\n",
              "      <td>1.018</td>\n",
              "      <td>4.956</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>14.29</td>\n",
              "      <td>14.09</td>\n",
              "      <td>0.9050</td>\n",
              "      <td>5.291</td>\n",
              "      <td>3.337</td>\n",
              "      <td>2.699</td>\n",
              "      <td>4.825</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>13.84</td>\n",
              "      <td>13.94</td>\n",
              "      <td>0.8955</td>\n",
              "      <td>5.324</td>\n",
              "      <td>3.379</td>\n",
              "      <td>2.259</td>\n",
              "      <td>4.805</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>16.14</td>\n",
              "      <td>14.99</td>\n",
              "      <td>0.9034</td>\n",
              "      <td>5.658</td>\n",
              "      <td>3.562</td>\n",
              "      <td>1.355</td>\n",
              "      <td>5.175</td>\n",
              "      <td>1</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",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>205</th>\n",
              "      <td>12.19</td>\n",
              "      <td>13.20</td>\n",
              "      <td>0.8783</td>\n",
              "      <td>5.137</td>\n",
              "      <td>2.981</td>\n",
              "      <td>3.631</td>\n",
              "      <td>4.870</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>206</th>\n",
              "      <td>11.23</td>\n",
              "      <td>12.88</td>\n",
              "      <td>0.8511</td>\n",
              "      <td>5.140</td>\n",
              "      <td>2.795</td>\n",
              "      <td>4.325</td>\n",
              "      <td>5.003</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>207</th>\n",
              "      <td>13.20</td>\n",
              "      <td>13.66</td>\n",
              "      <td>0.8883</td>\n",
              "      <td>5.236</td>\n",
              "      <td>3.232</td>\n",
              "      <td>8.315</td>\n",
              "      <td>5.056</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>208</th>\n",
              "      <td>11.84</td>\n",
              "      <td>13.21</td>\n",
              "      <td>0.8521</td>\n",
              "      <td>5.175</td>\n",
              "      <td>2.836</td>\n",
              "      <td>3.598</td>\n",
              "      <td>5.044</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>209</th>\n",
              "      <td>12.30</td>\n",
              "      <td>13.34</td>\n",
              "      <td>0.8684</td>\n",
              "      <td>5.243</td>\n",
              "      <td>2.974</td>\n",
              "      <td>5.637</td>\n",
              "      <td>5.063</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>199 rows × 8 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-f7eb1d44-9a51-4936-bb6a-e7cff54a8561')\"\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-f7eb1d44-9a51-4936-bb6a-e7cff54a8561 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-f7eb1d44-9a51-4936-bb6a-e7cff54a8561');\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": 7
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "X_class = CLASSIFICATION.loc[:, CLASSIFICATION.columns != 'class'].to_numpy()\n",
        "y_class = CLASSIFICATION[\"class\"].to_numpy()\n",
        "X_class.shape, y_class.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "SS9X5AIhUr0Z",
        "outputId": "510b9bac-c36e-4bdb-f9ba-d88d98eee250"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((199, 7), (199,))"
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# split regression data\n",
        "X_class_train, X_class_test, y_class_train, y_class_test = train_test_split(\n",
        "    X_class, y_class, test_size=0.33, random_state=0, shuffle=True)\n",
        "X_class_train.shape, X_class_test.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YzQcEDH5VBEg",
        "outputId": "7255c525-0a34-4b7a-961f-1d1b12648a6d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((133, 7), (66, 7))"
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Простое дерево решений (simple decision tree)"
      ],
      "metadata": {
        "id": "EPPA6y6WSoQu"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Пусть нам даны пары наблюдений $(X_1,Y_1),...,(X_n,Y_n)$, где $X_i \\in R^p$, a $Y_i$ может принимать количественные или качественные значения.\n",
        "\n",
        "Суть подхода: разделить пространство независимых переменных (предикторов) $X^1, X^2, X^3,..,X^p$ на $J$ различных непересекющихся областей $R_1, R_2, ..., R_J$. Для набора данных $\\in R_J$ выполняется единственное предсказание зависимой переменной."
      ],
      "metadata": {
        "id": "V4rdGrAwqYOI"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "![image.png](data:image/png;base64,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)\n",
        "\n",
        "https://en.wikipedia.org/wiki/Decision_tree_learning#/media/File:Cart_tree_kyphosis.png"
      ],
      "metadata": {
        "id": "mM--kn_zv2LW"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Регрессия\n",
        "\n",
        "Цель регрессии через решающие деревья найти такой набор $R_1, R_2, ..., R_J$, который минимизирует целевую функцию:\n",
        "\n",
        "$RSS = \\sum_{j=1}^J\\sum_{i \\in R_j}(y_i - \\hat{y}_{R_j})^2$\n",
        "\n",
        "### Алгоритм рекурсивного бинарного разделения (\"жадный\")\n",
        "\n",
        "Выбор предиктора $X^j$ и выборт точки разделения $s$ так что простарнство **всех** предикторов образует две области: $\\{X|X^j<s\\}$ и $\\{X|X^j>s\\}$. Выбор предиктора и точки выполняется таким образом, что бы на данном шаге **максимально** уменшить RSS. Таким образом, рассматриваются все предикторы $X^1,..,X^J$ и все варианты $s$ для каждого предиктора. \n",
        "\n",
        "То есть, для каждого $j$ и $s$ задаётся пара:\n",
        "\n",
        "$R_1(j,s)=\\{X|X^j<s\\}$ и $R_2(j,s)=\\{X|X^j \\geq s\\}$ таким образом, что $j$ и $s$ минимизируют целевую функции в виде:\n",
        "\n",
        "$\\sum_{i:x_i \\in R_1(j,s)}(y_i - \\hat{y}_{R_1})^2 + \\sum_{i:x_i \\in R_2(j,s)}(y_i - \\hat{y}_{R_2})^2$, где $\\hat{y}_{R_{1,2}}$ - это выборочное среднее в $R_{1,2}(j,s)$.\n",
        "\n",
        "Далее, процее повторяется, но уже для разделения одного из $R_j$ полученных на предыдущих шагах до выполнения условия остановки. Например, условием остановки может быть то, каждая область включает на более 10 наблюдений."
      ],
      "metadata": {
        "id": "NA2USke2TmsR"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# https://scikit-learn.org/stable/modules/generated/sklearn.tree.DecisionTreeRegressor.html#sklearn.tree.DecisionTreeRegressor\n",
        "simple_dt_reg = tree.DecisionTreeRegressor(random_state=0, max_depth=3)\n",
        "simple_dt_reg.fit(X_reg_train, y_reg_train)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7GNOIVLZSg7r",
        "outputId": "51474eb6-7709-41b7-92d3-210e47e2c9f4"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "DecisionTreeRegressor(max_depth=3, random_state=0)"
            ]
          },
          "metadata": {},
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# https://scikit-learn.org/stable/modules/tree.html\n",
        "plt.rcParams['figure.figsize'] = [15, 15]\n",
        "tree.plot_tree(simple_dt_reg, fontsize=10)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "T9PN5kqSW-l2",
        "outputId": "6fe3dcaf-1ad0-4337-8dab-346d38197879"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[Text(0.5, 0.875, 'X[0] <= 130.25\\nsquared_error = 27.346\\nsamples = 134\\nvalue = 14.33'),\n",
              " Text(0.25, 0.625, 'X[0] <= 33.3\\nsquared_error = 7.449\\nsamples = 55\\nvalue = 9.985'),\n",
              " Text(0.125, 0.375, 'X[1] <= 31.45\\nsquared_error = 2.474\\nsamples = 16\\nvalue = 6.894'),\n",
              " Text(0.0625, 0.125, 'squared_error = 1.528\\nsamples = 10\\nvalue = 6.08'),\n",
              " Text(0.1875, 0.125, 'squared_error = 1.109\\nsamples = 6\\nvalue = 8.25'),\n",
              " Text(0.375, 0.375, 'X[2] <= 49.0\\nsquared_error = 3.959\\nsamples = 39\\nvalue = 11.254'),\n",
              " Text(0.3125, 0.125, 'squared_error = 1.996\\nsamples = 33\\nvalue = 10.67'),\n",
              " Text(0.4375, 0.125, 'squared_error = 2.556\\nsamples = 6\\nvalue = 14.467'),\n",
              " Text(0.75, 0.625, 'X[1] <= 26.85\\nsquared_error = 18.91\\nsamples = 79\\nvalue = 17.354'),\n",
              " Text(0.625, 0.375, 'X[1] <= 9.7\\nsquared_error = 3.248\\nsamples = 36\\nvalue = 13.311'),\n",
              " Text(0.5625, 0.125, 'squared_error = 0.853\\nsamples = 19\\nvalue = 11.921'),\n",
              " Text(0.6875, 0.125, 'squared_error = 1.352\\nsamples = 17\\nvalue = 14.865'),\n",
              " Text(0.875, 0.375, 'X[0] <= 210.75\\nsquared_error = 6.876\\nsamples = 43\\nvalue = 20.74'),\n",
              " Text(0.8125, 0.125, 'squared_error = 2.269\\nsamples = 17\\nvalue = 18.724'),\n",
              " Text(0.9375, 0.125, 'squared_error = 5.493\\nsamples = 26\\nvalue = 22.058')]"
            ]
          },
          "metadata": {},
          "execution_count": 11
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x1080 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "simple_dt_reg = tree.DecisionTreeRegressor(random_state=0)\n",
        "simple_dt_reg.fit(X_reg_train, y_reg_train)\n",
        "y_true = y_reg_test\n",
        "y_pred = simple_dt_reg.predict(X_reg_test)\n",
        "r2_score(y_true, y_pred), mean_squared_error(y_true, y_pred)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CxTyl7-6V0ET",
        "outputId": "50135782-5434-45ca-85ac-0f5e6cf5b326"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(0.923650758203275, 1.9833333333333334)"
            ]
          },
          "metadata": {},
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "tree.plot_tree(simple_dt_reg)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "OoMPgDIJYB9O",
        "outputId": "d0697ea4-d3ed-46ee-8906-ae6c5ab52fc1"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[Text(0.43275375939849625, 0.9583333333333334, 'X[0] <= 130.25\\nsquared_error = 27.346\\nsamples = 134\\nvalue = 14.33'),\n",
              " Text(0.1756578947368421, 0.875, 'X[0] <= 33.3\\nsquared_error = 7.449\\nsamples = 55\\nvalue = 9.985'),\n",
              " Text(0.04962406015037594, 0.7916666666666666, 'X[1] <= 31.45\\nsquared_error = 2.474\\nsamples = 16\\nvalue = 6.894'),\n",
              " Text(0.02406015037593985, 0.7083333333333334, 'X[0] <= 18.0\\nsquared_error = 1.528\\nsamples = 10\\nvalue = 6.08'),\n",
              " Text(0.012030075187969926, 0.625, 'X[0] <= 4.75\\nsquared_error = 0.938\\nsamples = 5\\nvalue = 5.08'),\n",
              " Text(0.006015037593984963, 0.5416666666666666, 'squared_error = 0.0\\nsamples = 1\\nvalue = 3.2'),\n",
              " Text(0.01804511278195489, 0.5416666666666666, 'X[0] <= 15.15\\nsquared_error = 0.067\\nsamples = 4\\nvalue = 5.55'),\n",
              " Text(0.012030075187969926, 0.4583333333333333, 'X[2] <= 5.75\\nsquared_error = 0.036\\nsamples = 3\\nvalue = 5.433'),\n",
              " Text(0.006015037593984963, 0.375, 'squared_error = 0.0\\nsamples = 1\\nvalue = 5.7'),\n",
              " Text(0.01804511278195489, 0.375, 'squared_error = -0.0\\nsamples = 2\\nvalue = 5.3'),\n",
              " Text(0.02406015037593985, 0.4583333333333333, 'squared_error = -0.0\\nsamples = 1\\nvalue = 5.9'),\n",
              " Text(0.03609022556390978, 0.625, 'X[2] <= 18.85\\nsquared_error = 0.118\\nsamples = 5\\nvalue = 7.08'),\n",
              " Text(0.03007518796992481, 0.5416666666666666, 'squared_error = 0.0\\nsamples = 1\\nvalue = 7.6'),\n",
              " Text(0.042105263157894736, 0.5416666666666666, 'X[0] <= 28.05\\nsquared_error = 0.062\\nsamples = 4\\nvalue = 6.95'),\n",
              " Text(0.03609022556390978, 0.4583333333333333, 'X[2] <= 36.35\\nsquared_error = 0.029\\nsamples = 3\\nvalue = 6.833'),\n",
              " Text(0.03007518796992481, 0.375, 'X[0] <= 23.45\\nsquared_error = 0.023\\nsamples = 2\\nvalue = 6.75'),\n",
              " Text(0.02406015037593985, 0.2916666666666667, 'squared_error = 0.0\\nsamples = 1\\nvalue = 6.6'),\n",
              " Text(0.03609022556390978, 0.2916666666666667, 'squared_error = 0.0\\nsamples = 1\\nvalue = 6.9'),\n",
              " Text(0.042105263157894736, 0.375, 'squared_error = 0.0\\nsamples = 1\\nvalue = 7.0'),\n",
              " Text(0.0481203007518797, 0.4583333333333333, 'squared_error = -0.0\\nsamples = 1\\nvalue = 7.3'),\n",
              " Text(0.07518796992481203, 0.7083333333333334, 'X[0] <= 14.45\\nsquared_error = 1.109\\nsamples = 6\\nvalue = 8.25'),\n",
              " Text(0.06015037593984962, 0.625, 'X[1] <= 37.9\\nsquared_error = 0.123\\nsamples = 2\\nvalue = 6.95'),\n",
              " Text(0.05413533834586466, 0.5416666666666666, 'squared_error = 0.0\\nsamples = 1\\nvalue = 7.3'),\n",
              " Text(0.06616541353383458, 0.5416666666666666, 'squared_error = 0.0\\nsamples = 1\\nvalue = 6.6'),\n",
              " Text(0.09022556390977443, 0.625, 'X[1] <= 38.3\\nsquared_error = 0.335\\nsamples = 4\\nvalue = 8.9'),\n",
              " Text(0.07819548872180451, 0.5416666666666666, 'X[0] <= 22.35\\nsquared_error = 0.16\\nsamples = 2\\nvalue = 8.4'),\n",
              " Text(0.07218045112781955, 0.4583333333333333, 'squared_error = 0.0\\nsamples = 1\\nvalue = 8.0'),\n",
              " Text(0.08421052631578947, 0.4583333333333333, 'squared_error = -0.0\\nsamples = 1\\nvalue = 8.8'),\n",
              " Text(0.10225563909774436, 0.5416666666666666, 'X[0] <= 21.4\\nsquared_error = 0.01\\nsamples = 2\\nvalue = 9.4'),\n",
              " Text(0.0962406015037594, 0.4583333333333333, 'squared_error = 0.0\\nsamples = 1\\nvalue = 9.3'),\n",
              " Text(0.10827067669172932, 0.4583333333333333, 'squared_error = 0.0\\nsamples = 1\\nvalue = 9.5'),\n",
              " Text(0.3016917293233083, 0.7916666666666666, 'X[2] <= 49.0\\nsquared_error = 3.959\\nsamples = 39\\nvalue = 11.254'),\n",
              " Text(0.2244360902255639, 0.7083333333333334, 'X[1] <= 12.85\\nsquared_error = 1.996\\nsamples = 33\\nvalue = 10.67'),\n",
              " Text(0.16466165413533834, 0.625, 'X[0] <= 98.95\\nsquared_error = 0.964\\nsamples = 16\\nvalue = 9.794'),\n",
              " Text(0.1368421052631579, 0.5416666666666666, 'X[0] <= 66.5\\nsquared_error = 0.521\\nsamples = 12\\nvalue = 9.367'),\n",
              " Text(0.12030075187969924, 0.4583333333333333, 'X[1] <= 4.75\\nsquared_error = 0.25\\nsamples = 2\\nvalue = 8.1'),\n",
              " Text(0.11428571428571428, 0.375, 'squared_error = 0.0\\nsamples = 1\\nvalue = 7.6'),\n",
              " Text(0.12631578947368421, 0.375, 'squared_error = 0.0\\nsamples = 1\\nvalue = 8.6'),\n",
              " Text(0.15338345864661654, 0.4583333333333333, 'X[1] <= 9.6\\nsquared_error = 0.19\\nsamples = 10\\nvalue = 9.62'),\n",
              " Text(0.13834586466165413, 0.375, 'X[1] <= 0.4\\nsquared_error = 0.085\\nsamples = 7\\nvalue = 9.429'),\n",
              " Text(0.13233082706766916, 0.2916666666666667, 'squared_error = 0.0\\nsamples = 1\\nvalue = 8.8'),\n",
              " Text(0.1443609022556391, 0.2916666666666667, 'X[0] <= 85.3\\nsquared_error = 0.022\\nsamples = 6\\nvalue = 9.533'),\n",
              " Text(0.13233082706766916, 0.20833333333333334, 'X[1] <= 5.05\\nsquared_error = 0.002\\nsamples = 2\\nvalue = 9.35'),\n",
              " Text(0.12631578947368421, 0.125, 'squared_error = 0.0\\nsamples = 1\\nvalue = 9.4'),\n",
              " Text(0.13834586466165413, 0.125, 'squared_error = -0.0\\nsamples = 1\\nvalue = 9.3'),\n",
              " Text(0.15639097744360902, 0.20833333333333334, 'X[1] <= 3.2\\nsquared_error = 0.007\\nsamples = 4\\nvalue = 9.625'),\n",
              " Text(0.15037593984962405, 0.125, 'X[1] <= 1.45\\nsquared_error = 0.002\\nsamples = 2\\nvalue = 9.55'),\n",
              " Text(0.1443609022556391, 0.041666666666666664, 'squared_error = 0.0\\nsamples = 1\\nvalue = 9.5'),\n",
              " Text(0.15639097744360902, 0.041666666666666664, 'squared_error = -0.0\\nsamples = 1\\nvalue = 9.6'),\n",
              " Text(0.162406015037594, 0.125, 'squared_error = 0.0\\nsamples = 2\\nvalue = 9.7'),\n",
              " Text(0.16842105263157894, 0.375, 'X[1] <= 10.35\\nsquared_error = 0.149\\nsamples = 3\\nvalue = 10.067'),\n",
              " Text(0.162406015037594, 0.2916666666666667, 'squared_error = 0.0\\nsamples = 1\\nvalue = 10.6'),\n",
              " Text(0.17443609022556392, 0.2916666666666667, 'X[0] <= 71.2\\nsquared_error = 0.01\\nsamples = 2\\nvalue = 9.8'),\n",
              " Text(0.16842105263157894, 0.20833333333333334, 'squared_error = 0.0\\nsamples = 1\\nvalue = 9.7'),\n",
              " Text(0.18045112781954886, 0.20833333333333334, 'squared_error = -0.0\\nsamples = 1\\nvalue = 9.9'),\n",
              " Text(0.1924812030075188, 0.5416666666666666, 'X[2] <= 40.0\\nsquared_error = 0.107\\nsamples = 4\\nvalue = 11.075'),\n",
              " Text(0.18646616541353384, 0.4583333333333333, 'X[2] <= 13.35\\nsquared_error = 0.02\\nsamples = 3\\nvalue = 10.9'),\n",
              " Text(0.18045112781954886, 0.375, 'squared_error = 0.0\\nsamples = 1\\nvalue = 10.7'),\n",
              " Text(0.1924812030075188, 0.375, 'squared_error = -0.0\\nsamples = 2\\nvalue = 11.0'),\n",
              " Text(0.19849624060150375, 0.4583333333333333, 'squared_error = -0.0\\nsamples = 1\\nvalue = 11.6'),\n",
              " Text(0.28421052631578947, 0.625, 'X[0] <= 75.8\\nsquared_error = 1.565\\nsamples = 17\\nvalue = 11.494'),\n",
              " Text(0.24060150375939848, 0.5416666666666666, 'X[0] <= 46.4\\nsquared_error = 0.409\\nsamples = 10\\nvalue = 10.79'),\n",
              " Text(0.21654135338345865, 0.4583333333333333, 'X[1] <= 33.0\\nsquared_error = 0.226\\nsamples = 5\\nvalue = 10.36'),\n",
              " Text(0.20451127819548873, 0.375, 'X[1] <= 26.3\\nsquared_error = 0.062\\nsamples = 2\\nvalue = 9.85'),\n",
              " Text(0.19849624060150375, 0.2916666666666667, 'squared_error = 0.0\\nsamples = 1\\nvalue = 9.6'),\n",
              " Text(0.21052631578947367, 0.2916666666666667, 'squared_error = 0.0\\nsamples = 1\\nvalue = 10.1'),\n",
              " Text(0.22857142857142856, 0.375, 'X[1] <= 39.8\\nsquared_error = 0.047\\nsamples = 3\\nvalue = 10.7'),\n",
              " Text(0.22255639097744362, 0.2916666666666667, 'squared_error = 0.0\\nsamples = 1\\nvalue = 10.4'),\n",
              " Text(0.23458646616541354, 0.2916666666666667, 'X[1] <= 40.7\\nsquared_error = 0.002\\nsamples = 2\\nvalue = 10.85'),\n",
              " Text(0.22857142857142856, 0.20833333333333334, 'squared_error = 0.0\\nsamples = 1\\nvalue = 10.9'),\n",
              " Text(0.24060150375939848, 0.20833333333333334, 'squared_error = -0.0\\nsamples = 1\\nvalue = 10.8'),\n",
              " Text(0.2646616541353383, 0.4583333333333333, 'X[1] <= 18.65\\nsquared_error = 0.222\\nsamples = 5\\nvalue = 11.22'),\n",
              " Text(0.25263157894736843, 0.375, 'X[1] <= 16.5\\nsquared_error = 0.04\\nsamples = 2\\nvalue = 10.7'),\n",
              " Text(0.24661654135338346, 0.2916666666666667, 'squared_error = 0.0\\nsamples = 1\\nvalue = 10.5'),\n",
              " Text(0.2586466165413534, 0.2916666666666667, 'squared_error = 0.0\\nsamples = 1\\nvalue = 10.9'),\n",
              " Text(0.27669172932330827, 0.375, 'X[0] <= 66.4\\nsquared_error = 0.042\\nsamples = 3\\nvalue = 11.567'),\n",
              " Text(0.2706766917293233, 0.2916666666666667, 'X[1] <= 39.9\\nsquared_error = 0.01\\nsamples = 2\\nvalue = 11.7'),\n",
              " Text(0.2646616541353383, 0.20833333333333334, 'squared_error = 0.0\\nsamples = 1\\nvalue = 11.8'),\n",
              " Text(0.27669172932330827, 0.20833333333333334, 'squared_error = 0.0\\nsamples = 1\\nvalue = 11.6'),\n",
              " Text(0.28270676691729324, 0.2916666666666667, 'squared_error = -0.0\\nsamples = 1\\nvalue = 11.3'),\n",
              " Text(0.32781954887218046, 0.5416666666666666, 'X[1] <= 28.0\\nsquared_error = 1.497\\nsamples = 7\\nvalue = 12.5'),\n",
              " Text(0.31278195488721805, 0.4583333333333333, 'X[1] <= 16.1\\nsquared_error = 0.058\\nsamples = 5\\nvalue = 11.74'),\n",
              " Text(0.3007518796992481, 0.375, 'X[2] <= 8.15\\nsquared_error = 0.047\\nsamples = 3\\nvalue = 11.6'),\n",
              " Text(0.29473684210526313, 0.2916666666666667, 'squared_error = 0.0\\nsamples = 1\\nvalue = 11.9'),\n",
              " Text(0.3067669172932331, 0.2916666666666667, 'X[1] <= 14.4\\nsquared_error = 0.003\\nsamples = 2\\nvalue = 11.45'),\n",
              " Text(0.3007518796992481, 0.20833333333333334, 'squared_error = 0.0\\nsamples = 1\\nvalue = 11.5'),\n",
              " Text(0.31278195488721805, 0.20833333333333334, 'squared_error = 0.0\\nsamples = 1\\nvalue = 11.4'),\n",
              " Text(0.324812030075188, 0.375, 'X[1] <= 22.45\\nsquared_error = 0.003\\nsamples = 2\\nvalue = 11.95'),\n",
              " Text(0.318796992481203, 0.2916666666666667, 'squared_error = 0.0\\nsamples = 1\\nvalue = 11.9'),\n",
              " Text(0.3308270676691729, 0.2916666666666667, 'squared_error = 0.0\\nsamples = 1\\nvalue = 12.0'),\n",
              " Text(0.34285714285714286, 0.4583333333333333, 'X[1] <= 37.65\\nsquared_error = 0.04\\nsamples = 2\\nvalue = 14.4'),\n",
              " Text(0.3368421052631579, 0.375, 'squared_error = 0.0\\nsamples = 1\\nvalue = 14.2'),\n",
              " Text(0.34887218045112783, 0.375, 'squared_error = 0.0\\nsamples = 1\\nvalue = 14.6'),\n",
              " Text(0.37894736842105264, 0.7083333333333334, 'X[0] <= 91.1\\nsquared_error = 2.556\\nsamples = 6\\nvalue = 14.467'),\n",
              " Text(0.3669172932330827, 0.625, 'X[1] <= 35.4\\nsquared_error = 0.082\\nsamples = 3\\nvalue = 12.933'),\n",
              " Text(0.3609022556390977, 0.5416666666666666, 'X[1] <= 30.25\\nsquared_error = 0.023\\nsamples = 2\\nvalue = 12.75'),\n",
              " Text(0.3548872180451128, 0.4583333333333333, 'squared_error = 0.0\\nsamples = 1\\nvalue = 12.9'),\n",
              " Text(0.3669172932330827, 0.4583333333333333, 'squared_error = -0.0\\nsamples = 1\\nvalue = 12.6'),\n",
              " Text(0.37293233082706767, 0.5416666666666666, 'squared_error = -0.0\\nsamples = 1\\nvalue = 13.3'),\n",
              " Text(0.39097744360902253, 0.625, 'X[0] <= 101.85\\nsquared_error = 0.327\\nsamples = 3\\nvalue = 16.0'),\n",
              " Text(0.3849624060150376, 0.5416666666666666, 'squared_error = 0.0\\nsamples = 1\\nvalue = 15.3'),\n",
              " Text(0.3969924812030075, 0.5416666666666666, 'X[0] <= 110.25\\nsquared_error = 0.122\\nsamples = 2\\nvalue = 16.35'),\n",
              " Text(0.39097744360902253, 0.4583333333333333, 'squared_error = 0.0\\nsamples = 1\\nvalue = 16.7'),\n",
              " Text(0.4030075187969925, 0.4583333333333333, 'squared_error = -0.0\\nsamples = 1\\nvalue = 16.0'),\n",
              " Text(0.6898496240601504, 0.875, 'X[1] <= 26.85\\nsquared_error = 18.91\\nsamples = 79\\nvalue = 17.354'),\n",
              " Text(0.5503759398496241, 0.7916666666666666, 'X[1] <= 9.7\\nsquared_error = 3.248\\nsamples = 36\\nvalue = 13.311'),\n",
              " Text(0.48270676691729325, 0.7083333333333334, 'X[1] <= 3.25\\nsquared_error = 0.853\\nsamples = 19\\nvalue = 11.921'),\n",
              " Text(0.4330827067669173, 0.625, 'X[0] <= 174.8\\nsquared_error = 0.426\\nsamples = 5\\nvalue = 10.84'),\n",
              " Text(0.42105263157894735, 0.5416666666666666, 'X[1] <= 1.7\\nsquared_error = 0.01\\nsamples = 2\\nvalue = 10.2'),\n",
              " Text(0.4150375939849624, 0.4583333333333333, 'squared_error = 0.0\\nsamples = 1\\nvalue = 10.1'),\n",
              " Text(0.4270676691729323, 0.4583333333333333, 'squared_error = 0.0\\nsamples = 1\\nvalue = 10.3'),\n",
              " Text(0.44511278195488724, 0.5416666666666666, 'X[2] <= 22.45\\nsquared_error = 0.249\\nsamples = 3\\nvalue = 11.267'),\n",
              " Text(0.43909774436090226, 0.4583333333333333, 'squared_error = 0.0\\nsamples = 1\\nvalue = 10.6'),\n",
              " Text(0.45112781954887216, 0.4583333333333333, 'X[1] <= 2.7\\nsquared_error = 0.04\\nsamples = 2\\nvalue = 11.6'),\n",
              " Text(0.44511278195488724, 0.375, 'squared_error = 0.0\\nsamples = 1\\nvalue = 11.8'),\n",
              " Text(0.45714285714285713, 0.375, 'squared_error = 0.0\\nsamples = 1\\nvalue = 11.4'),\n",
              " Text(0.5323308270676692, 0.625, 'X[1] <= 7.65\\nsquared_error = 0.439\\nsamples = 14\\nvalue = 12.307'),\n",
              " Text(0.5052631578947369, 0.5416666666666666, 'X[0] <= 235.95\\nsquared_error = 0.21\\nsamples = 10\\nvalue = 11.98'),\n",
              " Text(0.48721804511278194, 0.4583333333333333, 'X[2] <= 16.25\\nsquared_error = 0.097\\nsamples = 6\\nvalue = 11.7'),\n",
              " Text(0.4691729323308271, 0.375, 'X[0] <= 151.35\\nsquared_error = 0.042\\nsamples = 3\\nvalue = 11.467'),\n",
              " Text(0.4631578947368421, 0.2916666666666667, 'squared_error = 0.0\\nsamples = 1\\nvalue = 11.2'),\n",
              " Text(0.47518796992481205, 0.2916666666666667, 'X[1] <= 5.25\\nsquared_error = 0.01\\nsamples = 2\\nvalue = 11.6'),\n",
              " Text(0.4691729323308271, 0.20833333333333334, 'squared_error = 0.0\\nsamples = 1\\nvalue = 11.5'),\n",
              " Text(0.48120300751879697, 0.20833333333333334, 'squared_error = 0.0\\nsamples = 1\\nvalue = 11.7'),\n",
              " Text(0.5052631578947369, 0.375, 'X[1] <= 4.75\\nsquared_error = 0.042\\nsamples = 3\\nvalue = 11.933'),\n",
              " Text(0.4992481203007519, 0.2916666666666667, 'X[2] <= 67.3\\nsquared_error = 0.01\\nsamples = 2\\nvalue = 11.8'),\n",
              " Text(0.4932330827067669, 0.20833333333333334, 'squared_error = 0.0\\nsamples = 1\\nvalue = 11.7'),\n",
              " Text(0.5052631578947369, 0.20833333333333334, 'squared_error = -0.0\\nsamples = 1\\nvalue = 11.9'),\n",
              " Text(0.5112781954887218, 0.2916666666666667, 'squared_error = 0.0\\nsamples = 1\\nvalue = 12.2'),\n",
              " Text(0.5233082706766917, 0.4583333333333333, 'X[2] <= 14.0\\nsquared_error = 0.085\\nsamples = 4\\nvalue = 12.4'),\n",
              " Text(0.5172932330827068, 0.375, 'squared_error = 0.0\\nsamples = 1\\nvalue = 12.8'),\n",
              " Text(0.5293233082706766, 0.375, 'X[0] <= 251.35\\nsquared_error = 0.042\\nsamples = 3\\nvalue = 12.267'),\n",
              " Text(0.5233082706766917, 0.2916666666666667, 'X[1] <= 4.6\\nsquared_error = 0.01\\nsamples = 2\\nvalue = 12.4'),\n",
              " Text(0.5172932330827068, 0.20833333333333334, 'squared_error = 0.0\\nsamples = 1\\nvalue = 12.3'),\n",
              " Text(0.5293233082706766, 0.20833333333333334, 'squared_error = 0.0\\nsamples = 1\\nvalue = 12.5'),\n",
              " Text(0.5353383458646617, 0.2916666666666667, 'squared_error = 0.0\\nsamples = 1\\nvalue = 12.0'),\n",
              " Text(0.5593984962406015, 0.5416666666666666, 'X[0] <= 216.35\\nsquared_error = 0.077\\nsamples = 4\\nvalue = 13.125'),\n",
              " Text(0.5473684210526316, 0.4583333333333333, 'X[0] <= 191.95\\nsquared_error = 0.002\\nsamples = 2\\nvalue = 12.85'),\n",
              " Text(0.5413533834586466, 0.375, 'squared_error = 0.0\\nsamples = 1\\nvalue = 12.8'),\n",
              " Text(0.5533834586466165, 0.375, 'squared_error = -0.0\\nsamples = 1\\nvalue = 12.9'),\n",
              " Text(0.5714285714285714, 0.4583333333333333, 'X[2] <= 32.6\\nsquared_error = 0.0\\nsamples = 2\\nvalue = 13.4'),\n",
              " Text(0.5654135338345865, 0.375, 'squared_error = 0.0\\nsamples = 1\\nvalue = 13.4'),\n",
              " Text(0.5774436090225564, 0.375, 'squared_error = 0.0\\nsamples = 1\\nvalue = 13.4'),\n",
              " Text(0.6180451127819548, 0.7083333333333334, 'X[0] <= 140.3\\nsquared_error = 1.352\\nsamples = 17\\nvalue = 14.865'),\n",
              " Text(0.5834586466165413, 0.625, 'X[1] <= 14.4\\nsquared_error = 0.276\\nsamples = 3\\nvalue = 12.933'),\n",
              " Text(0.5774436090225564, 0.5416666666666666, 'squared_error = 0.0\\nsamples = 1\\nvalue = 12.2'),\n",
              " Text(0.5894736842105263, 0.5416666666666666, 'X[1] <= 16.85\\nsquared_error = 0.01\\nsamples = 2\\nvalue = 13.3'),\n",
              " Text(0.5834586466165413, 0.4583333333333333, 'squared_error = 0.0\\nsamples = 1\\nvalue = 13.4'),\n",
              " Text(0.5954887218045113, 0.4583333333333333, 'squared_error = -0.0\\nsamples = 1\\nvalue = 13.2'),\n",
              " Text(0.6526315789473685, 0.625, 'X[1] <= 22.45\\nsquared_error = 0.612\\nsamples = 14\\nvalue = 15.279'),\n",
              " Text(0.631578947368421, 0.5416666666666666, 'X[0] <= 227.75\\nsquared_error = 0.296\\nsamples = 12\\nvalue = 15.083'),\n",
              " Text(0.6075187969924812, 0.4583333333333333, 'X[0] <= 181.8\\nsquared_error = 0.117\\nsamples = 6\\nvalue = 14.7'),\n",
              " Text(0.5954887218045113, 0.375, 'X[2] <= 24.9\\nsquared_error = 0.04\\nsamples = 2\\nvalue = 14.3'),\n",
              " Text(0.5894736842105263, 0.2916666666666667, 'squared_error = 0.0\\nsamples = 1\\nvalue = 14.1'),\n",
              " Text(0.6015037593984962, 0.2916666666666667, 'squared_error = -0.0\\nsamples = 1\\nvalue = 14.5'),\n",
              " Text(0.6195488721804512, 0.375, 'X[2] <= 57.8\\nsquared_error = 0.035\\nsamples = 4\\nvalue = 14.9'),\n",
              " Text(0.6135338345864662, 0.2916666666666667, 'X[0] <= 188.15\\nsquared_error = 0.007\\nsamples = 3\\nvalue = 14.8'),\n",
              " Text(0.6075187969924812, 0.20833333333333334, 'squared_error = 0.0\\nsamples = 1\\nvalue = 14.7'),\n",
              " Text(0.6195488721804512, 0.20833333333333334, 'X[0] <= 207.8\\nsquared_error = 0.002\\nsamples = 2\\nvalue = 14.85'),\n",
              " Text(0.6135338345864662, 0.125, 'squared_error = 0.0\\nsamples = 1\\nvalue = 14.9'),\n",
              " Text(0.6255639097744361, 0.125, 'squared_error = -0.0\\nsamples = 1\\nvalue = 14.8'),\n",
              " Text(0.6255639097744361, 0.2916666666666667, 'squared_error = -0.0\\nsamples = 1\\nvalue = 15.2'),\n",
              " Text(0.6556390977443609, 0.4583333333333333, 'X[1] <= 12.25\\nsquared_error = 0.182\\nsamples = 6\\nvalue = 15.467'),\n",
              " Text(0.643609022556391, 0.375, 'X[2] <= 13.9\\nsquared_error = 0.01\\nsamples = 2\\nvalue = 14.9'),\n",
              " Text(0.637593984962406, 0.2916666666666667, 'squared_error = 0.0\\nsamples = 1\\nvalue = 15.0'),\n",
              " Text(0.649624060150376, 0.2916666666666667, 'squared_error = 0.0\\nsamples = 1\\nvalue = 14.8'),\n",
              " Text(0.6676691729323309, 0.375, 'X[2] <= 24.55\\nsquared_error = 0.027\\nsamples = 4\\nvalue = 15.75'),\n",
              " Text(0.6616541353383458, 0.2916666666666667, 'squared_error = 0.0\\nsamples = 2\\nvalue = 15.9'),\n",
              " Text(0.6736842105263158, 0.2916666666666667, 'X[1] <= 16.2\\nsquared_error = 0.01\\nsamples = 2\\nvalue = 15.6'),\n",
              " Text(0.6676691729323309, 0.20833333333333334, 'squared_error = 0.0\\nsamples = 1\\nvalue = 15.7'),\n",
              " Text(0.6796992481203008, 0.20833333333333334, 'squared_error = 0.0\\nsamples = 1\\nvalue = 15.5'),\n",
              " Text(0.6736842105263158, 0.5416666666666666, 'X[2] <= 25.1\\nsquared_error = 0.902\\nsamples = 2\\nvalue = 16.45'),\n",
              " Text(0.6676691729323309, 0.4583333333333333, 'squared_error = 0.0\\nsamples = 1\\nvalue = 17.4'),\n",
              " Text(0.6796992481203008, 0.4583333333333333, 'squared_error = 0.0\\nsamples = 1\\nvalue = 15.5'),\n",
              " Text(0.8293233082706767, 0.7916666666666666, 'X[0] <= 210.75\\nsquared_error = 6.876\\nsamples = 43\\nvalue = 20.74'),\n",
              " Text(0.7684210526315789, 0.7083333333333334, 'X[1] <= 43.35\\nsquared_error = 2.269\\nsamples = 17\\nvalue = 18.724'),\n",
              " Text(0.7308270676691729, 0.625, 'X[2] <= 4.8\\nsquared_error = 0.699\\nsamples = 13\\nvalue = 18.062'),\n",
              " Text(0.724812030075188, 0.5416666666666666, 'squared_error = 0.0\\nsamples = 1\\nvalue = 19.6'),\n",
              " Text(0.7368421052631579, 0.5416666666666666, 'X[0] <= 192.4\\nsquared_error = 0.544\\nsamples = 12\\nvalue = 17.933'),\n",
              " Text(0.7097744360902256, 0.4583333333333333, 'X[1] <= 36.2\\nsquared_error = 0.272\\nsamples = 8\\nvalue = 17.537'),\n",
              " Text(0.6917293233082706, 0.375, 'X[2] <= 28.45\\nsquared_error = 0.027\\nsamples = 4\\nvalue = 17.15'),\n",
              " Text(0.6857142857142857, 0.2916666666666667, 'squared_error = 0.0\\nsamples = 2\\nvalue = 17.3'),\n",
              " Text(0.6977443609022557, 0.2916666666666667, 'X[0] <= 170.15\\nsquared_error = 0.01\\nsamples = 2\\nvalue = 17.0'),\n",
              " Text(0.6917293233082706, 0.20833333333333334, 'squared_error = 0.0\\nsamples = 1\\nvalue = 16.9'),\n",
              " Text(0.7037593984962406, 0.20833333333333334, 'squared_error = 0.0\\nsamples = 1\\nvalue = 17.1'),\n",
              " Text(0.7278195488721805, 0.375, 'X[2] <= 52.2\\nsquared_error = 0.217\\nsamples = 4\\nvalue = 17.925'),\n",
              " Text(0.7218045112781954, 0.2916666666666667, 'X[2] <= 37.4\\nsquared_error = 0.142\\nsamples = 3\\nvalue = 17.733'),\n",
              " Text(0.7157894736842105, 0.20833333333333334, 'squared_error = 0.0\\nsamples = 2\\nvalue = 18.0'),\n",
              " Text(0.7278195488721805, 0.20833333333333334, 'squared_error = -0.0\\nsamples = 1\\nvalue = 17.2'),\n",
              " Text(0.7338345864661654, 0.2916666666666667, 'squared_error = 0.0\\nsamples = 1\\nvalue = 18.5'),\n",
              " Text(0.7639097744360902, 0.4583333333333333, 'X[2] <= 42.35\\nsquared_error = 0.147\\nsamples = 4\\nvalue = 18.725'),\n",
              " Text(0.7518796992481203, 0.375, 'X[2] <= 24.0\\nsquared_error = 0.002\\nsamples = 2\\nvalue = 18.35'),\n",
              " Text(0.7458646616541353, 0.2916666666666667, 'squared_error = 0.0\\nsamples = 1\\nvalue = 18.4'),\n",
              " Text(0.7578947368421053, 0.2916666666666667, 'squared_error = -0.0\\nsamples = 1\\nvalue = 18.3'),\n",
              " Text(0.7759398496240602, 0.375, 'X[2] <= 60.8\\nsquared_error = 0.01\\nsamples = 2\\nvalue = 19.1'),\n",
              " Text(0.7699248120300752, 0.2916666666666667, 'squared_error = 0.0\\nsamples = 1\\nvalue = 19.0'),\n",
              " Text(0.7819548872180451, 0.2916666666666667, 'squared_error = -0.0\\nsamples = 1\\nvalue = 19.2'),\n",
              " Text(0.806015037593985, 0.625, 'X[2] <= 58.85\\nsquared_error = 1.317\\nsamples = 4\\nvalue = 20.875'),\n",
              " Text(0.8, 0.5416666666666666, 'X[0] <= 190.15\\nsquared_error = 0.509\\nsamples = 3\\nvalue = 21.433'),\n",
              " Text(0.793984962406015, 0.4583333333333333, 'X[0] <= 183.75\\nsquared_error = 0.062\\nsamples = 2\\nvalue = 20.95'),\n",
              " Text(0.7879699248120301, 0.375, 'squared_error = 0.0\\nsamples = 1\\nvalue = 21.2'),\n",
              " Text(0.8, 0.375, 'squared_error = 0.0\\nsamples = 1\\nvalue = 20.7'),\n",
              " Text(0.806015037593985, 0.4583333333333333, 'squared_error = -0.0\\nsamples = 1\\nvalue = 22.4'),\n",
              " Text(0.8120300751879699, 0.5416666666666666, 'squared_error = 0.0\\nsamples = 1\\nvalue = 19.2'),\n",
              " Text(0.8902255639097745, 0.7083333333333334, 'X[1] <= 35.3\\nsquared_error = 5.493\\nsamples = 26\\nvalue = 22.058'),\n",
              " Text(0.8300751879699249, 0.625, 'X[0] <= 219.1\\nsquared_error = 0.721\\nsamples = 11\\nvalue = 19.809'),\n",
              " Text(0.8240601503759398, 0.5416666666666666, 'squared_error = 0.0\\nsamples = 1\\nvalue = 18.0'),\n",
              " Text(0.8360902255639098, 0.5416666666666666, 'X[1] <= 27.6\\nsquared_error = 0.433\\nsamples = 10\\nvalue = 19.99'),\n",
              " Text(0.8195488721804511, 0.4583333333333333, 'X[0] <= 252.1\\nsquared_error = 0.18\\nsamples = 3\\nvalue = 19.2'),\n",
              " Text(0.8135338345864662, 0.375, 'squared_error = 0.0\\nsamples = 2\\nvalue = 18.9'),\n",
              " Text(0.825563909774436, 0.375, 'squared_error = 0.0\\nsamples = 1\\nvalue = 19.8'),\n",
              " Text(0.8526315789473684, 0.4583333333333333, 'X[0] <= 229.35\\nsquared_error = 0.159\\nsamples = 7\\nvalue = 20.329'),\n",
              " Text(0.837593984962406, 0.375, 'X[1] <= 33.35\\nsquared_error = 0.062\\nsamples = 2\\nvalue = 19.85'),\n",
              " Text(0.8315789473684211, 0.2916666666666667, 'squared_error = 0.0\\nsamples = 1\\nvalue = 20.1'),\n",
              " Text(0.843609022556391, 0.2916666666666667, 'squared_error = -0.0\\nsamples = 1\\nvalue = 19.6'),\n",
              " Text(0.8676691729323308, 0.375, 'X[0] <= 268.2\\nsquared_error = 0.07\\nsamples = 5\\nvalue = 20.52'),\n",
              " Text(0.8556390977443609, 0.2916666666666667, 'X[2] <= 10.6\\nsquared_error = 0.056\\nsamples = 3\\nvalue = 20.367'),\n",
              " Text(0.849624060150376, 0.20833333333333334, 'squared_error = 0.0\\nsamples = 1\\nvalue = 20.7'),\n",
              " Text(0.8616541353383459, 0.20833333333333334, 'X[2] <= 18.1\\nsquared_error = 0.0\\nsamples = 2\\nvalue = 20.2'),\n",
              " Text(0.8556390977443609, 0.125, 'squared_error = 0.0\\nsamples = 1\\nvalue = 20.2'),\n",
              " Text(0.8676691729323308, 0.125, 'squared_error = 0.0\\nsamples = 1\\nvalue = 20.2'),\n",
              " Text(0.8796992481203008, 0.2916666666666667, 'X[2] <= 30.75\\nsquared_error = 0.002\\nsamples = 2\\nvalue = 20.75'),\n",
              " Text(0.8736842105263158, 0.20833333333333334, 'squared_error = 0.0\\nsamples = 1\\nvalue = 20.7'),\n",
              " Text(0.8857142857142857, 0.20833333333333334, 'squared_error = 0.0\\nsamples = 1\\nvalue = 20.8'),\n",
              " Text(0.9503759398496241, 0.625, 'X[0] <= 258.9\\nsquared_error = 2.566\\nsamples = 15\\nvalue = 23.707'),\n",
              " Text(0.924812030075188, 0.5416666666666666, 'X[1] <= 46.45\\nsquared_error = 1.534\\nsamples = 10\\nvalue = 22.95'),\n",
              " Text(0.9097744360902256, 0.4583333333333333, 'X[2] <= 22.85\\nsquared_error = 0.273\\nsamples = 7\\nvalue = 22.229'),\n",
              " Text(0.9037593984962407, 0.375, 'squared_error = 0.0\\nsamples = 1\\nvalue = 23.2'),\n",
              " Text(0.9157894736842105, 0.375, 'X[2] <= 36.7\\nsquared_error = 0.136\\nsamples = 6\\nvalue = 22.067'),\n",
              " Text(0.9037593984962407, 0.2916666666666667, 'X[2] <= 29.6\\nsquared_error = 0.116\\nsamples = 3\\nvalue = 21.833'),\n",
              " Text(0.8977443609022556, 0.20833333333333334, 'squared_error = 0.0\\nsamples = 1\\nvalue = 22.3'),\n",
              " Text(0.9097744360902256, 0.20833333333333334, 'X[0] <= 220.75\\nsquared_error = 0.01\\nsamples = 2\\nvalue = 21.6'),\n",
              " Text(0.9037593984962407, 0.125, 'squared_error = 0.0\\nsamples = 1\\nvalue = 21.7'),\n",
              " Text(0.9157894736842105, 0.125, 'squared_error = -0.0\\nsamples = 1\\nvalue = 21.5'),\n",
              " Text(0.9278195488721804, 0.2916666666666667, 'X[2] <= 54.4\\nsquared_error = 0.047\\nsamples = 3\\nvalue = 22.3'),\n",
              " Text(0.9218045112781955, 0.20833333333333334, 'squared_error = 0.0\\nsamples = 1\\nvalue = 22.6'),\n",
              " Text(0.9338345864661655, 0.20833333333333334, 'X[1] <= 37.15\\nsquared_error = 0.002\\nsamples = 2\\nvalue = 22.15'),\n",
              " Text(0.9278195488721804, 0.125, 'squared_error = 0.0\\nsamples = 1\\nvalue = 22.2'),\n",
              " Text(0.9398496240601504, 0.125, 'squared_error = 0.0\\nsamples = 1\\nvalue = 22.1'),\n",
              " Text(0.9398496240601504, 0.4583333333333333, 'X[0] <= 215.55\\nsquared_error = 0.429\\nsamples = 3\\nvalue = 24.633'),\n",
              " Text(0.9338345864661655, 0.375, 'squared_error = 0.0\\nsamples = 1\\nvalue = 23.8'),\n",
              " Text(0.9458646616541353, 0.375, 'X[0] <= 231.75\\nsquared_error = 0.123\\nsamples = 2\\nvalue = 25.05'),\n",
              " Text(0.9398496240601504, 0.2916666666666667, 'squared_error = 0.0\\nsamples = 1\\nvalue = 24.7'),\n",
              " Text(0.9518796992481203, 0.2916666666666667, 'squared_error = 0.0\\nsamples = 1\\nvalue = 25.4'),\n",
              " Text(0.9759398496240601, 0.5416666666666666, 'X[1] <= 40.8\\nsquared_error = 1.194\\nsamples = 5\\nvalue = 25.22'),\n",
              " Text(0.9639097744360903, 0.4583333333333333, 'X[2] <= 78.35\\nsquared_error = 0.09\\nsamples = 2\\nvalue = 24.1'),\n",
              " Text(0.9578947368421052, 0.375, 'squared_error = 0.0\\nsamples = 1\\nvalue = 24.4'),\n",
              " Text(0.9699248120300752, 0.375, 'squared_error = -0.0\\nsamples = 1\\nvalue = 23.8'),\n",
              " Text(0.98796992481203, 0.4583333333333333, 'X[1] <= 46.35\\nsquared_error = 0.536\\nsamples = 3\\nvalue = 25.967'),\n",
              " Text(0.9819548872180451, 0.375, 'X[1] <= 42.9\\nsquared_error = 0.002\\nsamples = 2\\nvalue = 25.45'),\n",
              " Text(0.9759398496240601, 0.2916666666666667, 'squared_error = 0.0\\nsamples = 1\\nvalue = 25.5'),\n",
              " Text(0.98796992481203, 0.2916666666666667, 'squared_error = -0.0\\nsamples = 1\\nvalue = 25.4'),\n",
              " Text(0.9939849624060151, 0.375, 'squared_error = 0.0\\nsamples = 1\\nvalue = 27.0')]"
            ]
          },
          "metadata": {},
          "execution_count": 13
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x1080 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "simple_dt_reg = tree.DecisionTreeRegressor(random_state=0)\n",
        "\n",
        "scores = cross_val_score(simple_dt_reg, X_reg, y_reg, cv=5)\n",
        "print(\"%0.2f R^2 with a standard deviation of %0.2f\" % (scores.mean(), scores.std()))\n",
        "\n",
        "scores = cross_val_score(simple_dt_reg, X_reg, y_reg, cv=5, scoring='neg_mean_squared_error')\n",
        "print(\"%0.2f MSE with a standard deviation of %0.2f\" % (scores.mean(), scores.std()))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Ry9vRQTlZKq2",
        "outputId": "b11a2215-cc60-467f-f79b-c1bec260792d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0.95 R^2 with a standard deviation of 0.01\n",
            "-1.23 MSE with a standard deviation of 0.26\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.rcParams['figure.figsize'] = [8, 8]\n",
        "simple_dt_reg = tree.DecisionTreeRegressor(random_state=0)\n",
        "simple_dt_reg.fit(X_reg_train, y_reg_train)\n",
        "importances = simple_dt_reg.feature_importances_\n",
        "feature_names = [\"TV\", \"Radio\", \"Newspaper\"]\n",
        "forest_importances = pd.Series(importances, index=feature_names)\n",
        "#std = np.std([tree.feature_importances_ for tree in simple_dt_reg.estimators_], axis=0)\n",
        "fig, ax = plt.subplots()\n",
        "forest_importances.plot.bar( ax=ax)\n",
        "ax.set_title(\"Feature importances\")\n",
        "fig.tight_layout()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 585
        },
        "id": "-2Wb3_CEqXat",
        "outputId": "c3d1fcc6-6b43-4050-8a76-19b0bfad71c5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 576x576 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Классификация\n",
        "\n",
        "Аналогично задаче регрессии выполняется жадный алгоритм бинарного деления, но целевая функция соответсвует качественным данных.\n",
        "\n",
        "Варианты целевой функции:\n",
        "1. Gini index $G = \\sum_{k=1}^K \\hat{p}_{mk}(1-\\hat{p}_{mk})$\n",
        "2. Entropy $D = -\\sum_{k=1}^K \\hat{p}_{mk}\\log(\\hat{p}_{mk})$\n",
        "\n",
        "где $\\hat{p}_{mk}$ это отношение частоты наблюдений класса $k$ к общему числу наблюдений в области $m$.\n"
      ],
      "metadata": {
        "id": "Cd2PaHLWZ32U"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# https://scikit-learn.org/stable/modules/generated/sklearn.tree.DecisionTreeClassifier.html\n",
        "simple_dt_class = tree.DecisionTreeClassifier(random_state=0, max_depth=3)\n",
        "simple_dt_class.fit(X_class_train, y_class_train)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "X3bMnPMHZwP3",
        "outputId": "262f06b3-bce6-4aea-a1cc-b682ab40bec4"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "DecisionTreeClassifier(max_depth=3, random_state=0)"
            ]
          },
          "metadata": {},
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.rcParams['figure.figsize'] = [15, 15]\n",
        "tree.plot_tree(simple_dt_class, fontsize=10)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "S6My912WcOV1",
        "outputId": "19e07c77-1612-4464-84e4-67b3f3aa7d7e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[Text(0.625, 0.875, 'X[6] <= 5.615\\ngini = 0.664\\nsamples = 133\\nvalue = [44, 49, 40]'),\n",
              " Text(0.5, 0.625, 'X[0] <= 13.115\\ngini = 0.51\\nsamples = 85\\nvalue = [44, 1, 40]'),\n",
              " Text(0.25, 0.375, 'X[6] <= 4.826\\ngini = 0.227\\nsamples = 46\\nvalue = [6, 0, 40]'),\n",
              " Text(0.125, 0.125, 'gini = 0.278\\nsamples = 6\\nvalue = [5, 0, 1]'),\n",
              " Text(0.375, 0.125, 'gini = 0.049\\nsamples = 40\\nvalue = [1, 0, 39]'),\n",
              " Text(0.75, 0.375, 'X[4] <= 3.544\\ngini = 0.05\\nsamples = 39\\nvalue = [38, 1, 0]'),\n",
              " Text(0.625, 0.125, 'gini = 0.0\\nsamples = 37\\nvalue = [37, 0, 0]'),\n",
              " Text(0.875, 0.125, 'gini = 0.5\\nsamples = 2\\nvalue = [1, 1, 0]'),\n",
              " Text(0.75, 0.625, 'gini = 0.0\\nsamples = 48\\nvalue = [0, 48, 0]')]"
            ]
          },
          "metadata": {},
          "execution_count": 17
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x1080 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "simple_dt_class = tree.DecisionTreeClassifier(random_state=0)\n",
        "scores = cross_val_score(simple_dt_class, X_class, y_class, cv=5, scoring='roc_auc_ovr')\n",
        "print(\"Simple DT has %0.4f OvR AUC with a standard deviation of %0.2f\" % (scores.mean(), scores.std()))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "w7ixK22Hcc47",
        "outputId": "6f587233-226c-4b4e-8acc-b6f65d11d16c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Simple DT has 0.9135 OvR AUC with a standard deviation of 0.03\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.rcParams['figure.figsize'] = [8, 8]\n",
        "simple_dt_class = tree.DecisionTreeClassifier(random_state=0)\n",
        "simple_dt_class.fit(X_class_train, y_class_train)\n",
        "importances = simple_dt_class.feature_importances_\n",
        "feature_names = [\"area\", \"perimeter\", \"compactness\", \"length\", \"width\", \"asymmetry\", \"length_groove\"]\n",
        "forest_importances = pd.Series(importances, index=feature_names)\n",
        "#std = np.std([tree.feature_importances_ for tree in simple_dt_reg.estimators_], axis=0)\n",
        "fig, ax = plt.subplots()\n",
        "forest_importances.plot.bar( ax=ax)\n",
        "ax.set_title(\"Feature importances\")\n",
        "fig.tight_layout()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 585
        },
        "id": "UOvCQhJDt2Vb",
        "outputId": "af078d15-972e-467d-b11c-cbfdc599829a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 576x576 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Ансабли моделей"
      ],
      "metadata": {
        "id": "AgBq6oZhdZHf"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# https://scikit-learn.org/stable/modules/ensemble.html"
      ],
      "metadata": {
        "id": "hPvjJhqUfrhp"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Bagging\n"
      ],
      "metadata": {
        "id": "JgCcijQlfftr"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Алгоритм:\n",
        "\n",
        "1. Симулировать $B$ различны тренирововчных данных (на основе ECDF).\n",
        "2. Построить $B$ моделей $\\hat{f}^1(x),\\hat{f}^2(x),...,\\hat{f}^B(x)$ с использование $B$ наборов данных методом бинарного деления.\n",
        "3. Взять среднее значение (для регрессии) или принять класс большинством предсказаний (классификация): $\\hat{f}^{bag}(x)=\\frac{1}{B}\\sum_{b=1}^B\\hat{f}^b(x)$\n",
        "\n"
      ],
      "metadata": {
        "id": "PcioAa3T6ZqX"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Регрессия"
      ],
      "metadata": {
        "id": "IxPQszz9f8D9"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.BaggingRegressor.html\n",
        "bagging_reg = BaggingRegressor(n_estimators=10, random_state=0)\n",
        "scores = cross_val_score(bagging_reg, X_reg, y_reg, cv=5)\n",
        "print(\"%0.2f R^2 with a standard deviation of %0.2f\" % (scores.mean(), scores.std()))\n",
        "\n",
        "scores = cross_val_score(bagging_reg, X_reg, y_reg, cv=5, scoring='neg_mean_squared_error')\n",
        "print(\"%0.2f MSE with a standard deviation of %0.2f\" % (scores.mean(), scores.std()))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Xv6jGaQbgee3",
        "outputId": "f95991b4-e78b-424d-9294-97af974bb75e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0.97 R^2 with a standard deviation of 0.01\n",
            "-0.84 MSE with a standard deviation of 0.28\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Классификация"
      ],
      "metadata": {
        "id": "gRMrVl5Hgf93"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.BaggingClassifier.html\n",
        "bagging_class = BaggingClassifier(random_state=0)\n",
        "scores = cross_val_score(bagging_class, X_class, y_class, cv=5, scoring='roc_auc_ovr')\n",
        "print(\"bagging has %0.4f OvR AUC with a standard deviation of %0.2f\" % (scores.mean(), scores.std()))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "eEfj_ADGf55i",
        "outputId": "c54d5a85-b718-4fe7-9a6a-61e0f51fe029"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "bagging has 0.9814 OvR AUC with a standard deviation of 0.01\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Random forest\n"
      ],
      "metadata": {
        "id": "3IxqI6VifSb-"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Алгоритм:\n",
        "\n",
        "1. Симулировать $B$ различных тренировочных данных (на основе ECDF).\n",
        "2. Построить $B$ моделей $\\hat{f}^1(x),\\hat{f}^2(x),...,\\hat{f}^B(x)$ с использование $B$ наборов данных методом бинарного деления. **Для каждого дерева выбираются $m$ предикторов $m \\leq p$**.\n",
        "3. Взять среднее значение (для регрессии) или принять класс большинством предсказаний (классификация): $\\hat{f}^{bag}(x)=\\frac{1}{B}\\sum_{b=1}^B\\hat{f}^b(x)$"
      ],
      "metadata": {
        "id": "k9ohp-QvBwVv"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Регрессия\n"
      ],
      "metadata": {
        "id": "29WRwXwEicqR"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.RandomForestRegressor.html#sklearn.ensemble.RandomForestRegressor\n",
        "random_forest_reg = RandomForestRegressor(random_state=0)\n",
        "scores = cross_val_score(random_forest_reg, X_reg, y_reg, cv=5)\n",
        "print(\"%0.2f R^2 with a standard deviation of %0.2f\" % (scores.mean(), scores.std()))\n",
        "\n",
        "scores = cross_val_score(random_forest_reg, X_reg, y_reg, cv=5, scoring='neg_mean_squared_error')\n",
        "print(\"%0.2f MSE with a standard deviation of %0.2f\" % (scores.mean(), scores.std()))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6HcICOj6ijMH",
        "outputId": "0086baad-c1b0-41f7-c0ad-37b23070d07e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0.98 R^2 with a standard deviation of 0.01\n",
            "-0.64 MSE with a standard deviation of 0.23\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.rcParams['figure.figsize'] = [8, 8]\n",
        "random_forest_reg = RandomForestRegressor(random_state=0)\n",
        "random_forest_reg.fit(X_reg_train, y_reg_train)\n",
        "importances = random_forest_reg.feature_importances_\n",
        "feature_names = [\"TV\", \"Radio\", \"Newspaper\"]\n",
        "forest_importances = pd.Series(importances, index=feature_names)\n",
        "std = np.std([tree.feature_importances_ for tree in random_forest_reg.estimators_], axis=0)\n",
        "fig, ax = plt.subplots()\n",
        "forest_importances.plot.bar(yerr=std,ax=ax)\n",
        "ax.set_title(\"Feature importances\")\n",
        "fig.tight_layout()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 585
        },
        "id": "VmqEKMjXunks",
        "outputId": "2e3166cf-b5ae-4d41-fe52-51b9fc9c0460"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 576x576 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Классификация"
      ],
      "metadata": {
        "id": "iS0NxlALigUc"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.RandomForestClassifier.html#sklearn.ensemble.RandomForestClassifier\n",
        "random_forest_class = RandomForestClassifier(random_state=0)\n",
        "scores = cross_val_score(random_forest_class, X_class, y_class, cv=5, scoring='roc_auc_ovr')\n",
        "print(\"Random Forest has %0.4f OvR AUC with a standard deviation of %0.2f\" % (scores.mean(), scores.std()))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2K52DSRxfY3A",
        "outputId": "25093335-3e10-48a1-8b93-485d5f93a498"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Random Forest has 0.9832 OvR AUC with a standard deviation of 0.02\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.rcParams['figure.figsize'] = [8, 8]\n",
        "random_forest_class = RandomForestClassifier(random_state=0)\n",
        "random_forest_class.fit(X_class_train, y_class_train)\n",
        "importances = random_forest_class.feature_importances_\n",
        "feature_names = [\"area\", \"perimeter\", \"compactness\", \"length\", \"width\", \"asymmetry\", \"length_groove\"]\n",
        "forest_importances = pd.Series(importances, index=feature_names)\n",
        "std = np.std([tree.feature_importances_ for tree in random_forest_class.estimators_], axis=0)\n",
        "fig, ax = plt.subplots()\n",
        "forest_importances.plot.bar(yerr=std,ax=ax)\n",
        "ax.set_title(\"Feature importances\")\n",
        "fig.tight_layout()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 585
        },
        "id": "zI_rRXmMvNC_",
        "outputId": "b8c6a45e-1a78-4608-883c-655b4ce2ad59"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 576x576 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Boosting"
      ],
      "metadata": {
        "id": "vPB9QDX7fWM2"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Алгоритм:\n",
        "\n",
        "1. Инициализация $f_0(x)=\\mathrm{argmin}_\\gamma\\sum_{i=1}^N L(y_i,\\gamma)$\n",
        "2. Для $m=1,2,...,M$:\n",
        "\n",
        "(а) вычисление псевдо-невязок (pseudo-residuals): $r_{im}=-\\left[ \\frac{\\partial L(y_i, f(x_i))}{\\partial f(x_i)} \\right]_{f=f_{m-1}}$ для $i=1,2,...,N$\n",
        "\n",
        "(б) подбор дерева на наборе $\\{(x_i,r_{im}) \\}_{i=1}^N$ для заданного набора областей $R_{jm}$, $j=1,2,...,J_m$\n",
        "\n",
        "(в) вычислить множитель $\\gamma$:\n",
        "\n",
        "$\\gamma_{m}=\\mathrm{argmin}_\\gamma \\sum_{x_i \\in R_{jm}}L(y_i,f_{m-1}(x_i)+\\gamma)$\n",
        "\n",
        "г) обновить модель: $f_m(x)=f_{m-1}(x)+\\alpha\\sum_{j=1}^{J_m}\\gamma_m I(x \\in R_{jm})$"
      ],
      "metadata": {
        "id": "5VfPPsgwCx4P"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Регрессия"
      ],
      "metadata": {
        "id": "ceoFCPtAmL6u"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.GradientBoostingRegressor.html#sklearn.ensemble.GradientBoostingRegressor\n",
        "boosting_reg = GradientBoostingRegressor(random_state=0, criterion='squared_error', learning_rate=0.1)\n",
        "scores = cross_val_score(boosting_reg, X_reg, y_reg, cv=5)\n",
        "print(\"%0.2f R^2 with a standard deviation of %0.2f\" % (scores.mean(), scores.std()))\n",
        "\n",
        "scores = cross_val_score(boosting_reg, X_reg, y_reg, cv=5, scoring='neg_mean_squared_error')\n",
        "print(\"%0.2f MSE with a standard deviation of %0.2f\" % (scores.mean(), scores.std()))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "O_b-JNbNmShW",
        "outputId": "d9586316-436c-45d6-8749-0db6ea745136"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0.98 R^2 with a standard deviation of 0.01\n",
            "-0.59 MSE with a standard deviation of 0.24\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        " ### Классификация "
      ],
      "metadata": {
        "id": "YgMZnPFOmOVZ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.GradientBoostingClassifier.html#sklearn.ensemble.GradientBoostingClassifier\n",
        "boosting_class = GradientBoostingClassifier(random_state=0, n_estimators=1000, max_depth=3, learning_rate=0.9)\n",
        "scores = cross_val_score(boosting_class, X_class, y_class, cv=5, scoring='roc_auc_ovr')\n",
        "print(\"Boosting has %0.4f OvR AUC with a standard deviation of %0.2f\" % (scores.mean(), scores.std()))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Jxi9Vg4SdGq8",
        "outputId": "d968ee29-33e5-4c2c-ba51-4b2bbfdcaac9"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Boosting has 0.9842 OvR AUC with a standard deviation of 0.03\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Выбор модели\n"
      ],
      "metadata": {
        "id": "_br7bMzZprvW"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Регрессия\n",
        "\n"
      ],
      "metadata": {
        "id": "-CuS7pElpvUV"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Model Name      | $r^2$            | Mean Squared Error|\n",
        "----------------|------------------|-------------------|\n",
        "LR              |$0.89\\pm0.04$     |$3.07\\pm1.28$      |\n",
        "LR poly 2       |$0.98\\pm0.01$     |$0.44\\pm0.39$      |\n",
        "LR poly 3       |$\\bf0.99\\pm0.01$  |$\\bf0.31\\pm0.24$   |\n",
        "NN              |$0.91\\pm1.61$     |$1.86\\pm1.49$      |\n",
        "Simple DT       |$0.95\\pm0.1$      |$1.23 \\pm0.26$     |\n",
        "Bagging         |$0.97 \\pm0.1$      |$0.84 \\pm0.28$     |  \n",
        "Random Forest   |$0.98\\pm0.1$      |$0.64 \\pm0.23$     |  \n",
        "Boosting        |$0.98\\pm0.1$      |$0.59 \\pm0.24$     |  "
      ],
      "metadata": {
        "id": "I5x60DCFp52H"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Классификация\n",
        "\n",
        "```\n",
        "Logit has            0.9961 ± 0.01 OvR AUC\n",
        "LDA has              0.9937 ± 0.01 OvR AUC\n",
        "QDA has              0.9888 ± 0.02 OvR AUC\n",
        "Simple DT has        0.9135 ± 0.03 OvR AUC\n",
        "bagging has          0.9814 ± 0.01 OvR AUC\n",
        "Random Forest has    0.9832 ± 0.02 OvR AUC\n",
        "Boosting has         0.9842 ± 0.03 OvR AUC\n",
        "```"
      ],
      "metadata": {
        "id": "GYj6zH9Bxhrh"
      }
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "a7SEwgWqyzW8"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}