Copia_di_Unbalanced_2classes,_Hypoxia_CLASSIFICATION,_hourly.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/bethweirich/IS-ENES-group_work/blob/main/Copia_di_Unbalanced_2classes%2C_Hypoxia_CLASSIFICATION%2C_hourly.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Zcar6gl9BVGZ"
      },
      "source": [
        "  **HYPOXIA - VENICE LAGOON - \n",
        "  IMBALANCED CLASSIFICATION PROBLEM**\n",
        "  \n",
        "\n",
        "*   Severe Imbalance -->  1:120\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "osjyY4jEFaG_"
      },
      "outputs": [],
      "source": [
        "import pandas as pd\n",
        "# Use numpy to convert to arrays\n",
        "import numpy as np\n",
        "#Import Random Forest Model\n",
        "from sklearn.ensemble import RandomForestClassifier\n",
        "from sklearn import neighbors\n",
        "from sklearn.metrics import accuracy_score, confusion_matrix, classification_report\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np \n",
        "# Import train_test_split function\n",
        "\n",
        "from sklearn.model_selection import train_test_split, RandomizedSearchCV\n",
        "\n",
        "from sklearn.metrics import mean_squared_error,confusion_matrix,accuracy_score, roc_auc_score, f1_score\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "import joblib\n",
        "#Import scikit-learn metrics module for accuracy calculation\n",
        "from sklearn import metrics\n",
        "from datetime import date, datetime"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "id": "B5WYeyNwA478",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5886f4a8-0e71-47a5-8be8-4ac63ec50e2d"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/gdrive\n"
          ]
        }
      ],
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/gdrive')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "c01h6ToRz3oo"
      },
      "source": [
        "# ▶ *Upload the dataset*"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "id": "pRtT9jIlKbiR",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 852
        },
        "outputId": "30aaeab0-b3ec-4ef7-e9b8-02e936faa91e"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "        Unnamed: 0  Unnamed: 0.1                 date         x         y  \\\n",
              "0                0             0  2008-01-01 00:00:00  17.56205  50.34165   \n",
              "1                1             1  2008-01-01 01:00:00  17.56205  50.34165   \n",
              "2                2             2  2008-01-01 02:00:00  17.56205  50.34165   \n",
              "3                3             3  2008-01-01 03:00:00  17.56205  50.34165   \n",
              "4                4             4  2008-01-01 04:00:00  17.56205  50.34165   \n",
              "...            ...           ...                  ...       ...       ...   \n",
              "638737      740059         51284  2017-07-25 03:00:00  17.54618  50.11736   \n",
              "638738      740060         51285  2017-07-25 09:00:00  17.54618  50.11736   \n",
              "638739      740061         51286  2017-07-25 10:00:00  17.54618  50.11736   \n",
              "638740      740062         51287  2017-07-25 11:00:00  17.54618  50.11736   \n",
              "638741      740063         51288  2017-07-25 12:00:00  17.54618  50.11736   \n",
              "\n",
              "        IDSonda  precipitation   humidity      radsol      wind  \\\n",
              "0             1            0.0  95.183425    0.000000  1.029575   \n",
              "1             1            0.0  92.060950    0.000000  1.075490   \n",
              "2             1            0.0  92.214959    0.000000  1.288724   \n",
              "3             1            0.0  95.018843    0.000000  1.288724   \n",
              "4             1            0.0  94.845991    0.000000  1.442809   \n",
              "...         ...            ...        ...         ...       ...   \n",
              "638737       10            0.0  88.671165    0.000000  0.916847   \n",
              "638738       10            0.0  56.040026  634.850998  1.234913   \n",
              "638739       10            0.0  53.637678  751.019047  1.385282   \n",
              "638740       10            0.0  53.663792  827.589861  1.289285   \n",
              "638741       10            0.0  57.469212  876.664363  1.505550   \n",
              "\n",
              "        w_temperature   salinity          DO  chla  \n",
              "0            7.540900  32.591700  100.380000  1.00  \n",
              "1            7.246600  32.566500   94.500000  1.00  \n",
              "2            6.984600  32.592999   94.820002  1.05  \n",
              "3            6.156700  32.415400   99.239999  1.04  \n",
              "4            6.093600  32.406200   99.850002  1.00  \n",
              "...               ...        ...         ...   ...  \n",
              "638737      25.444500  26.997200   39.809999  0.70  \n",
              "638738      25.282701  24.244000   51.799999  0.59  \n",
              "638739      25.831100  24.966500   59.730000  0.65  \n",
              "638740      25.997000  26.298800   61.799999  0.65  \n",
              "638741      26.809000  27.982400   91.380002  0.55  \n",
              "\n",
              "[638742 rows x 14 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-49dec719-1162-4cf9-854d-052d5942f5d2\">\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>Unnamed: 0</th>\n",
              "      <th>Unnamed: 0.1</th>\n",
              "      <th>date</th>\n",
              "      <th>x</th>\n",
              "      <th>y</th>\n",
              "      <th>IDSonda</th>\n",
              "      <th>precipitation</th>\n",
              "      <th>humidity</th>\n",
              "      <th>radsol</th>\n",
              "      <th>wind</th>\n",
              "      <th>w_temperature</th>\n",
              "      <th>salinity</th>\n",
              "      <th>DO</th>\n",
              "      <th>chla</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>2008-01-01 00:00:00</td>\n",
              "      <td>17.56205</td>\n",
              "      <td>50.34165</td>\n",
              "      <td>1</td>\n",
              "      <td>0.0</td>\n",
              "      <td>95.183425</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.029575</td>\n",
              "      <td>7.540900</td>\n",
              "      <td>32.591700</td>\n",
              "      <td>100.380000</td>\n",
              "      <td>1.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>2008-01-01 01:00:00</td>\n",
              "      <td>17.56205</td>\n",
              "      <td>50.34165</td>\n",
              "      <td>1</td>\n",
              "      <td>0.0</td>\n",
              "      <td>92.060950</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.075490</td>\n",
              "      <td>7.246600</td>\n",
              "      <td>32.566500</td>\n",
              "      <td>94.500000</td>\n",
              "      <td>1.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2008-01-01 02:00:00</td>\n",
              "      <td>17.56205</td>\n",
              "      <td>50.34165</td>\n",
              "      <td>1</td>\n",
              "      <td>0.0</td>\n",
              "      <td>92.214959</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.288724</td>\n",
              "      <td>6.984600</td>\n",
              "      <td>32.592999</td>\n",
              "      <td>94.820002</td>\n",
              "      <td>1.05</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>2008-01-01 03:00:00</td>\n",
              "      <td>17.56205</td>\n",
              "      <td>50.34165</td>\n",
              "      <td>1</td>\n",
              "      <td>0.0</td>\n",
              "      <td>95.018843</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.288724</td>\n",
              "      <td>6.156700</td>\n",
              "      <td>32.415400</td>\n",
              "      <td>99.239999</td>\n",
              "      <td>1.04</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>2008-01-01 04:00:00</td>\n",
              "      <td>17.56205</td>\n",
              "      <td>50.34165</td>\n",
              "      <td>1</td>\n",
              "      <td>0.0</td>\n",
              "      <td>94.845991</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.442809</td>\n",
              "      <td>6.093600</td>\n",
              "      <td>32.406200</td>\n",
              "      <td>99.850002</td>\n",
              "      <td>1.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>638737</th>\n",
              "      <td>740059</td>\n",
              "      <td>51284</td>\n",
              "      <td>2017-07-25 03:00:00</td>\n",
              "      <td>17.54618</td>\n",
              "      <td>50.11736</td>\n",
              "      <td>10</td>\n",
              "      <td>0.0</td>\n",
              "      <td>88.671165</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.916847</td>\n",
              "      <td>25.444500</td>\n",
              "      <td>26.997200</td>\n",
              "      <td>39.809999</td>\n",
              "      <td>0.70</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>638738</th>\n",
              "      <td>740060</td>\n",
              "      <td>51285</td>\n",
              "      <td>2017-07-25 09:00:00</td>\n",
              "      <td>17.54618</td>\n",
              "      <td>50.11736</td>\n",
              "      <td>10</td>\n",
              "      <td>0.0</td>\n",
              "      <td>56.040026</td>\n",
              "      <td>634.850998</td>\n",
              "      <td>1.234913</td>\n",
              "      <td>25.282701</td>\n",
              "      <td>24.244000</td>\n",
              "      <td>51.799999</td>\n",
              "      <td>0.59</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>638739</th>\n",
              "      <td>740061</td>\n",
              "      <td>51286</td>\n",
              "      <td>2017-07-25 10:00:00</td>\n",
              "      <td>17.54618</td>\n",
              "      <td>50.11736</td>\n",
              "      <td>10</td>\n",
              "      <td>0.0</td>\n",
              "      <td>53.637678</td>\n",
              "      <td>751.019047</td>\n",
              "      <td>1.385282</td>\n",
              "      <td>25.831100</td>\n",
              "      <td>24.966500</td>\n",
              "      <td>59.730000</td>\n",
              "      <td>0.65</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>638740</th>\n",
              "      <td>740062</td>\n",
              "      <td>51287</td>\n",
              "      <td>2017-07-25 11:00:00</td>\n",
              "      <td>17.54618</td>\n",
              "      <td>50.11736</td>\n",
              "      <td>10</td>\n",
              "      <td>0.0</td>\n",
              "      <td>53.663792</td>\n",
              "      <td>827.589861</td>\n",
              "      <td>1.289285</td>\n",
              "      <td>25.997000</td>\n",
              "      <td>26.298800</td>\n",
              "      <td>61.799999</td>\n",
              "      <td>0.65</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>638741</th>\n",
              "      <td>740063</td>\n",
              "      <td>51288</td>\n",
              "      <td>2017-07-25 12:00:00</td>\n",
              "      <td>17.54618</td>\n",
              "      <td>50.11736</td>\n",
              "      <td>10</td>\n",
              "      <td>0.0</td>\n",
              "      <td>57.469212</td>\n",
              "      <td>876.664363</td>\n",
              "      <td>1.505550</td>\n",
              "      <td>26.809000</td>\n",
              "      <td>27.982400</td>\n",
              "      <td>91.380002</td>\n",
              "      <td>0.55</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>638742 rows × 14 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-49dec719-1162-4cf9-854d-052d5942f5d2')\"\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-49dec719-1162-4cf9-854d-052d5942f5d2 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-49dec719-1162-4cf9-854d-052d5942f5d2');\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
        }
      ],
      "source": [
        "path_stations = '/content/gdrive/My Drive/Dati/DatiOriginali/Lagoon_WaterQuality/Elaborati/Dataset_hypoxia/Hourly/'\n",
        "\n",
        "features = pd.read_excel(path_stations + 'hypoxia_dataset_2008-2019_hourly.xlsx')\n",
        "features"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# selecting rows based on condition\n",
        "#features =  features.loc[features['IDSonda'] < 2]  #only station 1"
      ],
      "metadata": {
        "id": "Wua42YVDWwK1"
      },
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "wk37xbUv4nrc",
        "outputId": "477dace3-8c22-493d-ba58-9628637ed689"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "array([[<matplotlib.axes._subplots.AxesSubplot object at 0x7f707e837310>,\n",
              "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f70819d5a50>,\n",
              "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f70819b0990>],\n",
              "       [<matplotlib.axes._subplots.AxesSubplot object at 0x7f707dcd1690>,\n",
              "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f707e98dc90>,\n",
              "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f70816cf2d0>],\n",
              "       [<matplotlib.axes._subplots.AxesSubplot object at 0x7f70816cb950>,\n",
              "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f708162ae90>,\n",
              "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f708162aed0>],\n",
              "       [<matplotlib.axes._subplots.AxesSubplot object at 0x7f7081450610>,\n",
              "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f7080f47150>,\n",
              "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f7080ee3750>]],\n",
              "      dtype=object)"
            ]
          },
          "execution_count": 5,
          "metadata": {},
          "output_type": "execute_result"
        },
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1080x1080 with 12 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "# histogram of all the variables\n",
        "features.hist(figsize=(15, 15))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 300
        },
        "id": "P7tJjn-45G6P",
        "outputId": "fe346664-b513-40bd-a499-def93dd1480f"
      },
      "outputs": [
        {
          "data": {
            "text/html": [
              "\n",
              "  <div id=\"df-ba8be01b-b820-489b-866d-80117fc3915d\">\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>Unnamed: 0</th>\n",
              "      <th>x</th>\n",
              "      <th>y</th>\n",
              "      <th>IDSonda</th>\n",
              "      <th>precipitation</th>\n",
              "      <th>humidity</th>\n",
              "      <th>radsol</th>\n",
              "      <th>wind</th>\n",
              "      <th>w_temperature</th>\n",
              "      <th>salinity</th>\n",
              "      <th>DO</th>\n",
              "      <th>chla</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>638742.000000</td>\n",
              "      <td>638742.000000</td>\n",
              "      <td>638742.000000</td>\n",
              "      <td>638742.000000</td>\n",
              "      <td>638742.000000</td>\n",
              "      <td>638742.000000</td>\n",
              "      <td>638742.000000</td>\n",
              "      <td>638742.000000</td>\n",
              "      <td>638742.000000</td>\n",
              "      <td>638742.000000</td>\n",
              "      <td>638742.000000</td>\n",
              "      <td>638742.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>37494.216930</td>\n",
              "      <td>17.597623</td>\n",
              "      <td>50.338748</td>\n",
              "      <td>5.142428</td>\n",
              "      <td>0.099139</td>\n",
              "      <td>77.623325</td>\n",
              "      <td>160.110315</td>\n",
              "      <td>1.666503</td>\n",
              "      <td>16.645369</td>\n",
              "      <td>29.256460</td>\n",
              "      <td>93.043836</td>\n",
              "      <td>2.402620</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>23706.846901</td>\n",
              "      <td>0.060637</td>\n",
              "      <td>0.096977</td>\n",
              "      <td>2.888879</td>\n",
              "      <td>0.709700</td>\n",
              "      <td>16.588074</td>\n",
              "      <td>249.764984</td>\n",
              "      <td>1.165202</td>\n",
              "      <td>7.744990</td>\n",
              "      <td>4.383993</td>\n",
              "      <td>23.033406</td>\n",
              "      <td>3.671078</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>17.500080</td>\n",
              "      <td>50.117360</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>12.262226</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>-1.997900</td>\n",
              "      <td>0.410000</td>\n",
              "      <td>0.083000</td>\n",
              "      <td>0.010000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>17299.000000</td>\n",
              "      <td>17.562050</td>\n",
              "      <td>50.243910</td>\n",
              "      <td>3.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>65.763998</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.789737</td>\n",
              "      <td>9.748300</td>\n",
              "      <td>27.493800</td>\n",
              "      <td>81.039999</td>\n",
              "      <td>0.784000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>35466.000000</td>\n",
              "      <td>17.601710</td>\n",
              "      <td>50.362470</td>\n",
              "      <td>5.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>79.877731</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.478159</td>\n",
              "      <td>16.572505</td>\n",
              "      <td>30.181000</td>\n",
              "      <td>91.596351</td>\n",
              "      <td>1.330000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>55140.750000</td>\n",
              "      <td>17.614530</td>\n",
              "      <td>50.406550</td>\n",
              "      <td>8.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>91.817623</td>\n",
              "      <td>248.604809</td>\n",
              "      <td>2.264058</td>\n",
              "      <td>23.504900</td>\n",
              "      <td>32.178200</td>\n",
              "      <td>102.440225</td>\n",
              "      <td>2.544188</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>93648.000000</td>\n",
              "      <td>17.724810</td>\n",
              "      <td>50.450370</td>\n",
              "      <td>10.000000</td>\n",
              "      <td>59.100270</td>\n",
              "      <td>100.000000</td>\n",
              "      <td>1093.621593</td>\n",
              "      <td>14.358378</td>\n",
              "      <td>34.123301</td>\n",
              "      <td>38.623600</td>\n",
              "      <td>482.422700</td>\n",
              "      <td>123.653999</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-ba8be01b-b820-489b-866d-80117fc3915d')\"\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-ba8be01b-b820-489b-866d-80117fc3915d 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-ba8be01b-b820-489b-866d-80117fc3915d');\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",
              "  "
            ],
            "text/plain": [
              "          Unnamed: 0              x              y        IDSonda  \\\n",
              "count  638742.000000  638742.000000  638742.000000  638742.000000   \n",
              "mean    37494.216930      17.597623      50.338748       5.142428   \n",
              "std     23706.846901       0.060637       0.096977       2.888879   \n",
              "min         0.000000      17.500080      50.117360       1.000000   \n",
              "25%     17299.000000      17.562050      50.243910       3.000000   \n",
              "50%     35466.000000      17.601710      50.362470       5.000000   \n",
              "75%     55140.750000      17.614530      50.406550       8.000000   \n",
              "max     93648.000000      17.724810      50.450370      10.000000   \n",
              "\n",
              "       precipitation       humidity         radsol           wind  \\\n",
              "count  638742.000000  638742.000000  638742.000000  638742.000000   \n",
              "mean        0.099139      77.623325     160.110315       1.666503   \n",
              "std         0.709700      16.588074     249.764984       1.165202   \n",
              "min         0.000000      12.262226       0.000000       0.000000   \n",
              "25%         0.000000      65.763998       0.000000       0.789737   \n",
              "50%         0.000000      79.877731       0.000000       1.478159   \n",
              "75%         0.000000      91.817623     248.604809       2.264058   \n",
              "max        59.100270     100.000000    1093.621593      14.358378   \n",
              "\n",
              "       w_temperature       salinity             DO           chla  \n",
              "count  638742.000000  638742.000000  638742.000000  638742.000000  \n",
              "mean       16.645369      29.256460      93.043836       2.402620  \n",
              "std         7.744990       4.383993      23.033406       3.671078  \n",
              "min        -1.997900       0.410000       0.083000       0.010000  \n",
              "25%         9.748300      27.493800      81.039999       0.784000  \n",
              "50%        16.572505      30.181000      91.596351       1.330000  \n",
              "75%        23.504900      32.178200     102.440225       2.544188  \n",
              "max        34.123301      38.623600     482.422700     123.653999  "
            ]
          },
          "execution_count": 6,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "features.describe()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LjD8kjoWCDQ3"
      },
      "source": [
        "# ▶ Classes of DO"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "id": "nQCBQ5MPKI4R",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "4cffdde8-05e0-427a-b90b-8974d920d041"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/ipykernel_launcher.py:6: SettingWithCopyWarning: \n",
            "A value is trying to be set on a copy of a slice from a DataFrame.\n",
            "Try using .loc[row_indexer,col_indexer] = value instead\n",
            "\n",
            "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
            "  \n",
            "/usr/local/lib/python3.7/dist-packages/ipykernel_launcher.py:7: SettingWithCopyWarning: \n",
            "A value is trying to be set on a copy of a slice from a DataFrame.\n",
            "Try using .loc[row_indexer,col_indexer] = value instead\n",
            "\n",
            "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
            "  import sys\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "       Unnamed: 0  Unnamed: 0.1                 date         x         y  \\\n",
              "2678         2688          2688  2008-05-03 06:00:00  17.56205  50.34165   \n",
              "2679         2689          2689  2008-05-03 07:00:00  17.56205  50.34165   \n",
              "2702         2712          2712  2008-05-04 07:00:00  17.56205  50.34165   \n",
              "2792         2802          2802  2008-05-08 07:00:00  17.56205  50.34165   \n",
              "2954         2964          2964  2008-05-15 04:00:00  17.56205  50.34165   \n",
              "...           ...           ...                  ...       ...       ...   \n",
              "85043       93644         93644  2019-12-12 12:00:00  17.56205  50.34165   \n",
              "85044       93645         93645  2019-12-12 13:00:00  17.56205  50.34165   \n",
              "85045       93646         93646  2019-12-12 14:00:00  17.56205  50.34165   \n",
              "85046       93647         93647  2019-12-12 15:00:00  17.56205  50.34165   \n",
              "85047       93648         93648  2019-12-12 16:00:00  17.56205  50.34165   \n",
              "\n",
              "       IDSonda  precipitation   humidity      radsol      wind  w_temperature  \\\n",
              "2678         1       0.000000  91.423767   96.921936  1.091830      18.936101   \n",
              "2679         1       0.000000  91.676986  237.833783  1.175490      18.959999   \n",
              "2702         1       0.000000  68.654206   79.858013  1.980554      19.298900   \n",
              "2792         1       0.000000  85.735148  125.314169  0.650979      20.829000   \n",
              "2954         1       0.000000  93.373390    0.000000  0.834639      22.987400   \n",
              "...        ...            ...        ...         ...       ...            ...   \n",
              "85043        1       1.375802  94.149193   21.922295  0.967645       8.207500   \n",
              "85044        1       1.706694  99.786218   41.228411  1.710613       8.258100   \n",
              "85045        1       0.856190  98.120992   33.729587  0.941293       8.477950   \n",
              "85046        1       0.298946  97.395269   26.230763  0.603919       8.925550   \n",
              "85047        1       0.787901  93.054051   13.153058  1.517072       9.518700   \n",
              "\n",
              "       salinity  DO    chla  \n",
              "2678   30.52850   0  1.1000  \n",
              "2679   30.57910   0  1.1300  \n",
              "2702   30.59900   0  1.1000  \n",
              "2792   29.06000   0  0.4700  \n",
              "2954   31.49560   0  0.7000  \n",
              "...         ...  ..     ...  \n",
              "85043  30.23945   1  0.9135  \n",
              "85044  30.11800   1  0.9465  \n",
              "85045  30.30360   1  0.8835  \n",
              "85046  30.61990   1  0.9225  \n",
              "85047  30.95920   1  1.1100  \n",
              "\n",
              "[85048 rows x 14 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-8a647f6b-dd6c-4be8-87b8-d2bb3de4b33e\">\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>Unnamed: 0</th>\n",
              "      <th>Unnamed: 0.1</th>\n",
              "      <th>date</th>\n",
              "      <th>x</th>\n",
              "      <th>y</th>\n",
              "      <th>IDSonda</th>\n",
              "      <th>precipitation</th>\n",
              "      <th>humidity</th>\n",
              "      <th>radsol</th>\n",
              "      <th>wind</th>\n",
              "      <th>w_temperature</th>\n",
              "      <th>salinity</th>\n",
              "      <th>DO</th>\n",
              "      <th>chla</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>2678</th>\n",
              "      <td>2688</td>\n",
              "      <td>2688</td>\n",
              "      <td>2008-05-03 06:00:00</td>\n",
              "      <td>17.56205</td>\n",
              "      <td>50.34165</td>\n",
              "      <td>1</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>91.423767</td>\n",
              "      <td>96.921936</td>\n",
              "      <td>1.091830</td>\n",
              "      <td>18.936101</td>\n",
              "      <td>30.52850</td>\n",
              "      <td>0</td>\n",
              "      <td>1.1000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2679</th>\n",
              "      <td>2689</td>\n",
              "      <td>2689</td>\n",
              "      <td>2008-05-03 07:00:00</td>\n",
              "      <td>17.56205</td>\n",
              "      <td>50.34165</td>\n",
              "      <td>1</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>91.676986</td>\n",
              "      <td>237.833783</td>\n",
              "      <td>1.175490</td>\n",
              "      <td>18.959999</td>\n",
              "      <td>30.57910</td>\n",
              "      <td>0</td>\n",
              "      <td>1.1300</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2702</th>\n",
              "      <td>2712</td>\n",
              "      <td>2712</td>\n",
              "      <td>2008-05-04 07:00:00</td>\n",
              "      <td>17.56205</td>\n",
              "      <td>50.34165</td>\n",
              "      <td>1</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>68.654206</td>\n",
              "      <td>79.858013</td>\n",
              "      <td>1.980554</td>\n",
              "      <td>19.298900</td>\n",
              "      <td>30.59900</td>\n",
              "      <td>0</td>\n",
              "      <td>1.1000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2792</th>\n",
              "      <td>2802</td>\n",
              "      <td>2802</td>\n",
              "      <td>2008-05-08 07:00:00</td>\n",
              "      <td>17.56205</td>\n",
              "      <td>50.34165</td>\n",
              "      <td>1</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>85.735148</td>\n",
              "      <td>125.314169</td>\n",
              "      <td>0.650979</td>\n",
              "      <td>20.829000</td>\n",
              "      <td>29.06000</td>\n",
              "      <td>0</td>\n",
              "      <td>0.4700</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2954</th>\n",
              "      <td>2964</td>\n",
              "      <td>2964</td>\n",
              "      <td>2008-05-15 04:00:00</td>\n",
              "      <td>17.56205</td>\n",
              "      <td>50.34165</td>\n",
              "      <td>1</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>93.373390</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.834639</td>\n",
              "      <td>22.987400</td>\n",
              "      <td>31.49560</td>\n",
              "      <td>0</td>\n",
              "      <td>0.7000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>85043</th>\n",
              "      <td>93644</td>\n",
              "      <td>93644</td>\n",
              "      <td>2019-12-12 12:00:00</td>\n",
              "      <td>17.56205</td>\n",
              "      <td>50.34165</td>\n",
              "      <td>1</td>\n",
              "      <td>1.375802</td>\n",
              "      <td>94.149193</td>\n",
              "      <td>21.922295</td>\n",
              "      <td>0.967645</td>\n",
              "      <td>8.207500</td>\n",
              "      <td>30.23945</td>\n",
              "      <td>1</td>\n",
              "      <td>0.9135</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>85044</th>\n",
              "      <td>93645</td>\n",
              "      <td>93645</td>\n",
              "      <td>2019-12-12 13:00:00</td>\n",
              "      <td>17.56205</td>\n",
              "      <td>50.34165</td>\n",
              "      <td>1</td>\n",
              "      <td>1.706694</td>\n",
              "      <td>99.786218</td>\n",
              "      <td>41.228411</td>\n",
              "      <td>1.710613</td>\n",
              "      <td>8.258100</td>\n",
              "      <td>30.11800</td>\n",
              "      <td>1</td>\n",
              "      <td>0.9465</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>85045</th>\n",
              "      <td>93646</td>\n",
              "      <td>93646</td>\n",
              "      <td>2019-12-12 14:00:00</td>\n",
              "      <td>17.56205</td>\n",
              "      <td>50.34165</td>\n",
              "      <td>1</td>\n",
              "      <td>0.856190</td>\n",
              "      <td>98.120992</td>\n",
              "      <td>33.729587</td>\n",
              "      <td>0.941293</td>\n",
              "      <td>8.477950</td>\n",
              "      <td>30.30360</td>\n",
              "      <td>1</td>\n",
              "      <td>0.8835</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>85046</th>\n",
              "      <td>93647</td>\n",
              "      <td>93647</td>\n",
              "      <td>2019-12-12 15:00:00</td>\n",
              "      <td>17.56205</td>\n",
              "      <td>50.34165</td>\n",
              "      <td>1</td>\n",
              "      <td>0.298946</td>\n",
              "      <td>97.395269</td>\n",
              "      <td>26.230763</td>\n",
              "      <td>0.603919</td>\n",
              "      <td>8.925550</td>\n",
              "      <td>30.61990</td>\n",
              "      <td>1</td>\n",
              "      <td>0.9225</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>85047</th>\n",
              "      <td>93648</td>\n",
              "      <td>93648</td>\n",
              "      <td>2019-12-12 16:00:00</td>\n",
              "      <td>17.56205</td>\n",
              "      <td>50.34165</td>\n",
              "      <td>1</td>\n",
              "      <td>0.787901</td>\n",
              "      <td>93.054051</td>\n",
              "      <td>13.153058</td>\n",
              "      <td>1.517072</td>\n",
              "      <td>9.518700</td>\n",
              "      <td>30.95920</td>\n",
              "      <td>1</td>\n",
              "      <td>1.1100</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>85048 rows × 14 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-8a647f6b-dd6c-4be8-87b8-d2bb3de4b33e')\"\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-8a647f6b-dd6c-4be8-87b8-d2bb3de4b33e 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-8a647f6b-dd6c-4be8-87b8-d2bb3de4b33e');\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": 5
        }
      ],
      "source": [
        "#HYPOXIA < 60 % SAT     #https://www.sincem.unibo.it/images/tesi/Tesi_bianchi.pdf   https://nmsfarallones.blob.core.windows.net/farallones-prod/media/archive/manage/pdf/sac/16_08/SAC_HypoxiaInSanctuaries_JLargier2.pdf\n",
        "low = features[features.DO <= 60]\n",
        "high = features[features.DO > 60]\n",
        "\n",
        "#assigns label to the classes\n",
        "low[\"DO\"] = 0 \n",
        "high[\"DO\"] = 1\n",
        "\n",
        "#feature_class = pd.concat([primo, secondo, terzo, quarto])\n",
        "feature_class = pd.concat([low, high])\n",
        "feature_class"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jJDzVT2OX9IK"
      },
      "source": [
        "# ▶ Data preparation\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "NZz8YC9jPd8x",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "f678681a-7d3c-45b7-9988-de1ebaa7d084"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Index(['Unnamed: 0', 'Unnamed: 0.1', 'date', 'x', 'y', 'IDSonda',\n",
              "       'precipitation', 'humidity', 'radsol', 'wind', 'w_temperature',\n",
              "       'salinity', 'DO', 'chla'],\n",
              "      dtype='object')"
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ],
      "source": [
        "features.columns"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "id": "TKmSuaSDN3v4",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "2340f037-5c0a-4560-f3e5-bf9108528c39"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Dataset shape (85048, 14)\n",
            "X_t shape (85048, 10)\n",
            "y shape (85048,)\n",
            "Etichette:  2678     0\n",
            "2679     0\n",
            "2702     0\n",
            "2792     0\n",
            "2954     0\n",
            "        ..\n",
            "85043    1\n",
            "85044    1\n",
            "85045    1\n",
            "85046    1\n",
            "85047    1\n",
            "Name: DO, Length: 85048, dtype: int64\n"
          ]
        }
      ],
      "source": [
        "# data preparation\n",
        "\n",
        "dataset = feature_class  #feature_class\n",
        "print(\"Dataset shape\", dataset.shape)\n",
        "\n",
        "# predictors\n",
        "X_t = dataset.drop(labels=['DO', 'date', 'Unnamed: 0', 'Unnamed: 0.1'], axis=1) #'STATION' 'ODmol', ,'Data'\n",
        "print(\"X_t shape\", X_t.shape)\n",
        "\n",
        "# target variable\n",
        "y = dataset['DO']  #ODmol\n",
        "\n",
        "print(\"y shape\", y.shape)\n",
        "print(\"Etichette: \", y)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MRDdE9ZrQnvV",
        "outputId": "52527b07-2941-40d1-bc23-ae355098d650"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "1    77466\n",
              "0     7582\n",
              "Name: DO, dtype: int64"
            ]
          },
          "metadata": {},
          "execution_count": 7
        }
      ],
      "source": [
        "y.value_counts()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "id": "p1sH70wsQxeG",
        "outputId": "61061acc-1f8c-4564-f052-c04c7ea8ece0"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f9c8a4cb490>"
            ]
          },
          "metadata": {},
          "execution_count": 8
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "y.value_counts().plot.pie(autopct='%.2f', colors = ['lightblue', 'orange'] )"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5Cen-kKwm5ci"
      },
      "source": [
        "# ▶ Divide training and testing\n",
        "*   Keep the test set separated!\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "metadata": {
        "id": "mQ6jXeqeYEna"
      },
      "outputs": [],
      "source": [
        "# Split dataset into training set and test set\n",
        "X_train, X_test, y_train, y_test = train_test_split(X_t, y, test_size=0.3, stratify=y) # 70% training and 30% test"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GfD5PH0aYTke"
      },
      "source": [
        "# UNDERSAMPLING TRIALS"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### *Random Undersampling*"
      ],
      "metadata": {
        "id": "Bch1N3uf-Rdc"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "7XW_ydpSYaZn"
      },
      "outputs": [],
      "source": [
        "'''\n",
        "from imblearn.under_sampling import RandomUnderSampler\n",
        "\n",
        "rus = RandomUnderSampler(sampling_strategy=1)\n",
        "X_train_res, y_train_res = rus.fit_resample(X_train, y_train)\n",
        "\n",
        "ax = y_train_res.value_counts().plot.pie(autopct='%.2f')\n",
        "_ = ax.set_title(\"Under-sampling\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "hjhW4cvxZvOM"
      },
      "outputs": [],
      "source": [
        "#y_train_res.value_counts()"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### *Tomek Links Undersampling*"
      ],
      "metadata": {
        "id": "lym_MuxDXqyn"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "'''\n",
        "from imblearn.under_sampling import TomekLinks\n",
        "\n",
        "tl = TomekLinks(sampling_strategy=0)\n",
        "X_train_tl, y_train_tl, id_tl = tl.fit_sample(X_train, y_train)\n",
        "\n",
        "print('Removed indexes:', id_tl)\n",
        "\n",
        "plot_2d_space(X_train_tl, y_train_tl, 'Tomek links under-sampling')"
      ],
      "metadata": {
        "id": "z_HxSdvtXt47"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aaXgrFV-O4Ay"
      },
      "source": [
        "# OVERSALMPLING TRIALS\n",
        "\n",
        "https://www.kaggle.com/code/rafjaa/resampling-strategies-for-imbalanced-datasets/notebook\n"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### *Random Oversampling*"
      ],
      "metadata": {
        "id": "DQSD0J4k-g8F"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 264
        },
        "id": "42B89_pmdtQu",
        "outputId": "cbe917f7-c71e-4616-b4d4-4bd0218e0594"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "from imblearn.over_sampling import RandomOverSampler\n",
        "\n",
        "ros = RandomOverSampler(sampling_strategy=\"not majority\")\n",
        "X_train_res, y_train_res = ros.fit_resample(X_train, y_train)\n",
        "\n",
        "ax = y_train_res.value_counts().plot.pie(autopct='%.2f')\n",
        "_ = ax.set_title(\"Over-sampling\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "id": "Rm58ty9MMmyL"
      },
      "outputs": [],
      "source": [
        "'''\n",
        "from collections import Counter\n",
        "from sklearn.datasets import make_classification\n",
        "from imblearn.over_sampling import SMOTE \n",
        "from matplotlib import pyplot\n",
        "from numpy import where"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "metadata": {
        "id": "W7dNW2stcPKk"
      },
      "outputs": [],
      "source": [
        "'''\n",
        "oversample = SMOTE()\n",
        "X_train_res, y_train_res = oversample.fit_resample(X_train, y_train)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "'''\n",
        "ax = y_train_res.value_counts().plot.pie(autopct='%.2f')\n",
        "_ = ax.set_title(\"Over-sampling_SMOTE\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 264
        },
        "id": "lWtoLvy2GKOn",
        "outputId": "f3464c5f-5c1b-4146-bb64-4d3cbe71600d"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAPUAAAD3CAYAAADFeRJuAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAZ4ElEQVR4nO3deXhU1f3H8fc3CySQEBarIKiDorVu1Lpv1dZapXFprfsGIu611trasfqrV2vtaG2t9am1VC0uVXGrWMddi0orKohrFUGMsivbJIEQSHJ+f5wbGCchE8LMnLn3fl/PM08yd+7M/U5yP3POPXMXMcaglAqPEtcFKKVyS0OtVMhoqJUKGQ21UiGjoVYqZDTUSoWMhjriRCQmIkZEyvz7T4nIaNd1qZ7TUAMiMkZE3hWRVSKySET+IiL9XdflgjFmlDHmrnwuQ0TOEpEPRaRBRBaLyJMiUu0/NsH/kDkm4zk3+dPHpE0bJiL/EJGlIrJSRF4XkSP9x7YWkca0m/Hnab9/kL+sNRnzvZ3P914IkQ+1iFwKXA/8HKgB9gW2AZ4TkV45XE5Zrl4ryETkYOA64GRjTDXwNWBixmwfAWekPacMOAH4OG3aQGAKsAbYGdgMuAm4T0SOM8Z8Zoypar/5TxuZNu0Vf9oN6fMZY0bm/l0XVqRDLSL9gKuBi4wxTxtj1hpj6rArUAz4mYg0+StQ+3N2F5ElIlLu3x8rIh+IyHIReUZEtkmb14jIhSIyC5jVyfJHiMhLIpLyX3Ni2mM3i8hcEakXkekiclDaY56IPCQi9/qt3bsisoOIXC4in/vP+27a/JNF5Ld+S1YvIpPS31NGTZNFZJz/+xgRmSIiN/rv7xMRGZU273ARedmv4XkR+bOI3Jvlz74X8KoxZgaAMWaZMeYuY0xD2jz/Ag4UkQH+/SOAd4BFafNcAjQCZxljFhljmowx9wO/AX4vIpKljtCKdKiB/YEK4NH0icaYRuBJYFfgVeCHaQ+fAjxsjFnrdxF/CRwLfAV4Bbg/YxnfB/YBdupk+b8GngUGAMOAW9IeewP4OjAQuA94SEQq0h4/CrjHf+4M4Bns/3MocA3w14xlnQGMBYYALcCfOqmnM/sAM7Et4Q3AHWmBuQ94HRgEeMDp3Xi914DDReRqETlARHp3Ms9qYBJwUlrtd2fMcxjwiDGmLWP6g8DWwA7dqCWcjDGRvQGnAYs28FgCeA4YB7zoTxNgLvBN//5T2Jai/TklwCpgG/++Ab7dxfLvBsYDw7pR63Js9xFsgJ5Le+wobKtV6t+v9pfd378/GUikzb8Ttttaiu2RGKAsbd5x/u9jgNlpz+vjzzsYG5wWoE/a4/cC93bjvYzCtsYr/Lr/kFb7BOBa4EDsB2p/YDFQie1uj/Hnmw2c18lrV/g1HpAx3QAjMqZNwH6ArEi73eV6vdzUW9Rb6iXAZhvY3h3iP/4IsJ+IDAG+CbRhW2Sw2943i8gKEVkBLMMGf2ja68wF8Adm2gdj3vcfu8yf/3UReV9ExrY/SUR+5nfrU/5r12Bby3aL035vApYYY1rT7gNUpc0zN+33T4HyjNfbkHVdXmPMqrTX3RJYljYtcxkbZIx5yhhzFLYXcgz2w2NcxjxTsL2fK4AnjDFNGS+zBPs/yjQk7fHuuNEY0z/tFviR/6iH+lWgGdt9XkdEqrCtyQvGmOXYLvKJ2K73A8b/mMeuxOdmrBSVxpj/pr2cbSaMecWsH4zZ2Z+2yBhztjFmS+Bc4FZ/O/sgbOBPAAYYY/oDKewHQE9tlfb71sBaur/id2YhMFBE+mxgGVkZY9qMMS8ALwK7dDLLvcCldOx6AzwPHCsimevwCdj/y0cbU0uYRDrUxpgUdqDsFhE5QkTKRSSG3S6bh91mBbvteAZwnP97u9uAy0VkZwARqRGR47u7fBE5XkSG+XeXYz8A2rDd5xbgC6BMRH4F9OvRm1zvNBHZyQ/hNdhxgdZsT9oQY8ynwDTAE5FeIrIfdjOgSyJyjIicJCIDxNobOBiY2snsf8JuO7/cyWM3YXsvd4jIYBGpEJGTsS37z9M+eCMn0qEGMMbcgB3suhGoxw7kzAUONcY0+7M9DmyP3f5+O+25/8R+HfaAiNQD72Fb+O7aC3hNRBr9ZVxsjJmDHfR6GtvafIrd7utW17YL92C3IRdhtzt/vImvB3AqsB+wFLsdPBHb8+nKcuBs7LcB9djW+HfGmH9kzmjsyPgLnQXUGLMUu91dAfzPr+GnwOnGmMyvyLpyWcb31JvSeykKEuEPtMgQkcnYAazb87ycicCHxpir8rkc1bXIt9Sq50RkLxHZTkRKROQI7KDXY67rijoNtdoUg7FfgTVit3/PN8bMEJFTM7q0maP+Ko+0+61UyGhLrVTIaKiVChkNtVIho6FWKmQ01EqFjIZaqZDRUCsVMhpqpUJGQ61UyGiolQoZDbVaR0Tu9E9c+J7rWlTPaahVugnYM3eqANNQq3WMMS9jz7OmAkxDrVTIaKiVChkNtVIho6FWKmQ01GodEbkfey70r4rIPBE5y3VNauPp6YyUChltqZUKGQ21UiGjoVYqZDTUSoVMZ5dwVQEXiyd7YS/p2n7bMuPnFtjrPZel3cBelK8Fe0XMldjrbi0EFnTyc3FdorbHF9hT+aOj3wEXiycHAN8A9gD29H8OZ9Mue9sdrcBM7JUvp/u3t+oStSvzvFyVhYY6QGLxZCn2KpP7sz7A2zot6svagA9ZH/RX6hK1M9yWFD0a6iIXiyersYdDHgV8DxjktqKNNhd4Anup3hfrErVrHNcTehrqIhSLJ7cCjvZvhwC9nBaUOw3As9iAJ+sStUsd1xNKGuoiEYsnK4FTgHOxF6MPu1bgReA2YJIOuuWOhtqxWDy5PXABMBoY4LgcV+YBfwPG1yVqF7kuJug01A74A15HY8N8KPkfqQ6KtcA/gVvrErUvuS4mqDTUBRSLJyuAHwEXA8Mcl1Ps3gd+B9xTl6htc11MkGioC8Bvmc8EPGCo22oC5z3girpE7eOuCwkKDXWexeLJHwLXAju6riXg/gPE6xK1U1wXUuw01HkSiye/BSSAvV3XEjJJ4PK6RO27rgspVhrqHIvFk9sCtwKHu64lxNqAe4Cf1iVq9ZTGGTTUORKLJwW4ENs693VcTlQsBs6rS9Q+5rqQYqKhzgG/db4Du/eXKrz7gIu01bY01JtAW+eioq22T0PdQ9o6F63It9oa6h6IxZPHYS8mp61zcVoIHFuXqJ3quhAXNNQbwe9ue8D/obt2Frtm4Jy6RO3drgspNA11N8Xiyb7A3cCxrmtRG+UPwGVROgpMQ90NsXgyBkwCdnNciuqZZ4CT6hK1K1wXUgga6ixi8eTBwMPAZq5rUZvkI+DoukTtTNeF5JueIrgLsXhyHPAcGugw2AF4LRZPfsd1Ifmmod6AWDz5E+yB++Wua1E5UwM8EYsnj3JdSD5pqDsRiyfjwE2u61B50Rt4xD96LpQ01Bli8eRVwG9d16HyqhyYGIsnT3ZdSD7oQFmaWDz5S+A3rutQBdOKHRV/2HUhuaSh9sXiyUuw32mqaFkL/LAuUfsv14XkioYaiMWT52JPVauiqRk4si5R+7zrQnIh8qGOxZOHAU8Bpa5rUU6lgH3C8D12pEMdiydHAK8T3fNtqy/7CBvsQO95FtlQx+LJfsBU4Guua8mHeX8ZS0mvSigpQUpKGTL6j7Q2NbBk0vW01C+mrN8WbPb9OKUVVR2e2/juC6RefQCAmv1OomrXQwFoXjSbpcmbMC1rqNxuTwYceg4ioTuu5WlsVzyw+4pH8vrUsXiyBHvcbSgD3W6Lk6+jtE/Nuvv1Ux+iIjaSmn2PJzX1IeqnPsSAQ8780nNamxpI/ec+Bo/+I4iwaMLFVG6/D6UVVSx79s8MOuIiem35VT5/yGP1nOlUbrdnod9Wvh0B3ABc6rqQnorq99TXAbWuiyi0VbNfo+8uttXtu8uhrJrV8XDj1Z+8SUVsd0orqymtqKIitjur50ynpXEZbc1N9B66IyJC1S7f7vT5IfHTWDx5husieipyoY7Fk6cAv3BdR96J8PmDv2LhhItpeOtpAFpXrqCsaiAApX0H0Lqy46ZjS8NSSvut39W9tHoQLQ1LaW1YSln1oC9Nb20M9UUrx8fiyX1cF9ETkep+x+LJnbGnIAq9wadeT1n1ZrSuXMHiiVdSPujLV/kRET3LQ9d6A/+MxZO7BO3USJFpqWPxZBn2FEQVjkspiLJq29qW9u1Pnx32o3nBR5T27U9Lo10/WxqXUdK3fyfPG0Rr/ZJ199tb6PYWO316adWgDs8PmSHALa6L2FiRCTVwGRC6UZ3OtK1ZTVvzqnW/r/5kBr2+sg19RuzDyvdeAGDley/QZ0TH3mXF8G/QVDeD1tWNtK5upKluBhXDv0FZ1UBKelfSPP9DjDE0vvcifbYPZO90Y50Siye/77qIjRGJr7T8bvebQC/XtRTC2hWL+OLRa+2dtjb67nQwNfufSGtTPUsmJWip/4Kyfpuz2TFxSiuraV44i8a3nmLQqB8D0PjOs6RefQiAmv1OoGq3wwBoXjiLpU/6X2ltuwcDvnNeGL/S6swiYOegdMNDH2q/2/0qEWmlVd7cV5eoPdV1Ed0Rhe53ZLrdKq8C0w0PdUsdtW63yrtAdMND21L75+i+Aw20yp3BwO9dF5FNaEMNHAdEYnhWFdQZsXhyV9dFdCWUofYHx651XYcKpRLsbsZFK5ShBsZiTwmrVD4cGYsnD3RdxIaELtSxeLISuMp1HSr0Eq4L2JDQhRr4MbCl6yJU6B1QrOcPD9VXWrF4cgAwB+i4U7NSufceMLIuUdvmupB0YWupf4EGWhXOLsBprovIFJpQx+LJKuAC13WoyLnMdQGZQhNq4HSg2nURKnJ29q+MWjTCFOrzXRegIquoeoihGCiLxZMHAS+7rkNF1lpg67pE7SLXhUB4Wuqi+qRUkVMOnO26iHaBb6lj8eQWwGfogRvKrXlArBjOFx6GlnocGmjl3jDgaNdFQMBD7Z+U/xzXdSjlK4rB2kCHGnto5daui1DK9+1YPDnQdRFBD3VRdHeU8pVSBFd+0VArlVvO18nAjn7H4sntgNmu61AqQwOwWV2ido2rAoLcUjv/RFSqE9XAt1wWoKFWKvecrpuBDLV/3HTRnk5GRZ7TkycEMtTAKCJ2xU4VKFvF4smvu1p4UEN9kOsClMrC2Toa1FDv4boApbJwdqmnwIU6Fk+WA7u5rkOpLJw1PIELNfa8UL1dF6FUFjvG4sk+LhYcxFBr11sFQSngZLBMQ61U/jhZVzXUSuWPk8GyQIVaB8lUwGhL3Q0j0EEyFRw7xuLJ0kIvNGih1mtkqSApBTYv9EI11ErlV8HX2aCFeojrApTaSAVfZ4MWam2pVdBoS52FttQqaLSlzkJbahU02lJnoS21ChptqbMY7LoApTZS8YVaREaJyMsissS/vSQi3ytEcZ3o62i5SvVUwY/U6vKUQCJyNnAucBkwzZ+8J5AQkWHGmPF5rm+dWDyppy9SQVRe6AVmC8olwIHGmGVp014UkVHAFKBgoUbPSaaCqeDrbbbut2QEGgBjzNI81dMVDbUKooKvt9kWWC8iI40xb6dPFJGR2CsRFMybvc81/Wns8AGjVDFrQ+pheUGXmS3UlwKPi8jfgen+tD2B0cBp+Sws00BpaAWcX1FQqY1RgkkVfpldMMZMwV4utgQY499KgH39xwqppcDLUyoXCr7edvsCeSLyFQBjzBd5ragrXk0wr+anouwDvNROhVxgly21WJ6IfAHMBGaKyBci8qvClNfBakfLVaqnCr7OZhv9vgQ4ANjbGDPQGDMQ2x0/QEQuyXt1HS12sEylNsWiQi8wW6hPB042xnzSPsEYMwc7SHZGPgvbgIUOlqnUpij4Opst1OXGmCWZE/3t6oLvKQMscLBMpTZFwdfZbKFe08PH8kVbahU0BV9ns31PPVJE6juZLkBFHurJRltqFTQFX2e7DLUxpuCnN81CW2oVNEW3TV1sNNQqaDTUWWj3WwWJQUOd1Wyg1XURSnXTHLzU2kIvNFih9lKrgA9cl6FUN03PPkvuBSvU1rTssyhVFDTU3eTkD6VUDzhpgDTUSuXPmy4WGsRQv40Olqni9zFeaoWLBQcv1DpYpoLBWY8yeKG2dLBMFTsN9Ub6r+sClMrC2Toa1FAnsXvrKFWMlgCvulp4MEPtpRago+CqeD2Jl3I2mBvMUFuPuy5AqQ1wum4GOdT/cl2AUp1oBp5xWUBwQ+2l3gI+c12GUhkm46UaXRYQ3FBb2lqrYuN8szDooXb+B1Qqg/N1MuihngzoRfNUsXgdLzXPdRHBDrWXWgP83XUZSvn+6roACHqorb+gO6Io95YD97suAsIQai/1MfCs6zJU5E3ASzW5LgLCEGrrVtcFqEgz2B5jUQhLqJ8APnVdhIqs5/FSs1wX0S4cofZSbcB412WoyCqqnmI4Qm3djpvre6lom0uR7QQVnlB7qc+Be12XoSLnZpdHZHUmPKG2POwO9UoVwjzgz66LyBSuUHupuRThH1mFloeXWu26iEzhCrV1HdDZ5XeVyqUPgAmui+hM+ELtpZYCv3Ndhgq9K4ttW7pd+EJt3QQscl2ECq3X8FKPui5iQ8IZai+1Evi16zJUaMVdF9CVcIba+hv20rdK5dLTeKnJrovoSnhDba8LfA56BJfKnZXABa6LyCa8oQbwUv+miHa0V4H3C7zUJ66LyCbcobYuA4r+H6GK3r8psn28N0SMiUDv1Kv5FvACIK5LKZTYHxuo7i2UCpSVwLRzqljWZDjx4VXUrTDE+gsPHteHAZUd/yR3vbWGa1+xu9FfeVAvRn+9FwDTF7QyZlITTWsN39u+nJuP6I1IJP6kK4Fdg9BKQzRa6sh2w/89ug9vnVfFtHOqAEhMaebQ4WXMuqiKQ4eXkZjScY/aZU2Gq19q5rVxfXl9XF+ufqmZ5U32g//8ZBN/O6qCWRdVMWtZK0/Pbino+3EoEN3udtEItRX5bvikmS2MHlkOwOiR5Tw2s2Mon5ndwmHbljGwUhhQKRy2bRlPz25hYUMb9c2w77AyRIQzduvFYx9GItSB6Xa3i06o7XfXY4nIaLgIfPeeVewxvpHx021XenFjG0Oq7b98cJWwuLGtw/PmN7SxVc361WJYvxLmN7Qxv8EwrJ+kTRfmN4T+T9kInIWXCtQbLXNdQEF5qcl4Nb8iAjumTDmzL0P7lfD5yjYOu2cVO2725c9vESEam8M9ZoDRQep2t4tOS93OS10LPOi6jHwb2s/+azfvW8IPdizj9fmtbFFVwsIG2zovbGhj874d//1Dq0uYm1rfgs+rb2NodQlDq4V59SZtumFodag/Fa4p5l1BuxK9UFtnAm+6LiJfVq4xNDSbdb8/+3Eru2xeytE7lHHX22sBuOvttRzz1Y4dtcNHlPHsnBaWNxmWNxmendPC4SPKGFJdQr/eMHVeC8YY7n5nDcfsGNqO3iPA1a6L6KlofKXVGa9mK+ANYAvXpeTanOVt/GDiKgBa2uCUXcq54pu9WbqqjRMebuKzlGGbGuHB4/swsFKYtqCV26at4fajKwG4c8YarnvFjoxfcVBvztzdfqU1bUErYx5roqnFMGpEGbeMqgjjV1pvAwf4YzCBFN1QA3g1BwAvAr1cl6KKwhfAXnipQJ+ZNqrdb8tL/Qc433UZqiisBY4LeqAh6qEG8FJ3Ar93XYZy7ny81Muui8gFDTWAl/oZcJvrMpQzP8FL3eG6iFzRUK93AUV6zimVV3G81M2ui8glDXU7u9fQWcB9rktRBXMVXup610XkmoY6nb18z+loix0Fl+OlrnFdRD5oqDPZYI+lSC4grvLiErxUwnUR+RLt76mz8WpuBC51XYbKmVbgQrxUqD+wNdTZeDVjscdi6w4qwbYMOBEv9bzrQvJNQ90dXs3+wKOEcJfSiPgfcAxeKhJnl9Vt6u7wUv8F9iLEB4GE2BPAvlEJNGiou89efO9AYKLrUlS3JbAtdIPrQgpJu9894dVcgT3RQugOUQqJJuwZS+53XYgLGuqe8moOAe4EhjuuRH3ZG8CZeKn3XRfiina/e8peemU37Enp9JPRvWbgl8B+UQ40aEudG/a84negrbYrkW+d02lLnQv2vOLaaheets6d0JY612yrPR4Y4bqUkJsKjNMwd6Shzgevphw4G/g/YLDjasJmJnAlXuph14UUKw11Pnk1fYGfYK8O0s9xNUE3H3uGzzvxUq2uiylmGupC8GoGAZcDFwIVjqsJmuXYnUhuwUs1uS4mCDTUhWRPS3wV9phtPUCka/XYgcfr8VIrXBcTJBpqF7yazbFnWTkX2MZxNcXmHexRcffipRpdFxNEGmqXvJoSoBZ7frTDie5up2uwV8W4FS81xXUxQaehLhZezXbAedhLAg1yXE2hfIY9w8zteKnPXRcTFhrqYmO/DjsEOBo4ivB1z98HHvdvrwXtMrFBoKEudl7NSNYHfE+C10VvAV6hPcheao7jekJPQx0kXs0Q4Ehgf2APYCeg1GlNHTVjB7umYcP8lI5eF5aGOsi8mkpgJDbge1L4oKcHeLp/ex8vtbZAy1ed0FCHjQ36CGBLYIh/2zLj52C63gnGAKuAhf5tQSc/FwAfa4CLj4Y6yryaUqAMKMcGuQVo0d0wg01DrVTI6PHUSoWMhlqpkNFQKxUyGuoIEpEjRGSmiMwWkbjrelRu6UBZxIhIKfARcBgwD3vSvpONMf9zWpjKGW2po2dvYLYxZo4xZg3wAHCM45pUDmmoo2coMDft/jx/mgoJDbVSIaOhjp75wFZp94f501RIaKij5w1gexEZLiK9gJOwh0WqkChzXYAqLGNMi4j8CHgGezTXncYYPSF+iOhXWkqFjHa/lQoZDbVSIaOhVipkNNRKhYyGWqmQ0VArFTIaaqVCRkOtVMhoqJUKGQ21UiGjoVYqZDTUSoWMhlqpkNFQKxUyGmqlQub/Ad0FyUBvKQAnAAAAAElFTkSuQmCC\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# OVER & UNDER SAMPLING with SMOTETomek\n"
      ],
      "metadata": {
        "id": "JkQgzcGdWjEE"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "'''\n",
        "from imblearn.combine import SMOTETomek\n",
        "\n",
        "smt = SMOTETomek(sampling_strategy='auto')\n",
        "X_train_smt, y_train_smt = smt.fit_resample(X_train, y_train)\n",
        "\n"
      ],
      "metadata": {
        "id": "YS7vzV_fWsHC"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "'''\n",
        "def plot_2d_space(x_train, y_train, label='Classes'):   \n",
        "    colors = ['#1F77B4', '#FF7F0E']\n",
        "    markers = ['o', 's']\n",
        "    for l, c, m in zip(np.unique(y), colors, markers):\n",
        "        plt.scatter(\n",
        "            X_train[y_train==l, 0],\n",
        "            X_train[y_train==l, 1],\n",
        "            c=c, label=l, marker=m\n",
        "        )\n",
        "    plt.title(label)\n",
        "    plt.legend(loc='upper right')\n",
        "    plt.show()"
      ],
      "metadata": {
        "id": "TvdeIQLkXHhZ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#plot_2d_space(X_train_smt, y_train_smt, 'SMOTE + Tomek links')"
      ],
      "metadata": {
        "id": "fPZIhl38XJMD"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# ▶ *train resampled and test to numpy*"
      ],
      "metadata": {
        "id": "JQTFecAA-dr1"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "X_train = X_train_res.to_numpy()\n",
        "y_train = y_train_res.to_numpy()\n",
        "#X_train = X_train_smt.to_numpy()\n",
        "#y_train =y_train_smt.to_numpy()"
      ],
      "metadata": {
        "id": "IxQh6UiCsz3j"
      },
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": 14,
      "metadata": {
        "id": "ZWPtc1dYgHXb"
      },
      "outputs": [],
      "source": [
        "X_test = X_test.to_numpy()\n",
        "y_test = y_test.to_numpy()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TtVyyhNgN5A5"
      },
      "source": [
        "#**RANDOM** **FOREST**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KyMPnyV5OqVA"
      },
      "source": [
        "## RF traditional method - **Train** **Model**  "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 16,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "57aCQvMtepJt",
        "outputId": "c5bc1568-4dae-4d93-f3f5-a61e51e5e3d1"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Fitting 10 folds for each of 10 candidates, totalling 100 fits\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "RandomizedSearchCV(cv=10, estimator=RandomForestClassifier(), n_jobs=-1,\n",
              "                   param_distributions={'max_depth': [3, 6, 10, None],\n",
              "                                        'min_samples_split': [3, 5, 7],\n",
              "                                        'n_estimators': [10, 50, 100, 500,\n",
              "                                                         700]},\n",
              "                   verbose=4)"
            ]
          },
          "metadata": {},
          "execution_count": 16
        }
      ],
      "source": [
        "#SETTING THE SEARCH SPACE\n",
        "\n",
        "# Number of trees in random forest\n",
        "n_estimators = [10, 50, 100, 500, 700]\n",
        "\n",
        "# Maximum number of levels in tree\n",
        "max_depth = [3, 6, 10, None]\n",
        "\n",
        "# Minimum number of samples required to split a node\n",
        "min_samples_split = [3, 5, 7]\n",
        "\n",
        "# Create the random grid\n",
        "random_grid = {'n_estimators': n_estimators,\n",
        "               'max_depth': max_depth,\n",
        "               'min_samples_split': min_samples_split}   #learning rate\n",
        "\n",
        "# Use the random grid to search for best hyperparameters\n",
        "rf = RandomForestClassifier()\n",
        "rf_random = RandomizedSearchCV(rf, random_grid, n_iter=10, cv = 10, verbose = 4, n_jobs = -1)\n",
        "# Fit the random search model\n",
        "rf_random.fit(X_train, y_train)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Kqe4VC03HnrK"
      },
      "source": [
        "## Performances"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 17,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2yrE0hIjfJXW",
        "outputId": "44f29edb-efe2-47bd-fe97-2f7322d7767f"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "{'n_estimators': 50, 'min_samples_split': 5, 'max_depth': None}\n"
          ]
        }
      ],
      "source": [
        "#best hyperparameters\n",
        "print(rf_random.best_params_)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 18,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JgN1chD4fR5u",
        "outputId": "6a95d512-79e2-4539-b321-cfa76a00cb57"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "RandomForestClassifier(max_depth=10, min_samples_split=5, n_estimators=50)"
            ]
          },
          "metadata": {},
          "execution_count": 18
        }
      ],
      "source": [
        "best_rf = RandomForestClassifier(n_estimators=50, max_depth=10 , min_samples_split=5, min_samples_leaf=1)   # class_weight='balanced_subsample\n",
        "best_rf.fit(X_train, y_train)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#create prediction basing on the test set\n",
        "y_pred = best_rf.predict(X_test)\n",
        "ytrain_pred = best_rf.predict(X_train)"
      ],
      "metadata": {
        "id": "tVyZR8v80epb"
      },
      "execution_count": 23,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": 20,
      "metadata": {
        "id": "_uGtu2Z1dN7D"
      },
      "outputs": [],
      "source": [
        "#create prediction OF THE PROBABILITIES basing on the test set\n",
        "#y_pred = best_rf.predict_proba(X_test)\n",
        "#ytrain_pred = best_rf.predict_proba(X_train)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#ytrain_pred"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FCwES15XyFsw",
        "outputId": "30ccef11-0614-42e3-d52a-a1f3043c5bfe"
      },
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0.02311613, 0.97688387],\n",
              "       [0.02913773, 0.97086227],\n",
              "       [0.44361804, 0.55638196],\n",
              "       ...,\n",
              "       [0.91775633, 0.08224367],\n",
              "       [0.89197839, 0.10802161],\n",
              "       [0.3716207 , 0.6283793 ]])"
            ]
          },
          "metadata": {},
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 24,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KpTgk_axsmM1",
        "outputId": "63a05748-a36b-4faa-d2c0-eff5627ce9b9"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Accuracy test:  0.7829120125416422\n"
          ]
        }
      ],
      "source": [
        "#print(\"Accuracy train: \", accuracy_score(y_train, ytrain_pred))\n",
        "print(\"Accuracy test: \", accuracy_score(y_test, y_pred))\n",
        "#print('F1_score train: ',f1_score(y_train, ytrain_pred))\n",
        "#print('Roc auc train: ',roc_auc_score(y_train, ytrain_pred))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 25,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 621
        },
        "id": "whLITJGnlhYG",
        "outputId": "70f9eabb-7c5c-43ba-cd64-ba6a6fcfc56c"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x720 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "features = X_t.columns\n",
        "importances = best_rf.feature_importances_\n",
        "indices = np.argsort(importances)\n",
        "plt.figure(figsize=(10,10))\n",
        "plt.title('Feature Importances')\n",
        "plt.barh(range(len(indices)), importances[indices], color='b', align='center')\n",
        "plt.yticks(range(len(indices)), [features[i] for i in indices])\n",
        "plt.xlabel('Relative Importance')\n",
        "plt.show()\n",
        "#plt.savefig(output_path+'features_importance.png')"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# get the probability distribution  #probas = rf_clf.predict_proba(X_test)\n",
        "# plot\n",
        "plt.figure(dpi=150)\n",
        "plt.hist(y_pred, bins=20)\n",
        "plt.title('Classification Probabilities')\n",
        "plt.xlabel('Probability')\n",
        "plt.ylabel('# of Instances')\n",
        "plt.xlim([0.5, 1.0])\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 611
        },
        "id": "4OKlrxQSAieT",
        "outputId": "8ff940d3-2885-43c8-dfb6-ea5e0622008f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:matplotlib.legend:No handles with labels found to put in legend.\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 900x600 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## ROC AUC"
      ],
      "metadata": {
        "id": "3RnkR3UpP0dN"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#ROC CURVE\n",
        "from sklearn.metrics import RocCurveDisplay\n",
        "best_rf_disp = RocCurveDisplay.from_estimator(best_rf, X_test, y_test)\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        },
        "id": "z8CjQ1Hx02u1",
        "outputId": "0c00e155-94f4-423a-80a5-06acb556322f"
      },
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEGCAYAAABo25JHAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3deXhU5fXA8e9JyErCvq9hCbJKkFQERKGIO9CqFbVacSl1q/6sYrFatWit1q0VF4pVcQVcaksVtYriAqgECRD2xSCRPUBICCHb+f1xb8ZJCJkbyMxkOZ/nyZO733Nnkjnzvu+97yuqijHGmIYrItwBGGOMCS9LBMYY08BZIjDGmAbOEoExxjRwlgiMMaaBaxTuAKqrVatWmpSUFO4wjDGmTlm6dOkeVW1d2bo6lwiSkpJIS0sLdxjGGFOniMiWo62zqiFjjGngLBEYY0wDZ4nAGGMaOEsExhjTwFkiMMaYBi5oiUBEXhCRXSKScZT1IiJPishGEVkhIicFKxZjjDFHF8wSwUzg7CrWnwMkuz+TgGeDGIsxxpijCNpzBKr6uYgkVbHJeOBldfrB/kpEmolIe1XdHqyYjDEm2EpLlcKSUvIOF3OosISiklKKSpSCohIOHi6mRJXDRaXsyj1MfHQkxaVKSWkpJaVQUlrqzivb9hfQonEUpQqlqpSWKqP7tGVg52Y1HnM4HyjrCGz1m89ylx2RCERkEk6pgS5duoQkOGNM/VdUUsrBw8XsyTtM1r5DFBaX8sP+QzSKEL7fm0+jyAgKi0vZkn2QxNgoCotLyT54mAOHiomKFDbuyqNpXBSFJUphcQkHCoqDGm+bJrH1LhF4pqozgBkAqampNpKOMcbncHEJew8WknOoiAOHilm3MxdUyczOJz46koKiEjbsymPngcMUlZSyJ+8wAuzLL/J0/MbRkcRERbI/v5DurROIjoygoLiEhJgYTuvVmpxDRSS1akx0ZATRjSLILSimS4t4IgQiI4RWCTFERUYQFSkUlyqtE2OIaRRBZIQQGxVJdGQEERFCowghQpzfkZFCpAjRjSKIECFCQESC9hqGMxH8AHT2m+/kLjPGNFCqyoFDxezNL2RP3mH25B7m+735FJcq2/Y739i/23MQgMzsfPbkHQ54zNioCBpFRNAoUmgRH82Ajk1pnRhDi/hoikpK6dwiHoCOzeLo3CKeuOhIEmIa0SQ2itioiKB+ANcW4UwEc4GbRGQ2MATIsfYBY+onVWV/fhE/7D/ENvdne04B2QcLWbvjANl5hWzPKajyGCKgCh2axlJUqvRpn0iT2BY0jY+iQ9NYurZsDEDrxBi6t2pMfEwjGkdHNogP8uMVtEQgIrOAkUArEckC7gWiAFR1OjAPOBfYCOQDVwUrFmNM8BQUlbA79zDrduTy3Z6D7DzgfKCv2naA3XmHKVVl+/4CDhWVlNsvulEErRpHOx/YMY0YkdyKtk1i6dgsjjZNYoiPjqRjs3gSYxvRtWU8cVH2oR4swbxr6NIA6xW4MVjnN8bUjMPFJazadoDt+wvYtv8QP+w/xPacQ2xz57MPFla6X0JMI+KiI2mdEMNlQ7rQoVkcHZvF0qFZHB2axdGycbR9sNcSdaKx2BgTfDn5RazalsPSLfuYl7GD/fmVV9c0jo70fZj379iEDk3jaJ0YQ1x0JP06NKFT83hioyLDcAXmWFkiMKYBKSgqIX3rft5fuZ0Nu/LYuCuPklI96rf6YT1a0rVlY3q1TeAnSS3o3CKeJrGN7Jt8PWOJwJh65HBxCet25PLDvkMcKirx1dl/t+cgy7fmUFhSWm77Dk1j6dE6gf4dm1BSCr3bJdK7fSIntEskppF9q28oLBEYU0flHCpiwbpdvLtiOztyCli9/QAlpUd/zKZjszj6dWjCoC7NOb1Xa3q3SyQiwr7ZG0sExtR6BwqK+GbzXrbszSdzz0E+Wr2THQeOrLtPbpPAgI5NOaFdIgM7NyMhphFtm8TSonE0kfaBb6pgicCYWub77Hxe/XoLX23OZkVWzhHrIwTaNonhp73bcELbRMb0a0fHZnFhiNTUF54SgYhEAAOBDsAhIENVdwUzMGPqs+KSUrbtL2D/oULmrdzBwo17OFhYzObdB8tt1zw+ilOTW3NacitO7NSMzi3iiI+272+mZlX5FyUiPYDfA2cAG4DdQCzQS0TygX8AL6lq6dGPYkzDparsyy9i46481u/MZe7ybWzPOcTWvYeO2LZX2wQuTu1EZIRwZr92nJ7c2urwTUgE+mrxAM44Ab9xHwDzEZE2wGXAFcBLwQnPmNqvsLiUtMy9ZGzLYUt2Ptv2H6KoRFm+dT+5hyvvjfK8E9vTOiGGlM7N6Ng8jgEdm9q99yZsqkwEVT0d7FYN/a3GIzKmFlNVvttzkLU7cpm3cjsfrd7J4eIjC8RN46IY0KkpjWMa0b1VY7q0jKdn6wROaJdI07gouw/f1CrHXNkoImNU9aOaDMaY2qigqIQXF2by9/nrKSg68kO/Z5sEzh3QnmE9WjKoSzOiIxtGj5Wm/jieVqfnARslxtQrhwpLWLvjAG8tzeL7vfls3JVXrpuF5vFR/GpoEsltExjYqZmvC2Nj6rJAjcVzj7YKaFnz4RgTGqrKV5v3Mn/NTjK25bB1r9OZWkVtEmP4SVJz+nVoym1n9iIxNioM0RoTXIFKBCOAy4G8CssFODkoERkTBBt35fLR6l0s3bKXb77be8SQgm0SY9x+deLp0TqBHq0TOKV7S+KirQHX1H+BEsFXQL6qflZxhYisC05Ixhy7vQcL+XbLPtZsP8B32QdZve0Aa3fkltsm0e37PrVrC84d0I4erRPsNk3ToAW6a+icKtadVvPhGOOdqrI8K4enPtnAzgOHWfnDkU/hRkdG0KFpLKf0aMnPB3VkSLeWRDeKCEO0xtRe9oiiqRMOF5eQuSefLzbs5r2V29m+v+CI/nb6tm9CmyYxjBvYgb4dmtCjdQJRkfahb0wglghMrXK42Ok6Of37/azfmcfO3ALSv99faUPu6b1ak9wmgXEpHTixU7MwRGtM/WCJwIRdfmEx0xds4slPNla6vlVCDKldmzO4a3NO6dGSod1b2lO4xtQgSwQm5EpLlUWbsnnq0w0s35pzxKDmd5x9Ar3bJZLSuTnN4+0pXGOCzXMiEJH7VPW+o80bU5XiklI+XLWTJ+dvYN3O8nfx9G3fhPMHtufq4d3sm74xYVCdEsHSAPPG+GzPOcTn63eTlrmPLzbsOaJh94pTunLlsK70bJMYpgiNMWU8JwJV/W9V88YUFpfyf3OWMW/ljiPWDe7anJTOzbj61G42iIoxtUygLiamAUcdBFVVb67xiEydkltQxFOfbGTWN9+Xe1r3jD5tGDuwA6clt6Z54+gwRmiMCSRQiSAtJFGYOqOkVJnx+WY+Wr2DDbvyyPX78O/eujHjB3Zk4rAkmsZbnzzG1BWBniwuN+CMiMSran5wQzK1zYGCIqYv2MSbS7PYnXvYtzwxphFDurVgRHIrrh3R3Rp6jamjvI5ZPBSn2+kEoIuIDMQZteyGYAZnwmPjrlzmLNnK+xk7yNpX/kGu5vFRXDW8G5NOsw9+Y+oLr43FfwPOAuYCqOpyEbG+huqRrzdn8+Qnzn39eX7DKya1jOfU5Fb0ad+EC0/qZB/+xtRD1blraGuFB3tKjratqRuKS0r555ff8dD7a8stH9y1Odee2o1zBrQPU2TGmFDymgi2isgwQEUkCrgFWBO8sEywlJQqCzfu4eEP1rJq2wHf8l5tE7h/fH+GdLfxhoxpaLwmguuAvwMdgW3Ah8CNwQrK1Lysffmc+vCn5ZZFR0Zw1alJ/PanySTEWG8jxjRUnv77VXUP8MvqHlxEzsZJIJHAP1X1oQrruwAvAc3cbaao6rzqnsdU7m8fr2dlVg7z1+4qt/zGUT0YN7AjJ7Szp3qNMd7vGuqO84F+Cs4DZouBW1V1cxX7RAJPA2OALGCJiMxV1dV+m90NvKGqz4pIX2AekHQsF2J+tDIrhwufXURhSSkAHZvF0al5HBee1ImLf9I5zNEZY2obr/UBr+N8qP/cnb8EmAUMqWKfk4GNZclCRGYD4wH/RKBAE3e6KU61kzlGWfvyuen1ZaRv3e9b9tWdo2nXNDaMURljajuviSBeVV/xm39VRCYH2KcjsNVvPosjE8d9wP9E5LdAY+CMyg4kIpOASQBdunTxGHLDsCfvME98tJ45S7ZSXOr0BpIY24hbz+jF1ad2C3N0xpi6IFBfQy3cyfdFZAowG+db/AScapzjdSkwU1Ufcx9ae0VE+qtqqf9GqjoDmAGQmpp61L6PGpJl3+/j1y8vZU/ej0/69mnfhOtO7864gR2sD39jjGeBSgRLcT74yz5VfuO3ToE7q9j3B8C/QrqTu8zfNcDZAKq6WERigVbALkylvs/O57RHyt/9M+Wc3vzmtO724W+MOSaB+ho6nrqFJUCyiHTDSQCXAJdV2OZ7YDQwU0T6ALHA7uM4Z722eXceP33sM9/8rF+fwtAedt+/Meb4VGeEsv5AX5wPawBU9eWjba+qxSJyE84zB5HAC6q6SkSmAmmqOhe4DXhORG7FKWFMVFWr+qkgO+8wgx/42Dc/pFsLZk86xUoAxpga4fX20XuBkTiJYB5wDvAlcNREAOA+EzCvwrJ7/KZXA8OrFXEDcaCgiCc+Ws8na3exJfvHDl+nX34SZ/e3rh+MMTXHa4ngImAgsExVrxKRtsCrwQurYXvg3dX888vvfPN92zdhTN+23DqmVxijMsbUV14TwSFVLRWRYhFpgtOYa08m1bDcgiIG3Pc/3/wVp3TlT+P6ERFhVUDGmODxmgjSRKQZ8BzOnUR5OE8XmxpQUFTCvf9ZxZy0Hx+7WH7vmTSNs1G+jDHB57WvobIBaKaLyAdAE1VdEbywGo53V2zjpteX+eZPaJvIh7faUA/GmNAJ9EDZSVWtU9Vvaz6khmHxpmwmvZxGrjsIzE+SmjNn0lCrBjLGhFygEsFjVaxT4Kc1GEuDcPBwMYMf+IiCoh8fnn7jN0M5uVuLKvYyxpjgCfRA2ahQBdIQXDJjMV9t3uubf/3aIQzr2SqMERljTDUeKDPHZ+hf5rM9pwCAXw7pwtTx/Ym0aiBjTC1giSDIikpK6XX3+5Q9L5129xm0SogJb1DGGOPHEkGQHCgoYvqCTTyzYJNv2YLbR1oSMMbUOl67mBCcoSq7q+pUd4jJdqr6TVCjq6PSt+7nZ08v9M03iW3E8nvPtL6BjDG1ktcSwTNAKc5dQlOBXOBt4CdBiqvOeviDtTzrlgJO7NSUt64bRnSjiDBHZYwxR+c1EQxR1ZNEZBmAqu4TkeggxlXn7DtYyKD7P/LNT7t0EGMHdghjRMYY443XRFDkDkavACLSGqeEYABVLZcEPp88ii4t48MYkTHGeOc1ETwJvAO0EZE/4/RGenfQoqpDKpYEMh86L4zRGGNM9Xnta+g1EVmKM5qYAD9T1TVBjayO8E8C6x44O4yRGGPMsfF619CTwGxVfTrI8dQp17+61DdtJQFjTF3l9XaWpcDdIrJJRB4VkdRgBlUXbNiZy/sZOwB4+/phYY7GGGOOnadEoKovqeq5OLeLrgMeFpENQY2sFvsgYwdjnvgcgPvH92Nw1+ZhjsgYY45ddW9w7wn0BroCa2s+nLrhOrdK6I/n9+WKoUnhDcYYY46Tp0QgIn91SwBTgQwgVVXHBjWyWmr8U1/6pq85tVsYIzHGmJrh9fbRTcBQVd0TzGBqu+tfXcryrBwAvrlrdJijMcaYmhFohLLeqroWWAJ0cfsY8mlII5R9vHqnr3H4reuG0iYxNswRGWNMzQhUIvgdMInKRyprMCOUPfrhOp76dCMAD/58AKlJNpqYMab+CDRC2SR38hxVLfBfJyIN4ivxN9/t9SWBG0b24LIhXQLsYYwxdYvXu4YWeVxWr+w9WMjF/1gMwHWn9+COs3uHOSJjjKl5gdoI2gEdgTgRGYTTvQRAE6Be96pWXFLKSW73EcN7tmTKOZYEjDH1U6A2grOAiUAn4HG/5bnAH4IUU63Q8673fdMvXXVyGCMxxpjgCtRG8BLwkohcqKpvhyimsFuwbpdv+ru/nGsjixlj6rVAVUOXq+qrQJKI/K7ielV9vJLd6rScQ0VMfHEJAM/9KtWSgDGm3gvUWNzY/Z0AJFbyUyUROVtE1onIRhGZcpRtLhaR1SKySkRer0bsQfHJ2p0AnNO/HWP6tg1zNMYYE3yBqob+4f7+U3UP7I5o9jQwBsgClojIXFVd7bdNMnAnMNwd/rJNdc9TkwqKSrh1znIAfn1a93CGYowxIVOdvoaaiEiUiMwXkd0icnmA3U4GNqrqZlUtBGYD4yts82vgaVXdB6Cquwij3n/8AIARya04qYv1KGqMaRi8PkdwpqoeAM4HMnF6IZ0cYJ+OwFa/+Sx3mb9eQC8RWSgiX4lIpUN8icgkEUkTkbTdu3d7DLl6Lpmx2Dc9/fLBQTmHMcbURl4TQVkV0nnAm6qaU0PnbwQkAyOBS4HnRKRZxY1UdYaqpqpqauvWrWvo1D8qKVW+2rwXgC/uGEXjGK998RljTN3nNRG8KyJrgcHAfBFpDRQE2OcHoLPffCd3mb8sYK6qFqnqd8B6nMQQUve/6zRbXHhSJzq3qNfPyRljzBG8jlA2BRiGMw5BEXCQI+v7K1oCJItINxGJBi4B5lbY5t84pQFEpBVOVdFmz9HXgLTMvcxclAnAL1I7hfLUxhhTK3gdvD4KuBw4zb2v/jNgelX7qGqxiNwEfAhEAi+o6ioRmQqkqepcd92ZIrIaKAEmq2r2MV9NNakqF0132gZe//UQTuneMlSnNsaYWsNrZfizQBTwjDt/hbvs2qp2UtV5wLwKy+7xm1acrq6PeFgtFF5evAWAZvFRDOvRKhwhGGNM2HlNBD9R1YF+85+IyPJgBBQqJaXKvXNXAfDezSPCHI0xxoSP18biEhHpUTYjIt1xqnLqrOyDhwFI6dyMjs3iwhyNMcaEj9cSwWTgUxHZjNMVdVfgqqBFFQJ/mbcWgPMGtA9zJMYYE14BE4F7q2gOzpPCZV1ArFPVw8EMLJi27s3nnWXOnawXp3YOsLUxxtRvVVYNici1wCpgGpAOJKnqirqcBADW7sgF4PqRPWgaHxXmaIwxJrwClQj+D+inqrvddoHXOPJZgDrnjTSn5wvrXdQYYwI3Fheq6m4AVd0MxAQ/pOD7PjsfgEGdj+jNwhhjGpxAJYJOIvLk0eZV9ebghBVc63bmEtMowgadMcYYAieCij2MLg1WIKGyaXceAJ2a2y2jxhgD3sYsrlcumfEVAFcOSwpvIMYYU0sEumvoORHpf5R1jUXkahH5ZXBCC47e7ZwRNq84pWuYIzHGmNohUNXQ08A9IjIAyAB2A7E4XUU3AV7AuZOozli/M5eBnZtZ+4AxxrgCVQ2lAxeLSAKQCrQHDgFrVHVdCOKrcVGREezMCTSUgjHGNByeuphQ1TxgQXBDCT5VJWvfIcYO7BDuUIwxptbw2ulcvVA2HOX3e/PDHIkxxtQeDSoRvLnUeaJ48pknhDkSY4ypPaqVCESkTg/ou+uA00XSqck2CI0xxpTxlAhEZJg7nORad36giDwTYLdap3nj6HCHYIwxtY7XEsETwFlANoCqLgdOC1ZQwZKdd5iklnW6UGOMMTXOc9WQqm6tsKjOjVC2bf8h9uUXhTsMY4ypVbwmgq0iMgxQEYkSkduBNUGMKygys/Pp1qpxuMMwxphaxWsiuA64EegI/ACkADcEK6hgSYxtRFSkPVFsjDH+vI5ZfIKqlutTSESGAwtrPqTgaRQh9GnfJNxhGGNMreK1RDDN4zJjjDF1TJUlAhEZCgwDWovI7/xWNQEigxmYMcaY0AhUNRQNJLjbJfotPwBcFKygjDHGhE6g3kc/Az4TkZmquiVEMRljjAkhr43F+SLyCNAPZzwCAFT1p0GJyhhjTMh4bSx+Dad7iW7An4BMYEmQYjLGGBNCXhNBS1V9HihS1c9U9WrASgPGGFMPeK0aKuuXYbuInAdsA1oEJyRjjDGh5LVE8ICINAVuA24H/gn8X6CdRORsEVknIhtFZEoV210oIioiqR7jMcYYU0O8DlX5rjuZA4wC35PFRyUikcDTwBggC1giInNVdXWF7RKBW4Cvqxe6McaYmlBliUBEIkXkUhG5XUT6u8vOF5FFwFMBjn0ysFFVN6tqITAbGF/JdvcDDwM2orwxxoRBoKqh54FrgZbAkyLyKvAo8FdVHRRg346Af9fVWe4yHxE5Ceisqu9VdSARmSQiaSKStnv37gCnNcYYUx2BqoZSgRNVtVREYoEdQA9VzT7eE4tIBPA4MDHQtqo6A5gBkJqaqsdyvpJSZV9+EXpMextjTP0VqERQqKqlAKpaAGyuRhL4AejsN9/JXVYmEegPLBCRTOAUYG6wGozzC4sBKCopDcbhjTGmzgpUIugtIivcaQF6uPMCqKqeWMW+S4BkEemGkwAuAS4rW6mqOYBvFHkRWQDcrqpp1b6KaujZJiGYhzfGmDonUCLoc6wHVtViEbkJ+BCnp9IXVHWViEwF0lR17rEe2xhjTM0J1OnccXU0p6rzgHkVlt1zlG1HHs+5jDHGHBvPg9cbY4ypnxpMIth30Oklo6jEbhsyxhh/nhOBiMSJyAnBDCaYcg87iaBJnNfulYwxpmHwlAhEZCyQDnzgzqeISJ1s7G2VEBPuEIwxplbxWiK4D6fLiP0AqpqOMzaBMcaYOs5rIihy7/v3Z5XtxhhTD3itMF8lIpcBkSKSDNwMLApeWMYYY0LFa4ngtzjjFR8GXsfpjjrgeATGGGNqP68lgt6qehdwVzCDCaaCIutjyBhjKuO1RPCYiKwRkfvLxiWoa7L25Yc7BGOMqZU8JQJVHYUzMtlu4B8islJE7g5qZDUsOtK51K4t48MciTHG1C6eHyhT1R2q+iRwHc4zBZX2GWSMMaZu8fpAWR8RuU9EVgLTcO4Y6hTUyIwxxoSE18biF4A5wFmqui2I8RhjjAkxT4lAVYcGOxBjjDHhUWUiEJE3VPVit0rI/0liLyOUGWOMqQMClQhucX+fH+xAjDHGhEeVjcWqut2dvEFVt/j/ADcEPzxjjDHB5vX20TGVLDunJgMxxhgTHoHaCK7H+ebfXURW+K1KBBYGMzBjjDGhEaiN4HXgfeAvwBS/5bmqujdoURljjAmZQIlAVTVTRG6suEJEWlgyMMaYus9LieB8YCnO7aPit06B7kGKyxhjTIhUmQhU9Xz3tw1LaYwx9ZTXvoaGi0hjd/pyEXlcRLoENzRjjDGh4PX20WeBfBEZCNwGbAJeCVpUxhhjQsZrIihWVQXGA0+p6tM4t5AaY4yp47z2PporIncCVwAjRCQCiApeWMYYY0LFa4lgAs7A9Ver6g6csQgeCVpUQbBuZ264QzDGmFrJ61CVO4DXgKYicj5QoKovBzWyGhYXFQlAx2ZxYY7EGGNqF693DV0MfAP8ArgY+FpELvKw39kisk5ENorIlErW/05EVovIChGZLyJdq3sB1RUZIYE3MsaYBsRrG8FdwE9UdReAiLQGPgbeOtoOIhIJPI3TYV0WsERE5qrqar/NlgGpqprv9mv0V5xqKGOMMSHitY0goiwJuLI97HsysFFVN6tqITAb564jH1X9VFXz3dmvsHGQjTEm5LyWCD4QkQ+BWe78BGBegH06Alv95rOAIVVsfw1OB3dHEJFJwCSALl3sOTZjjKlJXscsniwiFwCnuotmqOo7NRWEiFwOpAKnH+X8M4AZAKmpqVrZNsYYY45NoPEIkoFHgR7ASuB2Vf3B47F/ADr7zXdyl1U8xxk4bRCnq+phj8c2xhhTQwLV878AvAtciNMD6bRqHHsJkCwi3UQkGrgEmOu/gYgMAv4BjKvQBmGMMSZEAlUNJarqc+70OhH51uuBVbVYRG4CPgQigRdUdZWITAXSVHUuzkNpCcCbIgLwvaqOq/ZVGGOMOWaBEkGs+6297Ob7OP95Va0yMajqPCo0KqvqPX7TZ1Q7YmOMMTUqUCLYDjzuN7/Db16BnwYjKGOMMaETaGCaUaEKxBhjTHh4faDMGGNMPWWJwBhjGjhLBMYY08B57X1U3LGK73Hnu4jIycENzRhjTCh4LRE8AwwFLnXnc3F6FjXGGFPHee10boiqniQiywBUdZ/7tLAxxpg6zmuJoMgdX0DBNx5BadCiMsYYEzJeE8GTwDtAGxH5M/Al8GDQojLGGBMyXruhfk1ElgKjcbqX+JmqrglqZMYYY0LCUyIQkS5APvBf/2Wq+n2wAjPGGBMaXhuL38NpHxAgFugGrAP6BSkuY4wxIeK1amiA/7yInATcEJSIjDHGhNQxPVnsdj9d1fjDxhhj6givbQS/85uNAE4CtgUlImOMMSHltY0g0W+6GKfN4O2aD8cYY0yoBUwE7oNkiap6ewjiMcYYE2JVthGISCNVLQGGhygeY4wxIRaoRPANTntAuojMBd4EDpatVNV/BTE2Y4wxIeC1jSAWyMYZo7jseQIFLBEYY0wdFygRtHHvGMrgxwRQRoMWlak3ioqKyMrKoqCgINyhGNMgxMbG0qlTJ6KiojzvEygRRAIJlE8AZSwRmICysrJITEwkKSkJkcr+jIwxNUVVyc7OJisri27dunneL1Ai2K6qU48vNNOQFRQUWBIwJkREhJYtW7J79+5q7RfoyWL77zXHzZKAMaFzLP9vgRLB6GMLxRhjTF1RZSJQ1b2hCsSYYImMjCQlJYX+/fszduxY9u/fXyPHnTlzJjfddFONHCspKYkBAwaQkpJCSkoKixYtqpHjVpSens68efPKLXv//fdJTU2lb9++DBo0iNtuuw2A++67j0cffbTGzj1s2DDf9OTJk+nXrx+TJ09m+vTpvPzyy8d17GXLlnHNNdeUW/azn/2MU045pdyyiRMn8tZbb5VblpCQ4Jtev3495557LsnJyZx00klcfPHF7Ny587hi27t3L2PGjCE5OZkxY8awb9++Sre744476NevH3369OHmm29G1WmGnTNnDieeeCL9+vXj97//vW/7p556ihdeeOG4YumcAEQAABVrSURBVCtzTJ3OGVOXxMXFkZ6eTkZGBi1atODpp58Od0iV+vTTT0lPTyc9Pb3ch2ZViouLq3WOiokgIyODm266iVdffZXVq1eTlpZGz549q3VMr/yT24wZM1ixYgWPPPII1113Hb/61a88H6eya37wwQe5+eabffP79+9n6dKl5OTksHnzZk/HLSgo4LzzzuP6669nw4YNfPvtt9xwww3Vrm+v6KGHHmL06NFs2LCB0aNH89BDDx2xzaJFi1i4cCErVqwgIyODJUuW8Nlnn5Gdnc3kyZOZP38+q1atYseOHcyfPx+Aq6++mmnTph1XbGW8PkdgzHH7039XsXrbgRo9Zt8OTbh3rPdhMYYOHcqKFSsA+Oabb7jlllsoKCggLi6OF198kRNOOIGZM2cyd+5c8vPz2bRpEz//+c/561//CsCLL77IX/7yF5o1a8bAgQOJiYkBIDMzk6uvvpo9e/bQunVrXnzxRbp06cLEiROJi4tj2bJl7Nq1ixdeeIGXX36ZxYsXM2TIEGbOnHnUWKs6ZmxsLMuWLWP48OHceOON3HjjjezevZv4+Hiee+45evfuzZtvvsmf/vQnIiMjadq0KR9//DH33HMPhw4d4ssvv+TOO+/kvffe46677qJ3796AU3q6/vrrj4jlueeeY8aMGRQWFtKzZ09eeeUV4uPjjzjH559/zqpVq7jqqqsoLCyktLSUt99+m+TkZBISEsjLy2PcuHHk5eUxePBg7rzzTtasWUNCQgK33347mzZtqvRaKl7z448/7ostNzeXFStWMHDgQN+yf/3rX4wdO5a2bdsye/Zs/vCHPwT823j99dcZOnQoY8eO9S0bOXJkwP0C+c9//sOCBQsAuPLKKxk5ciQPP/xwuW1EhIKCAgoLC1FVioqKaNu2LZs3byY5OZnWrVsDcMYZZ/D2228zevRo4uPjSUpK4ptvvuHkk08+rhitRGAajJKSEubPn8+4ceMA6N27N1988QXLli1j6tSp5T4s0tPTmTNnDitXrmTOnDls3bqV7du3c++997Jw4UK+/PJLVq9e7dv+t7/9LVdeeSUrVqzgl7/8Zblvp/v27WPx4sU88cQTjBs3jltvvZVVq1axcuVK0tPTfduNGjWKlJQUhgwZEvCYWVlZLFq0iMcff5xJkyYxbdo0li5dyqOPPsoNNzhDhUydOpUPP/yQ5cuXM3fuXKKjo5k6dSoTJkwgPT2dCRMmkJGRweDBgwO+dhdccAFLlixh+fLl9OnTh+eff77ScwBMnz6dW265hfT0dNLS0ujUqVO5Y82dO9dXSpswYUK5dUe7lorX7C8tLY3+/fuXWzZr1iwuvfRSLr30UmbNmhXw+gDPr0Vubq6vCq/ij//fRJmdO3fSvn17ANq1a1dpVdPQoUMZNWoU7du3p3379px11ln06dOHnj17sm7dOjIzMykuLubf//43W7du9e2XmprKF1984en6qmIlAhMy1fnmXpMOHTpESkoKP/zwA3369GHMmDEA5OTkcOWVV7JhwwZEhKKiIt8+o0ePpmnTpgD07duXLVu2sGfPHkaOHOn7djZhwgTWr18PwOLFi/nXv5wH7a+44gruuOMO37HGjh2LiDBgwADatm3LgAHOOE/9+vUjMzOTlJQUwKkaatWqlW+/qo75i1/8gsjISPLy8li0aBG/+MUvfOsOHz4MwPDhw5k4cSIXX3wxF1xwwXG9hhkZGdx9993s37+fvLw8zjrrrKOeY+jQofz5z38mKyuLCy64gOTkZE/nqOpa/K+5ou3bt/veE3A+eDds2MCpp56KiBAVFUVGRgb9+/ev9I6a6t5lk5iYWC6BV4eIVHq+jRs3smbNGrKysgAYM2YMX3zxBSNGjODZZ59lwoQJREREMGzYMDZt2uTbr02bNqxdu/aYYvEX1BKBiJwtIutEZKOITKlkfYyIzHHXfy0iScGMxzRMZd8+t2zZgqr62gj++Mc/MmrUKDIyMvjvf/9b7unnsiofcKpLqlsX76/sWBEREeWOGxERcczHbdy4MQClpaU0a9bM17aQnp7OmjVrAOeb+QMPPMDWrVsZPHgw2dnZRxynX79+LF26NOD5Jk6cyFNPPcXKlSu59957fa9VZee47LLLfN/6zz33XD755BNP11TVtfhfc0VxcXHl3rs33niDffv20a1bN5KSksjMzPSVClq2bFmusXbv3r2+5Ov1tahuiaBt27Zs374dcJJWmzZtjtjmnXfe4ZRTTiEhIYGEhATOOeccFi9eDDhfJL7++msWL17MCSecQK9evXz7lVVrHq+gJQK3++qngXOAvsClItK3wmbXAPtUtSfwBPAwxgRJfHw8Tz75JI899hjFxcXk5OTQsWNHgCrr6ssMGTLE14BXVFTEm2++6Vs3bNgwZs+eDcBrr73GiBEjjjteL8ds0qQJ3bp188WiqixfvhyATZs2MWTIEKZOnUrr1q3ZunUriYmJ5Obm+vafPHkyDz74oK9kU1payvTp0484T25uLu3bt6eoqIjXXnvNt7yyc2zevJnu3btz8803M378eF+bTCBVXUtV+vTpw8aNG33zs2bN4oMPPiAzM5PMzEyWLl3qex1HjhzJnDlzKCwsBJz3fdSoUQBcdtllLFq0iPfee893rM8//5yMjIxy5ysrEVT207dvxY84GDduHC+99BIAL730EuPHjz9imy5duvDZZ59RXFxMUVERn332GX369AFg165dgFPF+Mwzz3Dttdf69lu/fv0R1WLHIpglgpOBjaq6WVULgdlAxVdgPPCSO/0WMFrs6SMTRIMGDeLEE09k1qxZ3HHHHdx5550MGjTI0zfz9u3bc9999zF06FCGDx/u+0cFmDZtGi+++CInnngir7zyCn//+9+PO1avx3zttdd4/vnnGThwIP369eM///kP4HzIDxgwgP79+zNs2DAGDhzIqFGjWL16NSkpKb7bEv/2t79x6aWX0qdPH/r371/pXTb3338/Q4YMYfjw4b6G5aOd44033qB///6kpKSQkZFRrTuCjnYtVenduzc5OTnk5uaSmZnJli1byt022q1bN5o2bcrXX3/N+eefz4gRIxg8eDApKSksXLjQ13AbFxfHu+++y7Rp00hOTqZv374888wz5aqdjsWUKVP46KOPSE5O5uOPP2bKFKdyJC0tzfehftFFF9GjRw8GDBjAwIEDGThwoK/R+pZbbqFv374MHz6cKVOmlCsRLFy40FfVeTyk7F7VmiYiFwFnq+q17vwVwBBVvclvmwx3myx3fpO7zZ4Kx5oETALo0qXL4C1btlQ7nv+t2sG/03/g8YtTiI06sp7RBMeaNWvKfWAaEwxPPPEEiYmJ5b4t13fLli3j8ccf55VXXjliXWX/dyKyVFVTKztWnbhrSFVnqGqqqqYea3Y+s187nvnlYEsCxtRD119/fbn2l4Zgz5493H///TVyrGDeNfQD0NlvvpO7rLJtskSkEdAUZ9wDY4zxLDY2liuuuCLcYYRUTVQJlQlmiWAJkCwi3UQkGrgEmFthm7nAle70RcAnGqy6KhM29pYaEzrH8v8WtESgqsXATcCHwBrgDVVdJSJTRWScu9nzQEsR2Qj8DjjiFlNTt8XGxpKdnW3JwJgQKBuPIDY2tlr7Ba2xOFhSU1M1LS0t3GEYj2yEMmNC62gjlFXVWGxPFpugioqKqtZIScaY0KsTdw0ZY4wJHksExhjTwFkiMMaYBq7ONRaLyG6g+o8WO1oBewJuVb/YNTcMds0Nw/Fcc1dVrfSJ3DqXCI6HiKQdrdW8vrJrbhjsmhuGYF2zVQ0ZY0wDZ4nAGGMauIaWCGaEO4AwsGtuGOyaG4agXHODaiMwxhhzpIZWIjDGGFOBJQJjjGng6mUiEJGzRWSdiGwUkSN6NBWRGBGZ467/WkSSQh9lzfJwzb8TkdUiskJE5otI13DEWZMCXbPfdheKiIpInb/V0Ms1i8jF7nu9SkReD3WMNc3D33YXEflURJa5f9/nhiPOmiIiL4jILncEx8rWi4g86b4eK0TkpOM+qarWqx8gEtgEdAeigeVA3wrb3ABMd6cvAeaEO+4QXPMoIN6dvr4hXLO7XSLwOfAVkBruuEPwPicDy4Dm7nybcMcdgmueAVzvTvcFMsMd93Fe82nASUDGUdafC7wPCHAK8PXxnrM+lghOBjaq6mZVLQRmA+MrbDMeeMmdfgsYLSISwhhrWsBrVtVPVTXfnf0KZ8S4uszL+wxwP/AwUB/6wfZyzb8GnlbVfQCquivEMdY0L9esQBN3uimwLYTx1ThV/RzYW8Um44GX1fEV0ExE2h/POetjIugIbPWbz3KXVbqNOgPo5AAtQxJdcHi5Zn/X4HyjqMsCXrNbZO6squ+FMrAg8vI+9wJ6ichCEflKRM4OWXTB4eWa7wMuF5EsYB7w29CEFjbV/X8PyMYjaGBE5HIgFTg93LEEk4hEAI8DE8McSqg1wqkeGolT6vtcRAao6v6wRhVclwIzVfUxERkKvCIi/VW1NNyB1RX1sUTwA9DZb76Tu6zSbUSkEU5xMjsk0QWHl2tGRM4A7gLGqerhEMUWLIGuORHoDywQkUycutS5dbzB2Mv7nAXMVdUiVf0OWI+TGOoqL9d8DfAGgKouBmJxOmerrzz9v1dHfUwES4BkEekmItE4jcFzK2wzF7jSnb4I+ETdVpg6KuA1i8gg4B84SaCu1xtDgGtW1RxVbaWqSaqahNMuMk5V6/I4p17+tv+NUxpARFrhVBVtDmWQNczLNX8PjAYQkT44iWB3SKMMrbnAr9y7h04BclR1+/EcsN5VDalqsYjcBHyIc8fBC6q6SkSmAmmqOhd4Hqf4uBGnUeaS8EV8/Dxe8yNAAvCm2y7+vaqOC1vQx8njNdcrHq/5Q+BMEVkNlACTVbXOlnY9XvNtwHMicitOw/HEuvzFTkRm4STzVm67x71AFICqTsdpBzkX2AjkA1cd9znr8OtljDGmBtTHqiFjjDHVYInAGGMaOEsExhjTwFkiMMaYBs4SgTHGNHCWCBoAESkRkXS/n6Qqts2rgfPNFJHv3HN96z7tWd1j/FNE+rrTf6iwbtHxxugep+x1yRCR/4pIswDbpxxLz5Yi0l5E3nWnR4pIjnveNSJy7zEcb1xZL5wi8rOy18mdn+o+OHhc3PfwogDbLKjOA3rutb/rYbtKe98UkUdF5Kdez2e8s0TQMBxS1RS/n8wQnHOyqqYAU3AeZKsWVb1WVVe7s3+osG5YDcQHP74u/XGeJ7kxwPYpOPdvV9fvgOf85r9wX5tUnD5yqtWNsKrOVdWH3Nmf4fS4WbbuHlX9+BhirE1mApX1kTQN5+/J1DBLBA2QiCSIMybBtyKyUkSO6LXT/Rb7ud835hHu8jNFZLG775sikhDgdJ8DPd19f+ceK0NE/s9d1lhE3hOR5e7yCe7yBSKSKiIPAXFuHK+56/Lc37NF5Dy/mGeKyEUiEikij4jIEnH6a/+Nh5dlMW7HXSJysnuNy0RkkYic4D7VOhWY4MYywY39BRH5xt22st5PAS4EPqi4UFUPAkuBnm5p4ys33ndEpLkby83y4zgSs91lE0XkKREZBowDHnFj6uH3GpwtIm/6vTa+b+PVfQ9F5B73tcwQkRki5XrqvcLvb+Rkd3uvr0uljtb7pqpuAVqKSLvqHM94EI7+tu0ntD84T5imuz/v4DxR3sRd1wrnCcWyhwvz3N+3AXe505E4ffe0wvlgb+wu/z1wTyXnmwlc5E7/AvgaGAysBBrjPOG8ChiE8yH5nN++Td3fC3DHDyiLyW+bshh/DrzkTkfj9MgYB0wC7naXxwBpQLdK4szzu743gbPd+SZAI3f6DOBtd3oi8JTf/g8Cl7vTzXD69Wlc4RzdgKV+8yOBd93plkAm0A9YAZzuLp8K/M2d3gbElJ2jYhz+r7X/vPsef+/3Xj0LXH6M72ELv+WvAGP93qPn3OnTcPvPP9rrUuHaU4F/VvE3m0Ql/fHjlKwuDPf/VH37qXddTJhKHVKnKgIAEYkCHhSR04BSnG/CbYEdfvssAV5wt/23qqaLyOk41RAL3S+F0TjfpCvziIjcjdPnyzU4fcG8o863YETkX8AInG/Kj4nIwzgfEl9U47reB/4uIjE4VQmfq+ohETkTONGvjrspTsdr31XYP05E0t3rXwN85Lf9SyKSjNNlQdRRzn8mME5EbnfnY4Eu7rHKtOfIfm9GiMgynNf+IZyO4pqp6mfu+pdwEhM4CeI1Efk3Tj9CnqjTNcMHwFgReQs4D7gDp9dZr+9hmVEicgcQD7TASeL/ddfNcs/3uYg0Eaed5Wivi398acC1Xq/Hzy6gwzHsZ6pgiaBh+iXQGhisqkXi9M4Z67+B+499Gs4HyEwReRzYB3ykqpd6OMdkVX2rbEZERle2kaqud+vIzwUeEJH5qjrVy0WoaoGILADOAibgDFoCzshNv1XVDwMc4pCqpohIPE5fNjcCT+IMZvOpqv5cnIb1BUfZX3C+na6r6hxUeG1x2gjO9x1EpGkV+5+H8217LHCXiAyoYtuKZgM34VSzpKlqrlut4/U9RERigWdwSmdbReQ+yl9PxT5qlKO8LiLSthqxH00szmtqapC1ETRMTYFdbhIYBRwxfrE4YxrvVNXngH/iDJ33FTBcRMrq/BuLSC+P5/wC+JmIxItIY5xqnS9EpAOQr6qv4nSMV1nDaZFbMqnMHJxOt8pKF+B8qF9fto+I9HLPWSl1Rm67GbhNfuyWvKxb34l+m+biVJGV+RD4bVmduTg9vFa0Hqea46hUNQfYJ247DHAF8Jk4Yyp0VtVPcapwmuJUq/mrGJO/z3Bez1/zY5Ks7ntY9qG/x21LqHgnUVmbzqk4vWDm4O11OVa9gErH8jXHzhJBw/QakCoiK4FfAWsr2WYksNytwpgA/F1Vd+N8MM4SkRU4VQq9vZxQVb/FqXf+BqfN4J+qugwYAHzjVtHcCzxQye4zgBXiNhZX8D+c6o6P1RnKEJzEtRr4VpxbEP9BgNKvG8sKnEFO/gr8xb12//0+BfqWNRbjlByi3NhWufMVj3sQ2FT2wVuFK3Gq01bg3J00Faft4lX3fVoGPKlHDjAzG5jsNsr2qHDuEuBd4Bz3N9V9D93zPYfz4fshTpWhvwL3dZqOUwUIHl4XcW4E+Gdl5xSn983FwAkikiUi17jLo3BuPKjLXYnXStb7qDFBJiI/x6mGuzvcsdRl7ut4kqr+Mdyx1DfWRmBMkKnqOyJSl8fEri0aAY+FO4j6yEoExhjTwFkbgTHGNHCWCIwxpoGzRGCMMQ2cJQJjjGngLBEYY0wD9/9mQW06hr6iCgAAAABJRU5ErkJggg==\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "'''\n",
        "#PRECISION AND RECALL CURVE\n",
        "from sklearn.metrics import precision_recall_curve, average_precision_score, auc\n",
        "# get precision and recall values\n",
        "precision, recall, thresholds = precision_recall_curve(y_test, probas[:,0], pos_label=0)\n",
        "# average precision score\n",
        "avg_precision = average_precision_score(y_test, probas[:,1])\n",
        "# precision auc\n",
        "pr_auc = auc(recall, precision)\n",
        "# plot\n",
        "plt.figure(dpi=150)\n",
        "plt.plot(recall, precision, lw=1, color='blue', label=f'AP={avg_precision:.3f}; AUC={pr_auc:.3f}')\n",
        "plt.fill_between(recall, precision, -1, facecolor='lightblue', alpha=0.5)\n",
        "plt.title('PR Curve for RF classifier')\n",
        "plt.xlabel('Recall (TPR)')\n",
        "plt.ylabel('Precision')\n",
        "plt.xlim([-0.05, 1.05])\n",
        "plt.ylim([-0.05, 1.05])\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 235
        },
        "id": "N2Ku2DQNlL0o",
        "outputId": "1baf8080-01fd-4b76-9589-cc524420406d"
      },
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "error",
          "ename": "NameError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-27-694d8410a552>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      2\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0msklearn\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmetrics\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mprecision_recall_curve\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maverage_precision_score\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mauc\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0;31m# get precision and recall values\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 4\u001b[0;31m \u001b[0mprecision\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrecall\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mthresholds\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mprecision_recall_curve\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mprobas\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpos_label\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      5\u001b[0m \u001b[0;31m# average precision score\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      6\u001b[0m \u001b[0mavg_precision\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0maverage_precision_score\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mprobas\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mNameError\u001b[0m: name 'probas' is not defined"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 28,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "F_AcufgWQ9oc",
        "outputId": "2600c35b-5dc7-42f6-f3ef-cf45cea73096"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Accuracy: 0.7829120125416422\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.27      0.84      0.41      2275\n",
            "           1       0.98      0.78      0.87     23240\n",
            "\n",
            "    accuracy                           0.78     25515\n",
            "   macro avg       0.62      0.81      0.64     25515\n",
            "weighted avg       0.92      0.78      0.83     25515\n",
            "\n"
          ]
        }
      ],
      "source": [
        "#Import scikit-learn metrics module for accuracy calculation\n",
        "from sklearn import metrics\n",
        "# Model Accuracy, how often is the classifier correct?\n",
        "print(\"Accuracy:\",metrics.accuracy_score(y_test, y_pred))\n",
        "print(metrics.classification_report(y_test, y_pred))\n",
        "#print(\"roc_auc:\",metrics.roc_auc_score(y_test, y_pred))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 30,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 281
        },
        "id": "-7_QR2-7gVKB",
        "outputId": "c48f1d2e-cbf5-4fbb-f3ab-0f967d17133a"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "conf_mat = confusion_matrix(y_test, y_pred)\n",
        "plt.title(\"Confusion Matrix\")\n",
        "sns.heatmap(conf_mat,annot=True,fmt='.0f')\n",
        "plt.show()\n",
        "#plt.savefig(output_path+'confusion_matrix.png')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "KiDtPQ5DjtPU"
      },
      "outputs": [],
      "source": [
        "#unique, counts = np.unique(y_pred, return_counts=True)\n",
        "#dict(zip(unique, counts))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uN6ApYjrIZLM"
      },
      "source": [
        "### *RF without best paramether method*"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "CGN9K9jkI7Ut"
      },
      "outputs": [],
      "source": [
        "'''\n",
        "from sklearn.linear_model import Ridge\n",
        "from sklearn.model_selection import GridSearchCV"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "iT9DgKRjI-Dl"
      },
      "outputs": [],
      "source": [
        "'''\n",
        "parameters = {'n_estimators': [750],'criterion':['gini', 'entropy']}"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "SHH4SgYeJAWk",
        "outputId": "b8b6ccae-2073-4394-e578-6ddd094c8061"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Accuracy Test: 0.9973049281314168\n"
          ]
        }
      ],
      "source": [
        "'''\n",
        "model = GridSearchCV(\n",
        "      RandomForestClassifier(),\n",
        "      param_grid=parameters,\n",
        "      n_jobs=10, cv=10)\n",
        "model.fit(X_train,y_train)\n",
        "rm_pred=model.predict(X_test)\n",
        "print(\"Accuracy Test:\",metrics.accuracy_score(y_test, rm_pred))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "SZ7BQbdtNGeH"
      },
      "outputs": [],
      "source": [
        "'''\n",
        "!pip install scikit-plot\n",
        "import scikitplot.metrics as splt\n",
        "splt.plot_confusion_matrix(y_test, rm_pred)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "kigYuh3kN3ub",
        "outputId": "853b017a-b40b-405a-84d4-6f3d8ac3adcb"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "           1       1.00      0.16      0.28        25\n",
            "           2       1.00      1.00      1.00      7767\n",
            "\n",
            "    accuracy                           1.00      7792\n",
            "   macro avg       1.00      0.58      0.64      7792\n",
            "weighted avg       1.00      1.00      1.00      7792\n",
            "\n"
          ]
        }
      ],
      "source": [
        "'''\n",
        "print(metrics.classification_report(y_test, rm_pred))"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# ⬛ Save and load the model"
      ],
      "metadata": {
        "id": "JrlB5wyDLm3l"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#salva il modello\n",
        "joblib.dump(best_rf, path_stations + \"RF_hypoxia_undersample.joblib\")\n",
        "#joblib.dump(best_rf, path + \"random_forest_full_pt_norm.joblib\")"
      ],
      "metadata": {
        "id": "YIvm-MZlkPs9",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "da76f958-18d8-488c-f54a-84ecbd5c2b64"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['/content/gdrive/My Drive/Dati/DatiOriginali/Lagoon_WaterQuality/Elaborati/Dataset_hypoxia/Hourly/RF_hypoxia_undersample.joblib']"
            ]
          },
          "metadata": {},
          "execution_count": 37
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "best_rf = joblib.load(path_stations + \"RF_hypoxia_undersample.joblib\")"
      ],
      "metadata": {
        "id": "iOinCxIukuli"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "--------------------------------------------------------------------------------\n",
        "# ⛔ FOR THODORIS, STOP HERE :)"
      ],
      "metadata": {
        "id": "XFG6-6q692Ij"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Random Forest for unbalanced datasets"
      ],
      "metadata": {
        "id": "8YiRrkFU_ZyE"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XWBJZZFz4NK7"
      },
      "source": [
        "### Random Forest for Imbalanced Classification"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "NaFQ_1SZ4PqH"
      },
      "outputs": [],
      "source": [
        "# class balanced random forest for imbalanced classification\n",
        "from numpy import mean\n",
        "from sklearn.datasets import make_classification\n",
        "from sklearn.model_selection import cross_val_score\n",
        "from sklearn.model_selection import RepeatedStratifiedKFold\n",
        "from sklearn.ensemble import RandomForestClassifier\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "2dDFNdn19xF9"
      },
      "outputs": [],
      "source": [
        "# define model\n",
        "model = RandomForestClassifier(n_estimators=10, class_weight='balanced')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "IOp-MV379YZ5"
      },
      "outputs": [],
      "source": [
        "# define evaluation procedure\n",
        "cv = RepeatedStratifiedKFold(n_splits=10, n_repeats=3, random_state=1)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "bhLmH7Dq9aP1"
      },
      "outputs": [],
      "source": [
        "# evaluate model\n",
        "scores = cross_val_score(model, X_train, y_train, scoring='roc_auc', cv=cv, n_jobs=-1)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZQeG2tEw9byK",
        "outputId": "08bb6164-4244-4772-f647-f9036468195a"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mean ROC AUC: 0.992\n"
          ]
        }
      ],
      "source": [
        "# summarize performance\n",
        "print('Mean ROC AUC: %.3f' % mean(scores))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qMBrwzAl9pwv"
      },
      "source": [
        "### *Class_weight argument to the value ‘balanced_subsample*"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "2I67YKZr9tKT"
      },
      "outputs": [],
      "source": [
        "model = RandomForestClassifier(n_estimators=10, class_weight='balanced_subsample')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "baFh68o49z5u"
      },
      "outputs": [],
      "source": [
        "# define evaluation procedure\n",
        "cv = RepeatedStratifiedKFold(n_splits=10, n_repeats=3, random_state=1)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "CzSgfPTe92Pw"
      },
      "outputs": [],
      "source": [
        "# evaluate model\n",
        "scores = cross_val_score(model, X_train, y_train, scoring='roc_auc', cv=cv, n_jobs=-1)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "l5S__35X92wg",
        "outputId": "fcb484af-657a-4cd0-af62-898ba8acad3c"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Mean ROC AUC: 0.764\n"
          ]
        }
      ],
      "source": [
        "# summarize performance\n",
        "print('Mean ROC AUC: %.3f' % mean(scores))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Euiu_bBm-7Vf"
      },
      "source": [
        "### *BalancedRandomForestClassifier class* "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "jSioSeWx_SWn"
      },
      "outputs": [],
      "source": [
        "from imblearn.ensemble import BalancedRandomForestClassifier"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "vi24cJo0-962"
      },
      "outputs": [],
      "source": [
        "# define model\n",
        "model = BalancedRandomForestClassifier(n_estimators=10)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "4Btabobt_A5K"
      },
      "outputs": [],
      "source": [
        "# define evaluation procedure\n",
        "cv = RepeatedStratifiedKFold(n_splits=10, n_repeats=3, random_state=1)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "bG0SGBvc_D1Z"
      },
      "outputs": [],
      "source": [
        "# evaluate model\n",
        "scores = cross_val_score(model, X_train, y_train, scoring='roc_auc', cv=cv, n_jobs=-1)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "AMbhw5Eo_ILL",
        "outputId": "12075537-f625-4622-8681-9ebb1f4f6d64"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mean ROC AUC: 0.879\n"
          ]
        }
      ],
      "source": [
        "# summarize performance\n",
        "print('Mean ROC AUC: %.3f' % mean(scores))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QEngG3ZsEESD"
      },
      "source": [
        "# COMPARE CLASSIFIERS"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "pR-D391bENxN"
      },
      "outputs": [],
      "source": [
        "from sklearn.neural_network import MLPClassifier\n",
        "from sklearn.neighbors import KNeighborsClassifier\n",
        "from sklearn.svm import SVC\n",
        "from sklearn.gaussian_process import GaussianProcessClassifier\n",
        "from sklearn.ensemble import GradientBoostingClassifier\n",
        "from sklearn.gaussian_process.kernels import RBF\n",
        "from sklearn.tree import DecisionTreeClassifier\n",
        "from sklearn.ensemble import ExtraTreesClassifier\n",
        "from sklearn.ensemble import RandomForestClassifier, AdaBoostClassifier\n",
        "from sklearn.naive_bayes import GaussianNB\n",
        "from sklearn.discriminant_analysis import QuadraticDiscriminantAnalysis\n",
        "from sklearn.linear_model import SGDClassifier"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GPBV24kIEYlF"
      },
      "source": [
        "Defining learning classifiers"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "g4bnrGYVEQWW"
      },
      "outputs": [],
      "source": [
        "names = [\"Linear_SVM\", \"Polynomial_SVM\", \"RBF_SVM\",  \"Neural_Net\", \"AdaBoost\",\n",
        "         \"Naive_Bayes\"]\n",
        "\n",
        "classifiers = [\n",
        "    SVC(kernel=\"linear\", C=0.025),\n",
        "    SVC(kernel=\"poly\", degree=3, C=0.025),\n",
        "    SVC(kernel=\"rbf\", C=1, gamma=2),\n",
        "    MLPClassifier(alpha=1, max_iter=1000),\n",
        "    AdaBoostClassifier(n_estimators=100),\n",
        "    GaussianNB()]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "B2Rn6VDTRdMF"
      },
      "outputs": [],
      "source": [
        "model_ada = AdaBoostClassifier(n_estimators=100)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4PJkl_dfEZwC"
      },
      "source": [
        "Build Model, Apply Model on Test Data & Record Accuracy Scores"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "vcXF5VjpEb9b"
      },
      "outputs": [],
      "source": [
        "scores = []\n",
        "for name, clf in zip(names, classifiers):\n",
        "    clf.fit(X_train, y_train)\n",
        "    score = clf.score(X_test, y_test)\n",
        "    scores.append(score)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CsqzXO-7Efb7",
        "outputId": "69c53086-77f1-42f0-92dc-4de14f5af28a"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "[0.9883213552361396,\n",
              " 0.9883213552361396,\n",
              " 0.9883213552361396,\n",
              " 0.9883213552361396,\n",
              " 0.9879363449691991,\n",
              " 0.8803901437371663]"
            ]
          },
          "execution_count": 17,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "scores"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "8ikxYXp-RmqT"
      },
      "outputs": [],
      "source": [
        "model_ada.fit(X_train, y_train)\n",
        "score_model_ada = model_ada.score(X_test, y_test) "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "L1lIdbHnSFme",
        "outputId": "7e082c43-d528-46b4-908f-072c19d180f5"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "0.9879363449691991"
            ]
          },
          "execution_count": 29,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "score_model_ada"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7G_VXfG9FVTE"
      },
      "source": [
        "Analysis of Model Performance"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "woOMqLvmFXgP"
      },
      "outputs": [],
      "source": [
        "import seaborn as sns"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6eX2GMGeFaBr"
      },
      "source": [
        "Create data frame of model performance"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 238
        },
        "id": "Ur4bHSu0Fbzb",
        "outputId": "d2e1e4f0-2138-45a1-90ea-4f0a07f8b025"
      },
      "outputs": [
        {
          "data": {
            "text/html": [
              "\n",
              "  <div id=\"df-ae7b26f2-a950-4a78-af89-ff4e47561af8\">\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>name</th>\n",
              "      <th>score</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Linear_SVM</td>\n",
              "      <td>0.988321</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Polynomial_SVM</td>\n",
              "      <td>0.988321</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>RBF_SVM</td>\n",
              "      <td>0.988321</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>Neural_Net</td>\n",
              "      <td>0.988321</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>AdaBoost</td>\n",
              "      <td>0.987936</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>Naive_Bayes</td>\n",
              "      <td>0.880390</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-ae7b26f2-a950-4a78-af89-ff4e47561af8')\"\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-ae7b26f2-a950-4a78-af89-ff4e47561af8 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-ae7b26f2-a950-4a78-af89-ff4e47561af8');\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",
              "  "
            ],
            "text/plain": [
              "             name     score\n",
              "0      Linear_SVM  0.988321\n",
              "1  Polynomial_SVM  0.988321\n",
              "2         RBF_SVM  0.988321\n",
              "3      Neural_Net  0.988321\n",
              "4        AdaBoost  0.987936\n",
              "5     Naive_Bayes  0.880390"
            ]
          },
          "execution_count": 19,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "df = pd.DataFrame()\n",
        "df['name'] = names\n",
        "df['score'] = scores\n",
        "df"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oYZ46A6IFgfm"
      },
      "source": [
        "Adding colors to the data frame"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 238
        },
        "id": "FliNCeEmFjEL",
        "outputId": "c26e656e-ba05-4d81-8039-56f064fad3c9"
      },
      "outputs": [
        {
          "data": {
            "text/html": [
              "<style type=\"text/css\">\n",
              "#T_d09c3_row0_col1, #T_d09c3_row1_col1, #T_d09c3_row2_col1, #T_d09c3_row3_col1, #T_d09c3_row4_col1 {\n",
              "  background-color: #008000;\n",
              "  color: #f1f1f1;\n",
              "}\n",
              "#T_d09c3_row5_col1 {\n",
              "  background-color: #ebf3eb;\n",
              "  color: #000000;\n",
              "}\n",
              "</style>\n",
              "<table id=\"T_d09c3_\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr>\n",
              "      <th class=\"blank level0\" >&nbsp;</th>\n",
              "      <th class=\"col_heading level0 col0\" >name</th>\n",
              "      <th class=\"col_heading level0 col1\" >score</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th id=\"T_d09c3_level0_row0\" class=\"row_heading level0 row0\" >0</th>\n",
              "      <td id=\"T_d09c3_row0_col0\" class=\"data row0 col0\" >Linear_SVM</td>\n",
              "      <td id=\"T_d09c3_row0_col1\" class=\"data row0 col1\" >0.988321</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th id=\"T_d09c3_level0_row1\" class=\"row_heading level0 row1\" >1</th>\n",
              "      <td id=\"T_d09c3_row1_col0\" class=\"data row1 col0\" >Polynomial_SVM</td>\n",
              "      <td id=\"T_d09c3_row1_col1\" class=\"data row1 col1\" >0.988321</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th id=\"T_d09c3_level0_row2\" class=\"row_heading level0 row2\" >2</th>\n",
              "      <td id=\"T_d09c3_row2_col0\" class=\"data row2 col0\" >RBF_SVM</td>\n",
              "      <td id=\"T_d09c3_row2_col1\" class=\"data row2 col1\" >0.988321</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th id=\"T_d09c3_level0_row3\" class=\"row_heading level0 row3\" >3</th>\n",
              "      <td id=\"T_d09c3_row3_col0\" class=\"data row3 col0\" >Neural_Net</td>\n",
              "      <td id=\"T_d09c3_row3_col1\" class=\"data row3 col1\" >0.988321</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th id=\"T_d09c3_level0_row4\" class=\"row_heading level0 row4\" >4</th>\n",
              "      <td id=\"T_d09c3_row4_col0\" class=\"data row4 col0\" >AdaBoost</td>\n",
              "      <td id=\"T_d09c3_row4_col1\" class=\"data row4 col1\" >0.987936</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th id=\"T_d09c3_level0_row5\" class=\"row_heading level0 row5\" >5</th>\n",
              "      <td id=\"T_d09c3_row5_col0\" class=\"data row5 col0\" >Naive_Bayes</td>\n",
              "      <td id=\"T_d09c3_row5_col1\" class=\"data row5 col1\" >0.880390</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n"
            ],
            "text/plain": [
              "<pandas.io.formats.style.Styler at 0x7f332b7add10>"
            ]
          },
          "execution_count": 20,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "#https://pandas.pydata.org/pandas-docs/stable/user_guide/style.html\n",
        "\n",
        "cm = sns.light_palette(\"green\", as_cmap=True)\n",
        "s = df.style.background_gradient(cmap=cm)\n",
        "s"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RjYdW4yMFk6O"
      },
      "source": [
        "Bar plot of model performance"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 285
        },
        "id": "ncMCTCJAFma-",
        "outputId": "adf86958-1cd6-499b-bc4b-05259791338c"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "sns.set(style=\"whitegrid\")\n",
        "ax = sns.barplot(y=\"name\", x=\"score\", data=df)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "iqtr6kMwSUGj"
      },
      "outputs": [],
      "source": [
        "#create prediction basing on the test set\n",
        "y_pred = model_ada.predict(X_test)\n",
        "ytrain_pred = model_ada.predict(X_train)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tYLL1PVYJyQy"
      },
      "source": [
        "vorrei capire se da qui posso ricavarmi le accuraci di ogni modello"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vhDjZ69DJy4i",
        "outputId": "60f1476c-23ec-4361-8ff4-6f82ba2a4ac8"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "           1       0.20      0.01      0.02        91\n",
            "           2       0.99      1.00      0.99      7701\n",
            "\n",
            "    accuracy                           0.99      7792\n",
            "   macro avg       0.59      0.51      0.51      7792\n",
            "weighted avg       0.98      0.99      0.98      7792\n",
            "\n"
          ]
        }
      ],
      "source": [
        "print(metrics.classification_report(y_test, y_pred))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 284
        },
        "id": "HR3a3oKNJy4i",
        "outputId": "f9488cc8-c895-4cec-ab46-766238bbfd72"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "conf_mat = confusion_matrix(y_test, y_pred)\n",
        "plt.title(\"Confusion Matrix\")\n",
        "sns.heatmap(conf_mat,annot=True,fmt='.0f')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "c6WyNGZDXClf"
      },
      "source": [
        "# LOGISTIC REGRESSION"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "6qEoH6LnW9os"
      },
      "outputs": [],
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.linear_model import LogisticRegression\n",
        "\n",
        "logistic = LogisticRegression(class_weight=\"balanced\").fit(X_train, y_train)\n",
        "y_pred = logistic.predict(X_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Qj5rOfNZ6Cn1",
        "outputId": "9c71e311-aecb-424e-cb8b-fc88307bccdd"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "In-sample accuracy: 0.712\n",
            "Out-of-sample accuracy: 0.710\n"
          ]
        }
      ],
      "source": [
        "print('In-sample accuracy: %0.3f' %\n",
        "      accuracy_score(y_train, logistic.predict(X_train)))\n",
        "print('Out-of-sample accuracy: %0.3f' %\n",
        "      accuracy_score(y_test, y_pred))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5o_A6YuGYhsu",
        "outputId": "08955cdb-5ce9-428b-fea2-50df4846239e"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.13      0.83      0.22      9418\n",
            "           1       0.99      0.70      0.82    182205\n",
            "\n",
            "    accuracy                           0.71    191623\n",
            "   macro avg       0.56      0.77      0.52    191623\n",
            "weighted avg       0.95      0.71      0.79    191623\n",
            "\n"
          ]
        }
      ],
      "source": [
        "print(metrics.classification_report(y_test, y_pred))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 281
        },
        "id": "GGvUprScYnmh",
        "outputId": "f13b5135-5148-4b7d-83f7-f2b4daf357e9"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "conf_mat = confusion_matrix(y_test, y_pred)\n",
        "plt.title(\"Confusion Matrix\")\n",
        "sns.heatmap(conf_mat,annot=True,fmt='.0f')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ANv-vLI5Y98i"
      },
      "outputs": [],
      "source": [
        "# Evaluate a score by cross-validation.\n",
        "scores = cross_val_score(logistic, X_train, y_train, scoring='roc_auc', cv=None, n_jobs=-1)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "uEzT4NCoY98j",
        "outputId": "f7a27150-8b03-4c47-8955-cea7e9bb1a6d"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mean ROC AUC: 0.842\n"
          ]
        }
      ],
      "source": [
        "# summarize performance\n",
        "print('Mean ROC AUC: %.3f' % mean(scores))"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#ROC CURVE\n",
        "from sklearn.metrics import RocCurveDisplay\n",
        "best_rf_disp = RocCurveDisplay.from_estimator(logistic, X_test, y_test)\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        },
        "id": "t-ADidqYP7jp",
        "outputId": "3c245b52-afc6-438f-af0f-1102a62b7dde"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-fgxcPpF7a1T"
      },
      "source": [
        "## Probability of each sample to belong to a class"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pLFzF73e8svV",
        "outputId": "b73c6826-b1cb-4ebd-f893-06673671a622"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "18.9361005783081 :   0.021\n",
            "18.9599990844727 :   0.072\n",
            "19.2988998413086 :   0.010\n",
            "20.8289999008179 :   0.010\n",
            "22.9874 :  -0.018\n",
            "22.7599 :   0.000\n",
            "22.6284 :   0.064\n",
            "22.6179 :  -0.240\n",
            "22.6129 :   0.080\n",
            "22.8765 :  -0.000\n"
          ]
        }
      ],
      "source": [
        "for var, coef in zip(dataset.w_temperature,\n",
        "                     logistic.coef_[0]):\n",
        "  print(\"%7s : %7.3f\" %(var, coef))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "klM1hc727vfk",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "aa61d12b-ab18-43b8-a0f2-71645b510ca3"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "classes: [0 1]\n",
            "\n",
            "Probs:\n",
            " [[0.4135112  0.5864888 ]\n",
            " [0.27948137 0.72051863]\n",
            " [0.51688145 0.48311855]\n",
            " [0.0362484  0.9637516 ]\n",
            " [0.58320854 0.41679146]\n",
            " [0.30024669 0.69975331]\n",
            " [0.02479241 0.97520759]\n",
            " [0.22634438 0.77365562]\n",
            " [0.07051642 0.92948358]\n",
            " [0.78834059 0.21165941]\n",
            " [0.09907678 0.90092322]\n",
            " [0.45986062 0.54013938]\n",
            " [0.08292662 0.91707338]\n",
            " [0.05305949 0.94694051]\n",
            " [0.83742423 0.16257577]\n",
            " [0.62773425 0.37226575]\n",
            " [0.19872659 0.80127341]\n",
            " [0.00317303 0.99682697]\n",
            " [0.06518817 0.93481183]\n",
            " [0.73362338 0.26637662]\n",
            " [0.52907152 0.47092848]\n",
            " [0.67768082 0.32231918]\n",
            " [0.59998445 0.40001555]\n",
            " [0.03475235 0.96524765]\n",
            " [0.35402924 0.64597076]\n",
            " [0.84776646 0.15223354]\n",
            " [0.24561857 0.75438143]\n",
            " [0.62135331 0.37864669]\n",
            " [0.32542888 0.67457112]\n",
            " [0.36860162 0.63139838]\n",
            " [0.40137691 0.59862309]\n",
            " [0.26425081 0.73574919]\n",
            " [0.43354116 0.56645884]\n",
            " [0.73692746 0.26307254]\n",
            " [0.74895475 0.25104525]\n",
            " [0.70320151 0.29679849]\n",
            " [0.07620577 0.92379423]\n",
            " [0.0314166  0.9685834 ]\n",
            " [0.0920795  0.9079205 ]\n",
            " [0.11111913 0.88888087]\n",
            " [0.02495119 0.97504881]\n",
            " [0.02916047 0.97083953]\n",
            " [0.10974496 0.89025504]\n",
            " [0.51712638 0.48287362]\n",
            " [0.02520925 0.97479075]\n",
            " [0.14988246 0.85011754]\n",
            " [0.0496441  0.9503559 ]\n",
            " [0.93387083 0.06612917]\n",
            " [0.20317539 0.79682461]\n",
            " [0.03147237 0.96852763]\n",
            " [0.10452178 0.89547822]\n",
            " [0.06100804 0.93899196]\n",
            " [0.00983086 0.99016914]\n",
            " [0.10981324 0.89018676]\n",
            " [0.82076884 0.17923116]\n",
            " [0.24394176 0.75605824]\n",
            " [0.00678494 0.99321506]\n",
            " [0.75222936 0.24777064]\n",
            " [0.16242789 0.83757211]\n",
            " [0.52950184 0.47049816]\n",
            " [0.23940351 0.76059649]\n",
            " [0.32270595 0.67729405]\n",
            " [0.70721864 0.29278136]\n",
            " [0.17662206 0.82337794]\n",
            " [0.4987184  0.5012816 ]\n",
            " [0.00960801 0.99039199]\n",
            " [0.121434   0.878566  ]\n",
            " [0.19965545 0.80034455]\n",
            " [0.66712313 0.33287687]\n",
            " [0.52120068 0.47879932]\n",
            " [0.36451171 0.63548829]\n",
            " [0.10116857 0.89883143]\n",
            " [0.13682032 0.86317968]\n",
            " [0.27117239 0.72882761]\n",
            " [0.42642254 0.57357746]\n",
            " [0.02439032 0.97560968]\n",
            " [0.75793904 0.24206096]\n",
            " [0.72932654 0.27067346]\n",
            " [0.80832089 0.19167911]\n",
            " [0.73082727 0.26917273]\n",
            " [0.03364681 0.96635319]\n",
            " [0.15225894 0.84774106]\n",
            " [0.12800512 0.87199488]\n",
            " [0.89049968 0.10950032]\n",
            " [0.27901286 0.72098714]\n",
            " [0.12110819 0.87889181]\n",
            " [0.05049426 0.94950574]\n",
            " [0.45073705 0.54926295]\n",
            " [0.10116014 0.89883986]\n",
            " [0.77365181 0.22634819]\n",
            " [0.63761866 0.36238134]\n",
            " [0.16633721 0.83366279]\n",
            " [0.25806131 0.74193869]\n",
            " [0.59527301 0.40472699]\n",
            " [0.81740525 0.18259475]\n",
            " [0.03279418 0.96720582]\n",
            " [0.02075079 0.97924921]\n",
            " [0.40414972 0.59585028]\n",
            " [0.67291386 0.32708614]\n",
            " [0.2914171  0.7085829 ]\n",
            " [0.04619684 0.95380316]\n",
            " [0.25561328 0.74438672]\n",
            " [0.01826104 0.98173896]\n",
            " [0.27861956 0.72138044]\n",
            " [0.13426185 0.86573815]\n",
            " [0.77596525 0.22403475]\n",
            " [0.53557699 0.46442301]\n",
            " [0.13443199 0.86556801]\n",
            " [0.29119014 0.70880986]\n",
            " [0.02274516 0.97725484]\n",
            " [0.33779139 0.66220861]\n",
            " [0.32133444 0.67866556]\n",
            " [0.79739408 0.20260592]\n",
            " [0.07994661 0.92005339]\n",
            " [0.02542091 0.97457909]\n",
            " [0.08142235 0.91857765]\n",
            " [0.82541104 0.17458896]\n",
            " [0.04088714 0.95911286]\n",
            " [0.11854551 0.88145449]\n",
            " [0.01284753 0.98715247]\n",
            " [0.02155343 0.97844657]\n",
            " [0.03672154 0.96327846]\n",
            " [0.82060189 0.17939811]\n",
            " [0.71215221 0.28784779]\n",
            " [0.83507842 0.16492158]\n",
            " [0.23183253 0.76816747]\n",
            " [0.0254607  0.9745393 ]\n",
            " [0.13351945 0.86648055]\n",
            " [0.03963582 0.96036418]\n",
            " [0.30007293 0.69992707]\n",
            " [0.10788245 0.89211755]\n",
            " [0.03297352 0.96702648]\n",
            " [0.05183515 0.94816485]\n",
            " [0.51149989 0.48850011]\n",
            " [0.58534498 0.41465502]\n",
            " [0.20564481 0.79435519]\n",
            " [0.47275568 0.52724432]\n",
            " [0.79397362 0.20602638]\n",
            " [0.20516489 0.79483511]\n",
            " [0.48598809 0.51401191]\n",
            " [0.44672087 0.55327913]\n",
            " [0.1049789  0.8950211 ]\n",
            " [0.85100062 0.14899938]\n",
            " [0.03387748 0.96612252]\n",
            " [0.31971732 0.68028268]\n",
            " [0.77589045 0.22410955]\n",
            " [0.23126203 0.76873797]\n",
            " [0.01733455 0.98266545]\n",
            " [0.07765997 0.92234003]\n",
            " [0.17720771 0.82279229]\n",
            " [0.42641723 0.57358277]\n",
            " [0.17868241 0.82131759]\n",
            " [0.4003441  0.5996559 ]\n",
            " [0.3700226  0.6299774 ]\n",
            " [0.61156326 0.38843674]\n",
            " [0.02947705 0.97052295]\n",
            " [0.03083913 0.96916087]\n",
            " [0.51769392 0.48230608]\n",
            " [0.63631224 0.36368776]\n",
            " [0.57672669 0.42327331]\n",
            " [0.12448802 0.87551198]\n",
            " [0.01141245 0.98858755]\n",
            " [0.14988773 0.85011227]\n",
            " [0.05894128 0.94105872]\n",
            " [0.85129195 0.14870805]\n",
            " [0.07305678 0.92694322]\n",
            " [0.29490455 0.70509545]\n",
            " [0.04011863 0.95988137]\n",
            " [0.01721    0.98279   ]\n",
            " [0.0096403  0.9903597 ]\n",
            " [0.6978482  0.3021518 ]\n",
            " [0.52409603 0.47590397]\n",
            " [0.04307546 0.95692454]\n",
            " [0.51885004 0.48114996]\n",
            " [0.4984881  0.5015119 ]\n",
            " [0.00927119 0.99072881]\n",
            " [0.19600209 0.80399791]\n",
            " [0.47025294 0.52974706]\n",
            " [0.62865232 0.37134768]\n",
            " [0.16160955 0.83839045]\n",
            " [0.27634994 0.72365006]\n",
            " [0.46350484 0.53649516]\n",
            " [0.30159821 0.69840179]\n",
            " [0.09943416 0.90056584]\n",
            " [0.00751153 0.99248847]\n",
            " [0.7472025  0.2527975 ]\n",
            " [0.1090917  0.8909083 ]\n",
            " [0.08073853 0.91926147]\n",
            " [0.08525564 0.91474436]\n",
            " [0.76450107 0.23549893]\n",
            " [0.13582328 0.86417672]\n",
            " [0.67985256 0.32014744]\n",
            " [0.06291972 0.93708028]\n",
            " [0.18775593 0.81224407]\n",
            " [0.08611976 0.91388024]\n",
            " [0.03938016 0.96061984]\n",
            " [0.17686622 0.82313378]\n",
            " [0.24043127 0.75956873]\n",
            " [0.1197696  0.8802304 ]\n",
            " [0.75135766 0.24864234]\n",
            " [0.05531804 0.94468196]\n",
            " [0.92738799 0.07261201]\n",
            " [0.02780063 0.97219937]\n",
            " [0.01624914 0.98375086]\n",
            " [0.50408491 0.49591509]\n",
            " [0.06262753 0.93737247]\n",
            " [0.11976794 0.88023206]\n",
            " [0.31848698 0.68151302]\n",
            " [0.72584251 0.27415749]\n",
            " [0.70288257 0.29711743]\n",
            " [0.89930912 0.10069088]\n",
            " [0.07976877 0.92023123]\n",
            " [0.22522925 0.77477075]\n",
            " [0.87416596 0.12583404]\n",
            " [0.04148214 0.95851786]\n",
            " [0.12167721 0.87832279]\n",
            " [0.91998356 0.08001644]\n",
            " [0.73241971 0.26758029]\n",
            " [0.80269794 0.19730206]\n",
            " [0.13434387 0.86565613]\n",
            " [0.04482497 0.95517503]\n",
            " [0.28019238 0.71980762]\n",
            " [0.06872605 0.93127395]\n",
            " [0.75497474 0.24502526]\n",
            " [0.12266959 0.87733041]\n",
            " [0.89288313 0.10711687]\n",
            " [0.18223626 0.81776374]\n",
            " [0.69867081 0.30132919]\n",
            " [0.3162093  0.6837907 ]\n",
            " [0.02409524 0.97590476]\n",
            " [0.76810607 0.23189393]\n",
            " [0.88116632 0.11883368]\n",
            " [0.05178441 0.94821559]\n",
            " [0.7483788  0.2516212 ]\n",
            " [0.84412043 0.15587957]\n",
            " [0.02980541 0.97019459]\n",
            " [0.11148869 0.88851131]\n",
            " [0.21982458 0.78017542]\n",
            " [0.04130326 0.95869674]\n",
            " [0.01484917 0.98515083]\n",
            " [0.15100516 0.84899484]\n",
            " [0.85759598 0.14240402]\n",
            " [0.3834429  0.6165571 ]\n",
            " [0.8253896  0.1746104 ]\n",
            " [0.02138454 0.97861546]\n",
            " [0.78549013 0.21450987]\n",
            " [0.05872513 0.94127487]\n",
            " [0.04754859 0.95245141]\n",
            " [0.05396012 0.94603988]\n",
            " [0.42563436 0.57436564]\n",
            " [0.24227046 0.75772954]\n",
            " [0.81249812 0.18750188]\n",
            " [0.72890476 0.27109524]\n",
            " [0.01197193 0.98802807]\n",
            " [0.08334837 0.91665163]\n",
            " [0.05423507 0.94576493]\n",
            " [0.02672575 0.97327425]\n",
            " [0.05276482 0.94723518]\n",
            " [0.48014975 0.51985025]\n",
            " [0.02255662 0.97744338]\n",
            " [0.45312644 0.54687356]\n",
            " [0.03012285 0.96987715]\n",
            " [0.06664588 0.93335412]\n",
            " [0.68445455 0.31554545]\n",
            " [0.01533478 0.98466522]\n",
            " [0.03948265 0.96051735]\n",
            " [0.20064072 0.79935928]\n",
            " [0.75278948 0.24721052]\n",
            " [0.76434368 0.23565632]\n",
            " [0.69678886 0.30321114]\n",
            " [0.226155   0.773845  ]\n",
            " [0.62344437 0.37655563]\n",
            " [0.07874017 0.92125983]\n",
            " [0.03436037 0.96563963]\n",
            " [0.59928727 0.40071273]\n",
            " [0.46092786 0.53907214]\n",
            " [0.28140435 0.71859565]\n",
            " [0.05212501 0.94787499]\n",
            " [0.02503644 0.97496356]\n",
            " [0.13754853 0.86245147]\n",
            " [0.02471823 0.97528177]\n",
            " [0.02330952 0.97669048]\n",
            " [0.03725642 0.96274358]\n",
            " [0.72970953 0.27029047]\n",
            " [0.06529918 0.93470082]\n",
            " [0.31484109 0.68515891]\n",
            " [0.41870973 0.58129027]\n",
            " [0.32738736 0.67261264]\n",
            " [0.03992199 0.96007801]\n",
            " [0.70234122 0.29765878]\n",
            " [0.41560274 0.58439726]\n",
            " [0.6076625  0.3923375 ]\n",
            " [0.66223129 0.33776871]\n",
            " [0.09056381 0.90943619]\n",
            " [0.13140372 0.86859628]\n",
            " [0.50861779 0.49138221]\n",
            " [0.02900929 0.97099071]\n",
            " [0.66801888 0.33198112]\n",
            " [0.47101677 0.52898323]\n",
            " [0.09792388 0.90207612]\n",
            " [0.03006466 0.96993534]\n",
            " [0.07467814 0.92532186]\n",
            " [0.52005478 0.47994522]\n",
            " [0.03323827 0.96676173]\n",
            " [0.48430567 0.51569433]\n",
            " [0.26249545 0.73750455]\n",
            " [0.2456181  0.7543819 ]\n",
            " [0.32114929 0.67885071]\n",
            " [0.12901234 0.87098766]\n",
            " [0.20281426 0.79718574]\n",
            " [0.68508659 0.31491341]\n",
            " [0.11573374 0.88426626]\n",
            " [0.02953712 0.97046288]\n",
            " [0.57397121 0.42602879]\n",
            " [0.17952362 0.82047638]\n",
            " [0.14967582 0.85032418]\n",
            " [0.05771191 0.94228809]\n",
            " [0.62361241 0.37638759]\n",
            " [0.59142826 0.40857174]\n",
            " [0.12129457 0.87870543]\n",
            " [0.83892547 0.16107453]\n",
            " [0.10048058 0.89951942]\n",
            " [0.0131416  0.9868584 ]\n",
            " [0.29200873 0.70799127]\n",
            " [0.07511174 0.92488826]\n",
            " [0.27946677 0.72053323]\n",
            " [0.4672795  0.5327205 ]\n",
            " [0.01025591 0.98974409]\n",
            " [0.59087289 0.40912711]\n",
            " [0.13878633 0.86121367]\n",
            " [0.03528984 0.96471016]\n",
            " [0.07747154 0.92252846]\n",
            " [0.74867627 0.25132373]\n",
            " [0.17625633 0.82374367]\n",
            " [0.46821554 0.53178446]\n",
            " [0.71861101 0.28138899]\n",
            " [0.86048904 0.13951096]\n",
            " [0.5020428  0.4979572 ]\n",
            " [0.05464795 0.94535205]\n",
            " [0.1713315  0.8286685 ]\n",
            " [0.12425876 0.87574124]\n",
            " [0.37227659 0.62772341]\n",
            " [0.56526878 0.43473122]\n",
            " [0.77229495 0.22770505]\n",
            " [0.5102696  0.4897304 ]\n",
            " [0.01437176 0.98562824]\n",
            " [0.03692562 0.96307438]\n",
            " [0.27223814 0.72776186]\n",
            " [0.04224512 0.95775488]\n",
            " [0.32265133 0.67734867]\n",
            " [0.04708738 0.95291262]\n",
            " [0.17385957 0.82614043]\n",
            " [0.02668333 0.97331667]\n",
            " [0.65417345 0.34582655]\n",
            " [0.16070685 0.83929315]\n",
            " [0.3431886  0.6568114 ]\n",
            " [0.06330895 0.93669105]\n",
            " [0.02100778 0.97899222]\n",
            " [0.05745386 0.94254614]\n",
            " [0.19143939 0.80856061]\n",
            " [0.78473406 0.21526594]\n",
            " [0.74425529 0.25574471]\n",
            " [0.23912894 0.76087106]\n",
            " [0.4991741  0.5008259 ]\n",
            " [0.68488776 0.31511224]]\n"
          ]
        }
      ],
      "source": [
        "print('\\nclasses:', logistic.classes_)\n",
        "print('\\nProbs:\\n', logistic.predict_proba(X_test)[:365,:])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IgKOLcZeDxCo"
      },
      "source": [
        "# WEIGHTED SVM "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "NGy5EuqlES6C"
      },
      "outputs": [],
      "source": [
        "# fit a svm on an imbalanced classification dataset\n",
        "from numpy import mean\n",
        "from sklearn.datasets import make_classification\n",
        "from sklearn.model_selection import cross_val_score\n",
        "from sklearn.model_selection import RepeatedStratifiedKFold\n",
        "from sklearn.svm import SVC"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "tdssvb3oD0Bq"
      },
      "outputs": [],
      "source": [
        "# define model\n",
        "#weights = {0:100.0, 1:1.0}\n",
        "SVM_model = SVC(gamma='scale', class_weight='balanced')  # , class_weight='balanced'"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "haPolehbEa9S"
      },
      "outputs": [],
      "source": [
        "# define evaluation procedure\n",
        "cv = RepeatedStratifiedKFold(n_splits=10, n_repeats=3, random_state=1)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "jeTqbIXhE9Bk"
      },
      "outputs": [],
      "source": [
        "# evaluate model\n",
        "scores = cross_val_score(SVM_model, X_t, y, scoring='roc_auc', cv=cv, n_jobs=-1)\n",
        "# summarize performance\n",
        "print('Mean ROC AUC: %.3f' % mean(scores))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "gIjmQp8NFmpt",
        "outputId": "6f849b00-a633-490e-9719-28ff5aa4a033"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "SVC(class_weight='balanced')"
            ]
          },
          "execution_count": 28,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "SVM_model.fit(X_train, y_train)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "2wlWv9mBFTQM"
      },
      "outputs": [],
      "source": [
        "y_pred = SVM_model.predict(X_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "c5TWu1uCFTQN",
        "outputId": "391c8ebc-0068-43a5-ec44-616ea050ec9e"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "In-sample accuracy: 0.616\n",
            "Out-of-sample accuracy: 0.600\n"
          ]
        }
      ],
      "source": [
        "print('In-sample accuracy: %0.3f' %\n",
        "      accuracy_score(y_train, SVM_model.predict(X_train)))\n",
        "print('Out-of-sample accuracy: %0.3f' %\n",
        "      accuracy_score(y_test, y_pred))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ocXpx7arFTQO",
        "outputId": "bd8afee3-8f42-4771-91b1-f3fcfd783a8d"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.05      1.00      0.09        20\n",
            "           1       1.00      0.59      0.74      1033\n",
            "\n",
            "    accuracy                           0.60      1053\n",
            "   macro avg       0.52      0.80      0.42      1053\n",
            "weighted avg       0.98      0.60      0.73      1053\n",
            "\n"
          ]
        }
      ],
      "source": [
        "print(metrics.classification_report(y_test, y_pred))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 281
        },
        "id": "L78AWV7jFTQO",
        "outputId": "3a315a8a-051f-45dd-f563-59d454b8cec4"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "conf_mat = confusion_matrix(y_test, y_pred)\n",
        "plt.title(\"Confusion Matrix\")\n",
        "sns.heatmap(conf_mat,annot=True,fmt='.0f')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Bagging Classifier for imbalanced classification"
      ],
      "metadata": {
        "id": "zy0y-1lW8y4p"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "oYvNC2R01OmT"
      },
      "outputs": [],
      "source": [
        "# bagged decision trees on an imbalanced classification problem\n",
        "from numpy import mean\n",
        "from sklearn.datasets import make_classification\n",
        "from sklearn.model_selection import cross_val_score\n",
        "from sklearn.model_selection import RepeatedStratifiedKFold\n",
        "from sklearn.ensemble import BaggingClassifier"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "6WUEIaEnz6jp"
      },
      "outputs": [],
      "source": [
        "...\n",
        "# define model\n",
        "model = BaggingClassifier()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Flicbzkx0Nf4"
      },
      "outputs": [],
      "source": [
        "...\n",
        "# define evaluation procedure\n",
        "cv = RepeatedStratifiedKFold(n_splits=10, n_repeats=3, random_state=1)\n",
        "# evaluate model\n",
        "scores = cross_val_score(model, X, y, scoring='roc_auc', cv=cv, n_jobs=-1)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ltc-ZOo60YOW",
        "outputId": "a82f0c32-32d8-4443-e756-e6386167910b"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Mean ROC AUC: 0.999\n"
          ]
        }
      ],
      "source": [
        "# summarize performance\n",
        "print('Mean ROC AUC: %.3f' % mean(scores))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LWKZ95fN10TO"
      },
      "source": [
        "# BalancedBaggingClassifier class."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "j4sVNrm_16JS"
      },
      "outputs": [],
      "source": [
        "# bagged decision trees with random undersampling for imbalanced classification\n",
        "from numpy import mean\n",
        "from sklearn.datasets import make_classification\n",
        "from sklearn.model_selection import cross_val_score\n",
        "from sklearn.model_selection import RepeatedStratifiedKFold\n",
        "from imblearn.ensemble import BalancedBaggingClassifier"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "VYhdLpv71-L1"
      },
      "outputs": [],
      "source": [
        "# define model\n",
        "model = BalancedBaggingClassifier()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ZqLBiCvj2A_j"
      },
      "outputs": [],
      "source": [
        "# define evaluation procedure\n",
        "cv = RepeatedStratifiedKFold(n_splits=10, n_repeats=3, random_state=1)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ewk4xjBC2EyE",
        "outputId": "31f406a3-13ee-4f6b-be38-38762225e941"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:372: FitFailedWarning: \n",
            "30 fits failed out of a total of 30.\n",
            "The score on these train-test partitions for these parameters will be set to nan.\n",
            "If these failures are not expected, you can try to debug them by setting error_score='raise'.\n",
            "\n",
            "Below are more details about the failures:\n",
            "--------------------------------------------------------------------------------\n",
            "30 fits failed with the following error:\n",
            "Traceback (most recent call last):\n",
            "  File \"/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py\", line 680, in _fit_and_score\n",
            "    estimator.fit(X_train, y_train, **fit_params)\n",
            "  File \"/usr/local/lib/python3.7/dist-packages/imblearn/ensemble/_bagging.py\", line 321, in fit\n",
            "    return self._fit(X, y, self.max_samples, sample_weight=None)\n",
            "  File \"/usr/local/lib/python3.7/dist-packages/sklearn/ensemble/_bagging.py\", line 337, in _fit\n",
            "    max_features = self.max_features * self.n_features_in_\n",
            "AttributeError: 'BalancedBaggingClassifier' object has no attribute 'n_features_in_'\n",
            "\n",
            "  warnings.warn(some_fits_failed_message, FitFailedWarning)\n"
          ]
        }
      ],
      "source": [
        "# evaluate model\n",
        "scores = cross_val_score(model, X_t, y, scoring='roc_auc', cv=cv, n_jobs=-1)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MJBrPVS-2FYQ",
        "outputId": "73ad2b26-f818-403b-c339-36e75c2d0035"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mean ROC AUC: nan\n"
          ]
        }
      ],
      "source": [
        "# summarize performance\n",
        "print('Mean ROC AUC: %.3f' % mean(scores))"
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "P0y5rv0SGoyn"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## EasyEnsembleClassifier"
      ],
      "metadata": {
        "id": "Ev6J0qLuGhvj"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "f8-jFWH1GGQL"
      },
      "outputs": [],
      "source": [
        "# easy ensemble for imbalanced classification\n",
        "from numpy import mean\n",
        "from sklearn.datasets import make_classification\n",
        "from sklearn.model_selection import cross_val_score\n",
        "from sklearn.model_selection import RepeatedStratifiedKFold\n",
        "from imblearn.ensemble import EasyEnsembleClassifier"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "VY5OzwwlHyQP"
      },
      "outputs": [],
      "source": [
        "# define model\n",
        "model = EasyEnsembleClassifier(n_estimators=10)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "1ug1YqG-HvTf"
      },
      "outputs": [],
      "source": [
        "# define evaluation procedure\n",
        "cv = RepeatedStratifiedKFold(n_splits=10, n_repeats=3, random_state=1)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "rvlPC-lrHEzH",
        "outputId": "aee3bec5-0c75-4f54-ca6f-820a44970925"
      },
      "outputs": [
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:372: FitFailedWarning: \n",
            "30 fits failed out of a total of 30.\n",
            "The score on these train-test partitions for these parameters will be set to nan.\n",
            "If these failures are not expected, you can try to debug them by setting error_score='raise'.\n",
            "\n",
            "Below are more details about the failures:\n",
            "--------------------------------------------------------------------------------\n",
            "30 fits failed with the following error:\n",
            "Traceback (most recent call last):\n",
            "  File \"/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py\", line 680, in _fit_and_score\n",
            "    estimator.fit(X_train, y_train, **fit_params)\n",
            "  File \"/usr/local/lib/python3.7/dist-packages/imblearn/ensemble/_easy_ensemble.py\", line 224, in fit\n",
            "    return self._fit(X, y, self.max_samples, sample_weight=None)\n",
            "  File \"/usr/local/lib/python3.7/dist-packages/sklearn/ensemble/_bagging.py\", line 337, in _fit\n",
            "    max_features = self.max_features * self.n_features_in_\n",
            "AttributeError: 'EasyEnsembleClassifier' object has no attribute 'n_features_in_'\n",
            "\n",
            "  warnings.warn(some_fits_failed_message, FitFailedWarning)\n"
          ]
        }
      ],
      "source": [
        "# evaluate model\n",
        "scores = cross_val_score(model, X_t, y, scoring='roc_auc', cv=cv, n_jobs=-1)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "OUeuwsjhGQBO",
        "outputId": "1980c42e-85ff-443b-bea9-6123a771dc9e"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Mean ROC AUC: nan\n"
          ]
        }
      ],
      "source": [
        "# summarize performance\n",
        "print('Mean ROC AUC: %.3f' % mean(scores))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "lHqnvny6-BAi",
        "outputId": "3b72b6d8-376e-44fa-9d8c-5d5a8ff5a349"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "0        1\n",
              "1        1\n",
              "2        1\n",
              "3        1\n",
              "4        1\n",
              "        ..\n",
              "26767    1\n",
              "26768    1\n",
              "26769    1\n",
              "26770    1\n",
              "26771    1\n",
              "Name: hypoxia_label, Length: 26772, dtype: int64"
            ]
          },
          "execution_count": 29,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "y"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4wNFjyZpEK5Q"
      },
      "source": [
        "# FIT THE MODEL AND SEE THE PERFORMANCES"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "otG4KZD-E1_1",
        "outputId": "2d601de8-94ef-4200-e300-108b2b1ba177"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "RandomForestClassifier(class_weight='balanced', n_estimators=10)"
            ]
          },
          "metadata": {},
          "execution_count": 60
        }
      ],
      "source": [
        "# Fit the random search model\n",
        "model.fit(X_train, y_train)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "E4G3U5qYFGHy"
      },
      "outputs": [],
      "source": [
        "#create prediction basing on the test set\n",
        "y_pred = model.predict(X_test)\n",
        "ytrain_pred = model.predict(X_train)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8kd7wjkhFjij",
        "outputId": "1f94b197-7e3c-492b-bd2e-522ac305f7dd"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Accuracy test:  0.9213559958877587\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.34      0.62      0.44      9418\n",
            "           1       0.98      0.94      0.96    182205\n",
            "\n",
            "    accuracy                           0.92    191623\n",
            "   macro avg       0.66      0.78      0.70    191623\n",
            "weighted avg       0.95      0.92      0.93    191623\n",
            "\n"
          ]
        }
      ],
      "source": [
        "print(\"Accuracy test: \", accuracy_score(y_test, y_pred))\n",
        "print(metrics.classification_report(y_test, y_pred))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 281
        },
        "id": "BGKVQIZKFyK8",
        "outputId": "3584def5-e0ae-4dc9-bcb3-31d0fc16e2a9"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "conf_mat = confusion_matrix(y_test, y_pred)\n",
        "plt.title(\"Confusion Matrix\")\n",
        "sns.heatmap(conf_mat,annot=True,fmt='.0f')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#ROC CURVE\n",
        "from sklearn.metrics import RocCurveDisplay\n",
        "best_rf_disp = RocCurveDisplay.from_estimator(model, X_test, y_test)\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        },
        "id": "99riy-J8_BVb",
        "outputId": "6abc297a-27ea-4175-e349-df58e9f1c967"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    }
  ],
  "metadata": {
    "colab": {
      "collapsed_sections": [
        "GfD5PH0aYTke",
        "lym_MuxDXqyn",
        "8YiRrkFU_ZyE",
        "XWBJZZFz4NK7",
        "qMBrwzAl9pwv",
        "c6WyNGZDXClf",
        "4wNFjyZpEK5Q"
      ],
      "provenance": [],
      "toc_visible": true,
      "include_colab_link": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}