3B. Modelling experiments.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "waterpumps - modelling experimental.ipynb",
      "provenance": [],
      "toc_visible": true,
      "authorship_tag": "ABX9TyOM6fJ2Meco/4RooK36jJM8",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/BrendaLoznik/waterpumps/blob/main/3B.%20Modelling%20experiments.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 1 Housekeeping"
      ],
      "metadata": {
        "id": "EBX12nVV7mCb"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 1.1 Load libraries"
      ],
      "metadata": {
        "id": "c15-Gokb7p9N"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "pip install catboost"
      ],
      "metadata": {
        "id": "QWdpeQ8P7qz6"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "%matplotlib inline\n",
        "import seaborn as sns\n",
        "\n",
        "from sklearn.preprocessing import StandardScaler\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.model_selection import GridSearchCV, RandomizedSearchCV\n",
        "from sklearn.metrics import confusion_matrix, classification_report, accuracy_score\n",
        "from sklearn.model_selection import KFold\n",
        "from sklearn.model_selection import cross_val_score\n",
        "\n",
        "#machine learning\n",
        "from sklearn.ensemble import StackingClassifier\n",
        "from sklearn.linear_model import LogisticRegression \n",
        "from sklearn.neighbors import KNeighborsClassifier\n",
        "from sklearn.tree import DecisionTreeClassifier\n",
        "from xgboost import XGBClassifier\n",
        "from sklearn.ensemble import BaggingClassifier\n",
        "from sklearn.ensemble import RandomForestClassifier\n",
        "from sklearn.ensemble import AdaBoostClassifier\n",
        "from sklearn.ensemble import GradientBoostingClassifier\n",
        "from sklearn.ensemble import VotingClassifier\n",
        "from catboost import CatBoostClassifier\n",
        "from sklearn.cluster import KMeans \n",
        "from sklearn.linear_model import LogisticRegression \n",
        "\n",
        "import warnings\n",
        "warnings.filterwarnings(\"ignore\")\n",
        "warnings.simplefilter(action='ignore', category=FutureWarning)"
      ],
      "metadata": {
        "id": "JjuIgxcx7sUK"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 1.2 Custom functions"
      ],
      "metadata": {
        "id": "gb7rtzx57w0a"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def confusion_matrix_plotter (y_test,y_pred):\n",
        "  fig, ax = plt.subplots(figsize=(8,5)) \n",
        "  data = confusion_matrix(y_test, y_pred)\n",
        "  df_cm = pd.DataFrame(data, columns=np.unique(y_test), index = np.unique(y_test))\n",
        "  ax = sns.heatmap(df_cm, cmap='Blues', fmt='g' ,annot=True,annot_kws={\"size\": 14})\n",
        "  ax.set_xlabel(\"Predicted\")\n",
        "  ax.set_ylabel (\"Actual\")\n",
        "  ax.set_xticklabels(ax.get_xticklabels(), rotation=45)\n",
        "  ax.set_yticklabels(ax.get_xticklabels(), rotation=0)"
      ],
      "metadata": {
        "id": "WNvlCxOu7u8y"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def feature_importances (model):\n",
        "  features= X_train.columns\n",
        "  importances = np.round(model.feature_importances_,3) \n",
        "  indices = np.argsort(importances)\n",
        "  num_features = len(features)\n",
        "  fig, ax = plt.subplots(figsize=(10,10)) \n",
        "  plt.barh(range(num_features), importances[indices[-num_features:]],  align='center')\n",
        "  plt.yticks(range(num_features), [features[i] for i in indices[-num_features:]])\n",
        "  plt.xlabel('Relative Importance')\n",
        "  plt.title('Feature Importance')\n",
        "plt.show()"
      ],
      "metadata": {
        "id": "MreGYsjU7uhS"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 1.3 Load data"
      ],
      "metadata": {
        "id": "eClak7a373q7"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive \n",
        "drive.mount('/content/gdrive')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CJS9VPhe70ci",
        "outputId": "9d5e05be-3653-4491-96b4-340ce7d1389a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/gdrive\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "train_labels = pd.read_csv('/content/gdrive/MyDrive/Kaggle/Waterpump-github/train_labels.csv')\n",
        "train_values = pd.read_csv('/content/gdrive/MyDrive/Kaggle/Waterpump-github/train_values.csv', parse_dates = ['date_recorded' ],  na_values = [0, '0'])\n",
        "test_values = pd.read_csv('/content/gdrive/MyDrive/Kaggle/Waterpump-github/test_values.csv', parse_dates = ['date_recorded' ] , na_values = [0, '0'])\n",
        "final_df = pd.read_csv('/content/gdrive/MyDrive/Kaggle/Waterpump-github/final_df.csv')\n",
        "final_df = final_df.drop('Unnamed: 0', axis=1)"
      ],
      "metadata": {
        "id": "5q6Wi6H275Oe"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "final_df.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 270
        },
        "id": "8oYBqJ3678Na",
        "outputId": "890d4e65-17a4-4f14-cd91-29b6280310c8"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "\n",
              "  <div id=\"df-ad221c84-2e69-490c-86c5-8b479a682bf1\">\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>id</th>\n",
              "      <th>status_group</th>\n",
              "      <th>funder</th>\n",
              "      <th>installer</th>\n",
              "      <th>latitude</th>\n",
              "      <th>basin</th>\n",
              "      <th>region_code</th>\n",
              "      <th>public_meeting</th>\n",
              "      <th>extraction_type</th>\n",
              "      <th>extraction_type_class</th>\n",
              "      <th>payment_type</th>\n",
              "      <th>water_quality</th>\n",
              "      <th>quantity</th>\n",
              "      <th>source</th>\n",
              "      <th>waterpoint_type</th>\n",
              "      <th>longitude</th>\n",
              "      <th>permit</th>\n",
              "      <th>scheme_management</th>\n",
              "      <th>gps_height</th>\n",
              "      <th>population</th>\n",
              "      <th>age</th>\n",
              "      <th>season</th>\n",
              "      <th>amount_tsh_missing</th>\n",
              "      <th>region_district</th>\n",
              "      <th>authority_scheme</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>69572</td>\n",
              "      <td>functional</td>\n",
              "      <td>107</td>\n",
              "      <td>108</td>\n",
              "      <td>-9.86</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>6</td>\n",
              "      <td>1</td>\n",
              "      <td>8</td>\n",
              "      <td>1</td>\n",
              "      <td>34.94</td>\n",
              "      <td>0</td>\n",
              "      <td>7</td>\n",
              "      <td>1390</td>\n",
              "      <td>109</td>\n",
              "      <td>12</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>20</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>8776</td>\n",
              "      <td>functional</td>\n",
              "      <td>45</td>\n",
              "      <td>47</td>\n",
              "      <td>-2.15</td>\n",
              "      <td>4</td>\n",
              "      <td>12</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>6</td>\n",
              "      <td>2</td>\n",
              "      <td>5</td>\n",
              "      <td>1</td>\n",
              "      <td>34.70</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>1399</td>\n",
              "      <td>280</td>\n",
              "      <td>3</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>55</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>34310</td>\n",
              "      <td>functional</td>\n",
              "      <td>95</td>\n",
              "      <td>148</td>\n",
              "      <td>-3.82</td>\n",
              "      <td>5</td>\n",
              "      <td>13</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>5</td>\n",
              "      <td>6</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>37.46</td>\n",
              "      <td>1</td>\n",
              "      <td>7</td>\n",
              "      <td>686</td>\n",
              "      <td>250</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>52</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>67743</td>\n",
              "      <td>non functional</td>\n",
              "      <td>136</td>\n",
              "      <td>129</td>\n",
              "      <td>-11.16</td>\n",
              "      <td>7</td>\n",
              "      <td>25</td>\n",
              "      <td>1</td>\n",
              "      <td>11</td>\n",
              "      <td>5</td>\n",
              "      <td>2</td>\n",
              "      <td>6</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>38.49</td>\n",
              "      <td>1</td>\n",
              "      <td>7</td>\n",
              "      <td>263</td>\n",
              "      <td>58</td>\n",
              "      <td>27</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>76</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>19728</td>\n",
              "      <td>functional</td>\n",
              "      <td>95</td>\n",
              "      <td>7</td>\n",
              "      <td>-1.83</td>\n",
              "      <td>4</td>\n",
              "      <td>9</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>6</td>\n",
              "      <td>3</td>\n",
              "      <td>5</td>\n",
              "      <td>1</td>\n",
              "      <td>31.13</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>1328</td>\n",
              "      <td>532</td>\n",
              "      <td>15</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>22</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-ad221c84-2e69-490c-86c5-8b479a682bf1')\"\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-ad221c84-2e69-490c-86c5-8b479a682bf1 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-ad221c84-2e69-490c-86c5-8b479a682bf1');\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": [
              "      id    status_group  ...  region_district  authority_scheme\n",
              "0  69572      functional  ...               20                 1\n",
              "1   8776      functional  ...               55                 2\n",
              "2  34310      functional  ...               52                 1\n",
              "3  67743  non functional  ...               76                 1\n",
              "4  19728      functional  ...               22                 2\n",
              "\n",
              "[5 rows x 25 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 2 Modelling"
      ],
      "metadata": {
        "id": "1iWnwAXy7_is"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#define train and test \n",
        "train_set = final_df [final_df [\"id\"].isin(train_values[\"id\"])]\n",
        "test_set =  final_df  [final_df  [\"id\"].isin(test_values[\"id\"])]"
      ],
      "metadata": {
        "id": "qMQMzeHO79jq"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#define X and y for training the model\n",
        "X= train_set.drop(['id', 'status_group'], axis=1)\n",
        "y = train_set['status_group']"
      ],
      "metadata": {
        "id": "KOYm_F458BAi"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### **2.1 Best submission**"
      ],
      "metadata": {
        "id": "AzAre9Az8ImF"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "####**Baseline model performance**"
      ],
      "metadata": {
        "id": "qwzFnyfZAUWn"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#define the models\n",
        "models=[LogisticRegression(max_iter = 1000000,  random_state=42),\n",
        "        KNeighborsClassifier(),\n",
        "        DecisionTreeClassifier( random_state=42),\n",
        "       RandomForestClassifier( random_state=42),\n",
        "       XGBClassifier( random_state=42),\n",
        "        BaggingClassifier( random_state=42), \n",
        "        AdaBoostClassifier( random_state=42),\n",
        "        GradientBoostingClassifier( random_state=42),\n",
        "       CatBoostClassifier( random_state=42)]\n",
        "\n",
        "#create a list of model names\n",
        "names =['Logistic Regression', 'KNN Classifier', 'Decision Tree Classifier', 'Random Forest Classifier',  'XGB Classifier', 'Bagging Classifier', 'AdaBoost Classifier', 'Gradient Boosting Classifier', 'Catboost Classifier']\n",
        "\n",
        "#create empty lists for the accuracy and standard deviation\n",
        "accuracy=[]\n",
        "std = []\n",
        "\n",
        "#peform 10-fold cross validaton on each model and append the results to the list\n",
        "for model in models:\n",
        "  cv = KFold(n_splits=10, shuffle = True,  random_state=1)\n",
        "  n_scores = cross_val_score(model, X, y, scoring='accuracy', cv=cv, n_jobs=-1, error_score='raise')\n",
        "  mean_score = np.mean(n_scores)\n",
        "  std_score = np.std(n_scores)\n",
        "  accuracy.append(mean_score)\n",
        "  std.append(std_score)\n",
        "\n",
        "#create a dataframe of the model performances\n",
        "d = {'Model':names,'Accuracy':accuracy, 'Std':std}   \n",
        "score_df = pd.DataFrame(d)\n",
        "score_df"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 331
        },
        "id": "-5LzwUB6A_4k",
        "outputId": "ca2e12f2-696f-44ad-e6a5-80da1452d834"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "\n",
              "  <div id=\"df-28bba190-410c-47a6-87dd-368ccb06c6cc\">\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>Model</th>\n",
              "      <th>Accuracy</th>\n",
              "      <th>Std</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Logistic Regression</td>\n",
              "      <td>0.666510</td>\n",
              "      <td>0.006001</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>KNN Classifier</td>\n",
              "      <td>0.701140</td>\n",
              "      <td>0.007301</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Decision Tree Classifier</td>\n",
              "      <td>0.756158</td>\n",
              "      <td>0.004169</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>Random Forest Classifier</td>\n",
              "      <td>0.804121</td>\n",
              "      <td>0.004219</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>XGB Classifier</td>\n",
              "      <td>0.742740</td>\n",
              "      <td>0.005622</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>Bagging Classifier</td>\n",
              "      <td>0.793599</td>\n",
              "      <td>0.003323</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>AdaBoost Classifier</td>\n",
              "      <td>0.723985</td>\n",
              "      <td>0.005083</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>Gradient Boosting Classifier</td>\n",
              "      <td>0.751174</td>\n",
              "      <td>0.005385</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>Catboost Classifier</td>\n",
              "      <td>0.796428</td>\n",
              "      <td>0.003599</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-28bba190-410c-47a6-87dd-368ccb06c6cc')\"\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-28bba190-410c-47a6-87dd-368ccb06c6cc 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-28bba190-410c-47a6-87dd-368ccb06c6cc');\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": [
              "                          Model  Accuracy       Std\n",
              "0           Logistic Regression  0.666510  0.006001\n",
              "1                KNN Classifier  0.701140  0.007301\n",
              "2      Decision Tree Classifier  0.756158  0.004169\n",
              "3      Random Forest Classifier  0.804121  0.004219\n",
              "4                XGB Classifier  0.742740  0.005622\n",
              "5            Bagging Classifier  0.793599  0.003323\n",
              "6           AdaBoost Classifier  0.723985  0.005083\n",
              "7  Gradient Boosting Classifier  0.751174  0.005385\n",
              "8           Catboost Classifier  0.796428  0.003599"
            ]
          },
          "metadata": {},
          "execution_count": 36
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "####**Hyperparameter range plots**"
      ],
      "metadata": {
        "id": "j6pgp-cUK5N6"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Visualize the optimal number of n_estimators for a random forest and a bagging classifier."
      ],
      "metadata": {
        "id": "W0gZ6RFeMOPu"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#define X and y for training the model\n",
        "X= train_set.drop(['id', 'status_group'], axis=1)\n",
        "y = train_set['status_group']\n",
        "X_train, X_test, y_train, y_test= train_test_split(X, y, test_size = 0.2, random_state=42, stratify=y)\n",
        "\n",
        "#determine accuracy for different levels of n_estimators\n",
        "n_estimators = [1,  25,  50, 100, 250, 500, 750, 1000]\n",
        "train_results = []\n",
        "test_results = []\n",
        "\n",
        "for estimator in n_estimators:\n",
        "   rf = RandomForestClassifier(n_estimators=estimator, n_jobs=-1)\n",
        "   rf.fit(X_train, y_train)\n",
        "  #train accuracy\n",
        "   train_pred = rf.predict(X_train)\n",
        "   train_results.append(accuracy_score(y_train,train_pred))\n",
        "  #test accuracy\n",
        "   test_pred = rf.predict(X_test)\n",
        "   test_results.append(accuracy_score(y_test, test_pred))\n",
        "\n",
        "#plot  \n",
        "from matplotlib.legend_handler import HandlerLine2D\n",
        "line1, = plt.plot(n_estimators, train_results, color = 'k', label=\"Train\")\n",
        "line2, = plt.plot(n_estimators, test_results, color =  '#66c2a5', label=\"Test\")\n",
        "plt.legend(handler_map={line1: HandlerLine2D(numpoints=2)})\n",
        "plt.ylabel(\"Accuracy score\")\n",
        "plt.xlabel(\"n_estimators\")\n",
        "plt.title(\"Random Forest Classifier\")\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 296
        },
        "id": "FXOnEufQKw_T",
        "outputId": "81601fbb-1f26-494f-f64d-db5514ebeaa5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### **Best model**"
      ],
      "metadata": {
        "id": "k2Z8AtYDBtnA"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "My best model resulted in an accuracy of 0.8235 on the public board (rank 504). This model consisted of a weighted vote of 4 tuned tree-based models."
      ],
      "metadata": {
        "id": "QJ3VHH1OBDQz"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#tuned models\n",
        "rf = RandomForestClassifier(max_depth=30, max_features='log2', min_samples_split=7, n_estimators=225, random_state=42, warm_start=True)\n",
        "xgb = XGBClassifier(max_depth =12, n_estimators= 200, colsample_bytree = 0.3, eta = 0.2, random_state = 42) \n",
        "cat = CatBoostClassifier(max_depth = 9, rsm = 0.2, iterations = 1000, random_state = 42)\n",
        "bag = BaggingClassifier(max_features=0.4, n_estimators=1000, random_state=42) "
      ],
      "metadata": {
        "id": "xxgSdOgJ8HuJ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Create training and test sets: test size 0.2\n",
        "X_train, X_test, y_train, y_test= train_test_split(X, y, test_size = 0.2, random_state=42, stratify=y)\n",
        "\n",
        "\n",
        "vc = VotingClassifier([('rf', rf), ('xgb', xgb),('cat', cat), ('bag', bag)], voting = 'soft', weights = [0.85, 1, 0.85, 0.75])\n",
        "                       \n",
        "#fit and predict\n",
        "vc.fit(X_train , y_train)\n",
        "pred= vc .predict(X_test)\n",
        "pred_train = vc .predict(X_train)\n"
      ],
      "metadata": {
        "id": "FbfGV0eK8XJk"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#print best model scores on test data\n",
        "print(\"Accuracy score train: {}\".format(accuracy_score(y_train ,pred_train)))\n",
        "print(\"Accuracy score test: {}\".format(accuracy_score(y_test, pred)))\n",
        "\n",
        "#print(confusion_matrix(y_test, y_pred))\n",
        "print(classification_report(y_test, pred))\n",
        "\n",
        "#heatmap\n",
        "fig, ax = plt.subplots(figsize=(8,5)) \n",
        "data = confusion_matrix(y_test, pred)\n",
        "df_cm = pd.DataFrame(data, columns=np.unique(y_test), index = np.unique(y_test))\n",
        "ax = sns.heatmap(df_cm, cmap='Blues', fmt='g' ,annot=True,annot_kws={\"size\": 14})\n",
        "ax.set_xlabel(\"Predicted\")\n",
        "ax.set_ylabel (\"Actual\")\n",
        "ax.set_xticklabels(ax.get_xticklabels(), rotation=45)\n",
        "ax.set_yticklabels(ax.get_xticklabels(), rotation=0)\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 620
        },
        "id": "f-M8Mofa8cmh",
        "outputId": "573b4590-652a-4dd1-b6c0-94b1c19f1ca5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Accuracy score train: 0.9242197857699026\n",
            "Accuracy score test: 0.8159090909090909\n",
            "                         precision    recall  f1-score   support\n",
            "\n",
            "             functional       0.80      0.91      0.85      6452\n",
            "functional needs repair       0.64      0.28      0.39       863\n",
            "         non functional       0.86      0.78      0.82      4565\n",
            "\n",
            "               accuracy                           0.82     11880\n",
            "              macro avg       0.77      0.66      0.69     11880\n",
            "           weighted avg       0.81      0.82      0.81     11880\n",
            "\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### **2.2 Experiment 1 - Kmeans**"
      ],
      "metadata": {
        "id": "346GYola8h9l"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Is it possible to cluster the water pumps based on their characteristics? For each cluster I can calculate the percentage of functional, non-functional and pumps that need repairing. Can I improve model performance my including these percentages as variables to the dataset?"
      ],
      "metadata": {
        "id": "kdq7wkd3_QBQ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#create a copy of the dataset\n",
        "train_set_kmeans = train_set.copy()\n",
        "\n",
        "#define X and y for training the model\n",
        "X= train_set_kmeans.drop(['id', 'status_group'], axis=1)\n",
        "y = train_set_kmeans['status_group']"
      ],
      "metadata": {
        "id": "JLojDCuJBQYl"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# transform data\n",
        "scaler = StandardScaler()\n",
        "scaled_features = scaler.fit_transform(X)\n",
        "\n",
        "#create scaled df\n",
        "scaled_data = pd.DataFrame(scaled_features , index=X.index, columns=X.columns)"
      ],
      "metadata": {
        "id": "jMLJMjvS8dbJ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#Create an ellbow plot to determine the optimal number of clusters\n",
        "ks = range(1, 20)\n",
        "inertias = []\n",
        "\n",
        "for k in ks:\n",
        "    # Create a KMeans instance with k clusters: model\n",
        "    kmeans = KMeans(n_clusters=k).fit(scaled_data)\n",
        "     \n",
        "    # Append the inertia to the list of inertias\n",
        "    inertias.append(kmeans.inertia_)\n",
        "    \n",
        "# Plot ks vs inertias\n",
        "plt.plot(ks, inertias, '-o',  color =  '#66c2a5')\n",
        "plt.xlabel('number of clusters, k')\n",
        "plt.ylabel('inertia')\n",
        "plt.title(\"Ellbow method\")\n",
        "plt.xticks(ks)\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "aFQnsRGg8_WZ",
        "outputId": "c38997e9-8d0c-4e53-a674-da3185b1ee9b"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#fit K means with 3 clusters\n",
        "kmeans = KMeans(n_clusters=3, random_state =11 )\n",
        "kmeans.fit(scaled_data)\n",
        "\n",
        "# Calculate the cluster labels: labels\n",
        "clusters = kmeans.predict(scaled_data)\n",
        "train_set_kmeans[\"cluster\"] = clusters"
      ],
      "metadata": {
        "id": "AewGIduH967_"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#calculate percentage of status groups per cluster\n",
        "grouped = train_set_kmeans.groupby(['cluster', 'status_group'])['id'].count().reset_index()\n",
        "\n",
        "pivot = grouped.pivot(index = ['cluster'], columns = 'status_group')\n",
        "mi=pivot .columns\n",
        "mi.tolist()\n",
        "ind = pd.Index([ str(e[1])  for e in mi.tolist()])\n",
        "pivot .columns = ind\n",
        "\n",
        "pivot['Pumps (#)'] = pivot['functional'] + pivot['non functional'] + pivot['functional needs repair']\n",
        "pivot['Functional pumps (%)'] = round(pivot['functional'] / pivot['Pumps (#)'],2)\n",
        "pivot['Non-functional pumps (%)'] = round(pivot['non functional'] / pivot['Pumps (#)'],2)\n",
        "pivot['Functional need repair pumps (%)'] = round(pivot['functional needs repair'] / pivot['Pumps (#)'],2)\n",
        "\n",
        "pivot = pivot.drop(['functional', 'functional needs repair', 'non functional'],axis=1)\n",
        "pivot = pivot.reset_index()\n",
        "pivot"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 144
        },
        "id": "UnJ7Gd8__aOc",
        "outputId": "aa7a56e9-2978-4268-ed33-882275258031"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "\n",
              "  <div id=\"df-ca53b725-e1f6-4629-bb0b-cdadfed7746c\">\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>cluster</th>\n",
              "      <th>Pumps (#)</th>\n",
              "      <th>Functional pumps (%)</th>\n",
              "      <th>Non-functional pumps (%)</th>\n",
              "      <th>Functional need repair pumps (%)</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0</td>\n",
              "      <td>13923</td>\n",
              "      <td>0.48</td>\n",
              "      <td>0.47</td>\n",
              "      <td>0.04</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1</td>\n",
              "      <td>23291</td>\n",
              "      <td>0.62</td>\n",
              "      <td>0.29</td>\n",
              "      <td>0.09</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2</td>\n",
              "      <td>22185</td>\n",
              "      <td>0.50</td>\n",
              "      <td>0.43</td>\n",
              "      <td>0.07</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-ca53b725-e1f6-4629-bb0b-cdadfed7746c')\"\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-ca53b725-e1f6-4629-bb0b-cdadfed7746c 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-ca53b725-e1f6-4629-bb0b-cdadfed7746c');\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": [
              "   cluster  ...  Functional need repair pumps (%)\n",
              "0        0  ...                              0.04\n",
              "1        1  ...                              0.09\n",
              "2        2  ...                              0.07\n",
              "\n",
              "[3 rows x 5 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 40
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#add the cluster means to the dataframes\n",
        "train_set_kmeans = train_set_kmeans.merge(pivot[['cluster', 'Functional pumps (%)', 'Non-functional pumps (%)', 'Functional need repair pumps (%)']], how = 'left', on = 'cluster')\n",
        "train_set_kmeans.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 321
        },
        "id": "7wVqYstg_kmW",
        "outputId": "827218bf-6841-4f1f-fb46-7e3e6fcc3efe"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "\n",
              "  <div id=\"df-ed4c2dcf-24f9-4a4d-9bd2-190be74fbc56\">\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>id</th>\n",
              "      <th>status_group</th>\n",
              "      <th>funder</th>\n",
              "      <th>installer</th>\n",
              "      <th>latitude</th>\n",
              "      <th>basin</th>\n",
              "      <th>region_code</th>\n",
              "      <th>public_meeting</th>\n",
              "      <th>extraction_type</th>\n",
              "      <th>extraction_type_class</th>\n",
              "      <th>payment_type</th>\n",
              "      <th>water_quality</th>\n",
              "      <th>quantity</th>\n",
              "      <th>source</th>\n",
              "      <th>waterpoint_type</th>\n",
              "      <th>longitude</th>\n",
              "      <th>permit</th>\n",
              "      <th>scheme_management</th>\n",
              "      <th>gps_height</th>\n",
              "      <th>population</th>\n",
              "      <th>age</th>\n",
              "      <th>season</th>\n",
              "      <th>amount_tsh_missing</th>\n",
              "      <th>region_district</th>\n",
              "      <th>authority_scheme</th>\n",
              "      <th>cluster</th>\n",
              "      <th>Functional pumps (%)</th>\n",
              "      <th>Non-functional pumps (%)</th>\n",
              "      <th>Functional need repair pumps (%)</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>69572</td>\n",
              "      <td>functional</td>\n",
              "      <td>107</td>\n",
              "      <td>108</td>\n",
              "      <td>-9.86</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>6</td>\n",
              "      <td>1</td>\n",
              "      <td>8</td>\n",
              "      <td>1</td>\n",
              "      <td>34.94</td>\n",
              "      <td>0</td>\n",
              "      <td>7</td>\n",
              "      <td>1390</td>\n",
              "      <td>109</td>\n",
              "      <td>12</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>20</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0.62</td>\n",
              "      <td>0.29</td>\n",
              "      <td>0.09</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>8776</td>\n",
              "      <td>functional</td>\n",
              "      <td>45</td>\n",
              "      <td>47</td>\n",
              "      <td>-2.15</td>\n",
              "      <td>4</td>\n",
              "      <td>12</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>6</td>\n",
              "      <td>2</td>\n",
              "      <td>5</td>\n",
              "      <td>1</td>\n",
              "      <td>34.70</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>1399</td>\n",
              "      <td>280</td>\n",
              "      <td>3</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>55</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>0.62</td>\n",
              "      <td>0.29</td>\n",
              "      <td>0.09</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>34310</td>\n",
              "      <td>functional</td>\n",
              "      <td>95</td>\n",
              "      <td>148</td>\n",
              "      <td>-3.82</td>\n",
              "      <td>5</td>\n",
              "      <td>13</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>5</td>\n",
              "      <td>6</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>37.46</td>\n",
              "      <td>1</td>\n",
              "      <td>7</td>\n",
              "      <td>686</td>\n",
              "      <td>250</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>52</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0.48</td>\n",
              "      <td>0.47</td>\n",
              "      <td>0.04</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>67743</td>\n",
              "      <td>non functional</td>\n",
              "      <td>136</td>\n",
              "      <td>129</td>\n",
              "      <td>-11.16</td>\n",
              "      <td>7</td>\n",
              "      <td>25</td>\n",
              "      <td>1</td>\n",
              "      <td>11</td>\n",
              "      <td>5</td>\n",
              "      <td>2</td>\n",
              "      <td>6</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>38.49</td>\n",
              "      <td>1</td>\n",
              "      <td>7</td>\n",
              "      <td>263</td>\n",
              "      <td>58</td>\n",
              "      <td>27</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>76</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0.48</td>\n",
              "      <td>0.47</td>\n",
              "      <td>0.04</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>19728</td>\n",
              "      <td>functional</td>\n",
              "      <td>95</td>\n",
              "      <td>7</td>\n",
              "      <td>-1.83</td>\n",
              "      <td>4</td>\n",
              "      <td>9</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>6</td>\n",
              "      <td>3</td>\n",
              "      <td>5</td>\n",
              "      <td>1</td>\n",
              "      <td>31.13</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>1328</td>\n",
              "      <td>532</td>\n",
              "      <td>15</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>22</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>0.62</td>\n",
              "      <td>0.29</td>\n",
              "      <td>0.09</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-ed4c2dcf-24f9-4a4d-9bd2-190be74fbc56')\"\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-ed4c2dcf-24f9-4a4d-9bd2-190be74fbc56 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-ed4c2dcf-24f9-4a4d-9bd2-190be74fbc56');\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": [
              "      id  ... Functional need repair pumps (%)\n",
              "0  69572  ...                             0.09\n",
              "1   8776  ...                             0.09\n",
              "2  34310  ...                             0.04\n",
              "3  67743  ...                             0.04\n",
              "4  19728  ...                             0.09\n",
              "\n",
              "[5 rows x 29 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 42
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "####**Baseline model performance**"
      ],
      "metadata": {
        "id": "yZc1tQvjBMV9"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#define the models\n",
        "models=[LogisticRegression(max_iter = 1000000,  random_state=42),\n",
        "        KNeighborsClassifier(),\n",
        "        DecisionTreeClassifier( random_state=42),\n",
        "       RandomForestClassifier( random_state=42),\n",
        "       XGBClassifier( random_state=42),\n",
        "        BaggingClassifier( random_state=42), \n",
        "        AdaBoostClassifier( random_state=42),\n",
        "        GradientBoostingClassifier( random_state=42),\n",
        "       CatBoostClassifier( random_state=42)]\n",
        "\n",
        "#create a list of model names\n",
        "names =['Logistic Regression', 'KNN Classifier', 'Decision Tree Classifier', 'Random Forest Classifier',  'XGB Classifier', 'Bagging Classifier', 'AdaBoost Classifier', 'Gradient Boosting Classifier', 'Catboost Classifier']\n",
        "\n",
        "#create empty lists for the accuracy and standard deviation\n",
        "accuracy=[]\n",
        "std = []\n",
        "\n",
        "#peform 10-fold cross validaton on each model and append the results to the list\n",
        "for model in models:\n",
        "  cv = KFold(n_splits=10, shuffle = True,  random_state=1)\n",
        "  n_scores = cross_val_score(model, X, y, scoring='accuracy', cv=cv, n_jobs=-1, error_score='raise')\n",
        "  mean_score = np.mean(n_scores)\n",
        "  std_score = np.std(n_scores)\n",
        "  accuracy.append(mean_score)\n",
        "  std.append(std_score)\n",
        "\n",
        "#create a dataframe of the model performances\n",
        "d = {'Model':names,'Accuracy':accuracy, 'Std':std}   \n",
        "score_df = pd.DataFrame(d)\n",
        "score_df"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 331
        },
        "id": "RiFJYa-PAzd9",
        "outputId": "17171aae-484f-4646-d365-1e32f4c302db"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "\n",
              "  <div id=\"df-3d31554a-f748-4b44-a407-0bbd89f111f7\">\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>Model</th>\n",
              "      <th>Accuracy</th>\n",
              "      <th>Std</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Logistic Regression</td>\n",
              "      <td>0.666510</td>\n",
              "      <td>0.006001</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>KNN Classifier</td>\n",
              "      <td>0.701140</td>\n",
              "      <td>0.007301</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Decision Tree Classifier</td>\n",
              "      <td>0.756158</td>\n",
              "      <td>0.004169</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>Random Forest Classifier</td>\n",
              "      <td>0.804121</td>\n",
              "      <td>0.004219</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>XGB Classifier</td>\n",
              "      <td>0.742740</td>\n",
              "      <td>0.005622</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>Bagging Classifier</td>\n",
              "      <td>0.793599</td>\n",
              "      <td>0.003323</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>AdaBoost Classifier</td>\n",
              "      <td>0.723985</td>\n",
              "      <td>0.005083</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>Gradient Boosting Classifier</td>\n",
              "      <td>0.751174</td>\n",
              "      <td>0.005385</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>Catboost Classifier</td>\n",
              "      <td>0.796428</td>\n",
              "      <td>0.003599</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-3d31554a-f748-4b44-a407-0bbd89f111f7')\"\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-3d31554a-f748-4b44-a407-0bbd89f111f7 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-3d31554a-f748-4b44-a407-0bbd89f111f7');\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": [
              "                          Model  Accuracy       Std\n",
              "0           Logistic Regression  0.666510  0.006001\n",
              "1                KNN Classifier  0.701140  0.007301\n",
              "2      Decision Tree Classifier  0.756158  0.004169\n",
              "3      Random Forest Classifier  0.804121  0.004219\n",
              "4                XGB Classifier  0.742740  0.005622\n",
              "5            Bagging Classifier  0.793599  0.003323\n",
              "6           AdaBoost Classifier  0.723985  0.005083\n",
              "7  Gradient Boosting Classifier  0.751174  0.005385\n",
              "8           Catboost Classifier  0.796428  0.003599"
            ]
          },
          "metadata": {},
          "execution_count": 43
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### **2.3 Experiment 2 - Specialized models**"
      ],
      "metadata": {
        "id": "gSI4DmYlB3Gx"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Is it possible to create three sepperate models that are each specialized in predicting one of the three classes and combining them together to create a final prediction? "
      ],
      "metadata": {
        "id": "Rc66xuhaB9IM"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#create a copy of the dataset\n",
        "train_set_sp = train_set.copy()\n",
        "\n",
        "#define X and y for training the model\n",
        "X= train_set_sp.drop(['id', 'status_group'], axis=1)\n",
        "y = train_set_sp['status_group']"
      ],
      "metadata": {
        "id": "bqk1T3tADEna"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#step 1: create three different y variables\n",
        "\n",
        "#functional y:  functional =1, other classes are 0\n",
        "y_functional = y.copy()\n",
        "y_functional = y_functional .replace(['non functional'], 0)\n",
        "y_functional  = y_functional.replace('functional needs repair', 0)\n",
        "y_functional  = y_functional.replace('functional', 1)\n",
        "\n",
        "#non-fuctional y: non-functional =1, other classes are 0\n",
        "y_non_functional = y.copy()\n",
        "y_non_functional = y_non_functional.replace(['non functional'], 1)\n",
        "y_non_functional  = y_non_functional.replace('functional needs repair', 0)\n",
        "y_non_functional = y_non_functional.replace('functional', 0)\n",
        "\n",
        "#repair y: repair = 1, other classes are 0\n",
        "y_repair = y.copy()\n",
        "y_repair = y_repair.replace(['non functional'], 0)\n",
        "y_repair  = y_repair.replace('functional needs repair', 1)\n",
        "y_repair = y_repair.replace('functional', 0)"
      ],
      "metadata": {
        "id": "SirvbdMcB2Ik"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Note: the best models provided below were created using a slightly different subset of features and a different encoding strategy. The hyperparamter settings may not be the optimal parameters for the current dataset. "
      ],
      "metadata": {
        "id": "R8Aw-qpgEtZy"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### **Functional model**"
      ],
      "metadata": {
        "id": "1YE2qSyeDrIA"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#running this script on Google Colab takes too long\n",
        "\n",
        "# Create training and test sets: test size 0.2\n",
        "X_train, X_test, y_train, y_test= train_test_split(X, y_functional, test_size = 0.2, random_state=42)\n",
        "\n",
        "\n",
        "param_grid = { 'n_estimators': [ 180, 200, 220, ],\n",
        "             'max_features':['log2', ],\n",
        "             'max_depth' : [70, 80, 90, 100],\n",
        "             'min_samples_split':[ 7, 8,9]}\n",
        "\n",
        "\n",
        "#Create a based model\n",
        "rf =     RandomForestClassifier( random_state=42, warm_start = True)  \n",
        "\n",
        "grid_search_rf = GridSearchCV(estimator = rf, param_grid = param_grid, cv = 5,  verbose = 2,  scoring = 'accuracy')\n",
        "\n",
        "#fitmodel\n",
        "grid_search_rf .fit(X_train, y_train)\n",
        "\n",
        "#print best parameters \n",
        "print('Best Score:', grid_search_rf .best_score_)\n",
        "print('Parameters:', grid_search_rf .best_params_)\n",
        "print('Best Model:', grid_search_rf .best_estimator_)\n",
        "\n"
      ],
      "metadata": {
        "id": "qEN_Z9TJDp7K"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#score model\n",
        "best_model = RandomForestClassifier(max_depth=70, max_features='log2', min_samples_split=7, n_estimators=180, random_state=42, warm_start=True)\n",
        "\n",
        "X_train, X_test, y_train, y_test= train_test_split(X, y_functional, test_size = 0.2, random_state=42 , stratify=y_functional)\n",
        "\n",
        "#fit the best model\n",
        "best_model.fit(X_train, y_train)\n",
        "y_pred = best_model.predict(X_test)\n",
        "y_pred_train = best_model.predict(X_train)\n",
        "\n",
        "\n",
        "#print best model scores on test data\n",
        "print(\"Accuracy score train: {}\".format(accuracy_score(y_train, y_pred_train)))\n",
        "print(\"Accuracy score test: {}\".format(accuracy_score(y_test, y_pred)))\n",
        "print()\n",
        "\n",
        "#print(confusion_matrix(y_test, y_pred))\n",
        "print(classification_report(y_test, y_pred))\n",
        "\n",
        "feature_importances(best_model)\n",
        "\n",
        "confusion_matrix_plotter(y_test, y_pred)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "LLJ1CK00VBw6",
        "outputId": "9962b4ad-681a-49ae-8f28-44733a04bb19"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Accuracy score train: 0.9417496159430965\n",
            "Accuracy score test: 0.8283670033670034\n",
            "\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.84      0.77      0.80      5428\n",
            "           1       0.82      0.87      0.85      6452\n",
            "\n",
            "    accuracy                           0.83     11880\n",
            "   macro avg       0.83      0.82      0.83     11880\n",
            "weighted avg       0.83      0.83      0.83     11880\n",
            "\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x720 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "qTH0x8Kvcalj"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### **Non-functional model**"
      ],
      "metadata": {
        "id": "NeCeUueWEU2F"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Create training and test sets: test size 0.2\n",
        "X_train, X_test, y_train, y_test= train_test_split(X, y_non_functional, test_size = 0.2, random_state=42)\n",
        "\n",
        "param_grid = { 'n_estimators': [ 180, 200, 220, ],\n",
        "             'max_features':['sqrt' ],\n",
        "             'max_depth' : [30,40,50],\n",
        "             'min_samples_split':[ 4,5,6]}\n",
        "\n",
        "#Create a based model\n",
        "rf =     RandomForestClassifier( random_state=42, warm_start = True)  \n",
        "\n",
        "grid_search_rf = GridSearchCV(estimator = rf, param_grid = param_grid, cv = 5,  verbose = 2,  scoring = 'accuracy')\n",
        "\n",
        "#fitmodel\n",
        "grid_search_rf .fit(X_train, y_train)\n",
        "\n",
        "#print best parameters \n",
        "print('Best Score:', grid_search_rf .best_score_)\n",
        "print('Parameters:', grid_search_rf .best_params_)\n",
        "print('Best Model:', grid_search_rf .best_estimator_)\n"
      ],
      "metadata": {
        "id": "rmevHBJIEWtp"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#score model\n",
        "best_model =  RandomForestClassifier(max_depth=30, max_features='sqrt', min_samples_split=5,   n_estimators=180, random_state=42, warm_start=True)\n",
        "\n",
        "X_train, X_test, y_train, y_test= train_test_split(X, y_non_functional, test_size = 0.2, random_state=42 , stratify=y_non_functional)\n",
        "\n",
        "#fit the best model\n",
        "best_model.fit(X_train, y_train)\n",
        "y_pred = best_model.predict(X_test)\n",
        "y_pred_train = best_model.predict(X_train)\n",
        "\n",
        "\n",
        "#print best model scores on test data\n",
        "print(\"Accuracy score train: {}\".format(accuracy_score(y_train, y_pred_train)))\n",
        "print(\"Accuracy score test: {}\".format(accuracy_score(y_test, y_pred)))\n",
        "print()\n",
        "\n",
        "#print(confusion_matrix(y_test, y_pred))\n",
        "print(classification_report(y_test, y_pred))\n",
        "\n",
        "feature_importances(best_model)\n",
        "\n",
        "confusion_matrix_plotter(y_test, y_pred)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "BQWslrAmEYI5",
        "outputId": "99219456-7538-4269-d67a-c7873cbb7279"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Accuracy score train: 0.9612365580083756\n",
            "Accuracy score test: 0.8640572390572391\n",
            "\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.86      0.92      0.89      7315\n",
            "           1       0.86      0.77      0.81      4565\n",
            "\n",
            "    accuracy                           0.86     11880\n",
            "   macro avg       0.86      0.85      0.85     11880\n",
            "weighted avg       0.86      0.86      0.86     11880\n",
            "\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAArIAAAJcCAYAAAAFN0dyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzde5ieVX3v//fHiOEcqlB2ZCtTMYocI4yoKDQcan8Vq1ixqJSTVjZoRfHCvdN6KFhtY3EXPFQxWsQDogWVUqgCihyMIkwgB4KAuxKrYFVUwklQwvf3x7PSPoyTw0xmMrkz79d1zTX3s+51r/W9n/zzyXrW/UyqCkmSJKlrHjfZBUiSJEljYZCVJElSJxlkJUmS1EkGWUmSJHWSQVaSJEmdZJCVJElSJxlkJUmS1EkGWUmaYEmWJ/lVkvv7fp48DmMeOl41rsN8pyX57Iaab02SHJfkm5Ndh6TJZ5CVpA3jj6tq676fuyazmCSPn8z5x6qrdUuaGAZZSZokSWYk+ackP05yZ5L3JJnWzu2S5MokP09yd5LzkmzXzn0GeCrwr211938nmZPkR8PG/69V27aiemGSzya5FzhuTfOvQ+2V5A1JvpfkviR/02r+VpJ7k/xzkie0vnOS/CjJX7V7WZ7kqGHvw6eT/CzJD5K8I8nj2rnjkixIcmaSnwNfAM4Gnt/u/Z7W77AkN7W5f5jktL7xB1q9xyb5j1bD2/vOT2u1/Xu7l4VJntLO7ZrkiiS/SHJbkj8d5T+zpAlkkJWkyXMu8AjwdODZwIuAP2/nAvwd8GTgWcBTgNMAqupo4D/471Xev1/H+V4GXAhsB5y3lvnXxR8C+wLPA/43MB/4s1brHsCr+/r+D2B7YCfgWGB+kme2cx8CZgBPA34fOAY4vu/a5wLfB3Zs458IfLvd+3atzwPtuu2Aw4CTkhw+rN4XAs8EDgHeleRZrf2trdYXA9sCrwUeTLIVcAXwOeB3gVcBH0my2yjeI0kTyCArSRvGRUnuaT8XJdmRXnB6S1U9UFU/Bc6kF5aoqv9XVVdU1cNV9TPgH+iFvPXx7aq6qKoepRfYVjv/Ovr7qrq3qpYBNwOXV9X3q2oF8BV64bjfO9v9XA1cCvxpWwF+FfCXVXVfVS0H/i9wdN91d1XVh6rqkar61UiFVNVVVbW0qh6tqiXA+fz2+3V6Vf2qqhYDi4G9W/ufA++oqtuqZ3FV/Rx4CbC8qj7Z5r4J+CLwylG8R5ImkHuNJGnDOLyqvrbqRZL9gM2AHydZ1fw44Ift/I7AB4ADgG3auV+uZw0/7DveeU3zr6Of9B3/aoTX/6Pv9S+r6oG+1z+gt9q8favjB8PO7bSaukeU5LnAPHorwU8ApgMXDOv2n33HDwJbt+OnAP8+wrA7A89dtX2heTzwmbXVI2nDcEVWkibHD4GHge2rarv2s21V7d7O/y1QwJ5VtS29j9TTd30NG+8BYMtVL9pK5w7D+vRfs7b5x9vvtI/qV3kqcBdwN/AbeqGx/9ydq6l7pNfQ+/j/YuApVTWD3j7ajNBvJD8EdllN+9V97892bTvDSes4rqQJZpCVpElQVT8GLgf+b5JtkzyuPSy16uPwbYD7gRVJdgLeNmyIn9DbU7rK7cDm7aGnzYB30FuVHOv8E+H0JE9IcgC9j+0vqKqVwD8D702yTZKd6e1ZXdNXff0E+J+rHiZrtgF+UVUPtdXu14yirk8Af5NkVnr2SvIk4BLgGUmOTrJZ+3lO395aSZPMICtJk+cYeh+D30Jv28CFwMx27nRgH2AFvf2kXxp27d8B72h7bk9t+1LfQC+U3UlvhfZHrNma5h9v/9nmuIveg2YnVtWt7dyb6NX7feCb9FZXz1nDWFcCy4D/THJ3a3sD8O4k9wHvoheO19U/tP6XA/cC/wRsUVX30XsA7lWt7v8E3sca/oMgacNK1Uif0EiSND6SzAE+W1X/c7JrkbRpcUVWkiRJnWSQlSRJUie5tUCSJEmd5IqsJEmSOsk/iDAFbb/99jUwMDDZZUiSJK3VwoUL766q4d+LDRhkp6SBgQGGhoYmuwxJkqS1SvKD1Z1za4EkSZI6ySArSZKkTjLISpIkqZMMspIkSeokg6wkSZI6ySArSZKkTjLISpIkqZMMspIkSeokg6wkSZI6ySArSZKkTjLISpIkqZMMspIkSeokg6wkSZI6ySArSZKkTjLISpIkqZMMspIkSeokg6wkSZI6ySArSZKkTjLISpIkqZMMspIkSeokg6wkSZI6ySArSZKkTjLISpIkqZMeP9kFaMNbeucKBuZeOtllSJKkjlo+77DJLgFwRVaSJEkdZZCVJElSJxlkJUmS1EkGWUmSJHWSQVaSJEmdZJDdSCV5S5It+17/W5Lt2s8bJrM2SZKkjYFBduP1FuC/gmxVvbiq7gG2AwyykiRpyjPIjlGStye5Pck3k5yf5NQkVyUZbOe3T7K8HQ8kuTbJje1n/9Y+p11zYZJbk5yXnpOBJwPfSPKN1nd5ku2BecAuSRYlOSPJp5Mc3lfXeUletoHfDkmSpA3OP4gwBkn2BV4FzKb3Ht4ILFzDJT8F/qCqHkoyCzgfGGznng3sDtwFLABeUFUfTPJW4KCqunvYWHOBPapqdqvl94FTgIuSzAD2B44doeYTgBMApm27w+hvWpIkaSPjiuzYHAB8uaoerKp7gYvX0n8z4ONJlgIXALv1nbu+qn5UVY8Ci4CB0RRSVVcDs5LsALwa+GJVPTJCv/lVNVhVg9O2nDGaKSRJkjZKrsiOr0f47/8cbN7XfgrwE2Dvdv6hvnMP9x2vZGz/Jp8G/ozeKvHxY7hekiSpc1yRHZtrgMOTbJFkG+CPW/tyYN92fERf/xnAj9uq69HAtHWY4z5gm3VsP5few2FU1S3rMLYkSVLnGWTHoKpuBL4ALAa+AtzQTr0fOCnJTcD2fZd8BDg2yWJgV+CBdZhmPvDVVQ979c39c2BBkpuTnNHafgJ8F/jk2O9KkiSpW1JVk11D5yU5Dbi/qt4/SfNvCSwF9qmqFWvrP33mrJp57FkTX5gkSdokLZ932AabK8nCqhoc6Zwrsh2X5FB6q7EfWpcQK0mStKnwYa9xUFWnTeLcXwN2nqz5JUmSJotBdgrac6cZDG3AjwQkSZImglsLJEmS1EkGWUmSJHWSQVaSJEmdZJCVJElSJ/mw1xS09M4VDMy9dLLLkDTFbMjvnZQ0NbgiK0mSpE4yyEqSJKmTDLKSJEnqJIOsJEmSOskgu56S3L+W89sleUPf6ycnubAdz07y4jHMeVqSU0dfrSRJ0qbDIDvxtgP+K8hW1V1VdUR7ORsYdZCVJEmSQXbcJNk6ydeT3JhkaZKXtVPzgF2SLEpyRpKBJDcneQLwbuDIdu7I4Sutrd9AO357ktuTfBN4Zl+fXZJ8NcnCJNcm2XWD3bQkSdIk8ntkx89DwMur6t4k2wPXJbkYmAvsUVWzAVYF06r6dZJ3AYNV9Rft3GkjDZxkX+BV9FZwHw/cCCxsp+cDJ1bV95I8F/gIcPAIY5wAnAAwbdsdxuN+JUmSJpVBdvwE+NskBwKPAjsBO47T2AcAX66qBwFaQCbJ1sD+wAVJVvWdPtIAVTWfXuhl+sxZNU51SZIkTRqD7Pg5CtgB2LeqfpNkObD5KMd4hMdu91jb9Y8D7lm12itJkjSVuEd2/MwAftpC7EHAzq39PmCb1Vwz/NxyYB+AJPsAv9farwEOT7JFkm2APwaoqnuBO5K8sl2TJHuP3y1JkiRtvAyy4+c8YDDJUuAY4FaAqvo5sKA9uHXGsGu+Aey26mEv4IvAE5MsA/4CuL2NcSPwBWAx8BXghr4xjgJel2QxsAx4GZIkSVNAqtwuOdVMnzmrZh571mSXIWmKWT7vsMkuQVIHJVlYVYMjnXNFVpIkSZ1kkJUkSVInGWQlSZLUSX791hS0504zGHKvmiRJ6jhXZCVJktRJBllJkiR1kkFWkiRJneQe2Slo6Z0rGJh76WSXIWkK8LtjJU0kV2QlSZLUSQZZSZIkdZJBVpIkSZ1kkJUkSVInGWQlSZLUSQbZtUhy/wSM+dIkc9vx4Ul2G8MYVyUZHO/aJEmSusIgOwmq6uKqmtdeHg6MOshKkiRNdQbZdZSeM5LcnGRpkiNb+5y2OnphkluTnJck7dyLW9vCJB9McklrPy7Jh5PsD7wUOCPJoiS79K+0Jtk+yfJ2vEWSzyf5bpIvA1v01faiJN9OcmOSC5JsvWHfHUmSpA3PP4iw7v4EmA3sDWwP3JDkmnbu2cDuwF3AAuAFSYaAjwEHVtUdSc4fPmBVfSvJxcAlVXUhQMvAIzkJeLCqnpVkL+DG1n974B3AoVX1QJL/A7wVeHf/xUlOAE4AmLbtDmN8CyRJkjYersiuuxcC51fVyqr6CXA18Jx27vqq+lFVPQosAgaAXYHvV9Udrc9vBdlROhD4LEBVLQGWtPbn0duasCDJIuBYYOfhF1fV/KoarKrBaVvOWM9SJEmSJp8rsuPj4b7jlazf+/oI//0fjM3XoX+AK6rq1esxpyRJUue4IrvurgWOTDItyQ70VkivX0P/24CnJRlor49cTb/7gG36Xi8H9m3HR/S1XwO8BiDJHsBerf06elsZnt7ObZXkGetwP5IkSZ1mkF13X6b3cf5i4Ergf1fVf66uc1X9CngD8NUkC+kF1hUjdP088LYkNyXZBXg/cFKSm+jtxV3lo8DWSb5Lb//rwjbPz4DjgPOTLAG+TW9bgyRJ0iYtVTXZNWyykmxdVfe3bzH4R+B7VXXmZNc1feasmnnsWZNdhqQpYPm8wya7BEkdl2RhVY343fmuyE6s17cHsJYBM+h9i4EkSZLGgQ97TaC2+jrpK7CSJEmbIldkJUmS1EmuyE5Be+40gyH3rUmSpI5zRVaSJEmdZJCVJElSJxlkJUmS1EnukZ2Clt65goG5l052GZIav2tVksbGFVlJkiR1kkFWkiRJnWSQlSRJUicZZCVJktRJBllJkiR1kkFWkiRJnWSQlSRJUicZZDdCSS5KsjDJsiQntLbXJbk9yfVJPp7kw619hyRfTHJD+3nB5FYvSZK0YfgHETZOr62qXyTZArghyaXAO4F9gPuAK4HFre8HgDOr6ptJngpcBjxr+IAtEJ8AMG3bHTbALUiSJE0sg+zG6eQkL2/HTwGOBq6uql8AJLkAeEY7fyiwW5JV126bZOuqur9/wKqaD8wHmD5zVk1w/ZIkSRPOILuRSTKHXjh9flU9mOQq4FZGWGVtHgc8r6oe2jAVSpIkbRzcI7vxmQH8soXYXYHnAVsBv5/kd5I8HnhFX//LgTetepFk9gatVpIkaZIYZDc+XwUen+S7wDzgOuBO4G+B64EFwHJgRet/MjCYZEmSW4ATN3jFkiRJk8CtBRuZqnoY+KPh7UmGqmp+W5H9MnBR6383cOSGrVKSJGnyuSLbHaclWQTcDNxBC7KSJElTlSuyHVFVp052DZIkSRsTg+wUtOdOMxiad9hklyFJkrRe3FogSZKkTjLISpIkqZMMspIkSeokg6wkSZI6yYe9pqCld65gYO6lk12G1HnLfWhSkiaVK7KSJEnqJIOsJEmSOskgK0mSpE4yyEqSJKmTDLKSJEnqpE0qyCZ5S5ItN8A8L00ydy19BpK8Zi19Zid58fhWJ0mSNDVsUkEWeAswqiCbZNpoJ6mqi6tq3lq6DQBrDLLAbMAgK0mSNAYbZZBN8rYkJ7fjM5Nc2Y4PTnJeko8mGUqyLMnp7dzJwJOBbyT5Rmt7UZJvJ7kxyQVJtm7ty5O8L8mNwCuTXJXkA0kWJbk5yX6t3xOTXJRkSZLrkuzV2o9L8uF2fG6SDyb5VpLvJzmi3cY84IA25ikj3OMTgHcDR7Y+Ryb5XpId2vnHJfl/SXZoc5zd7vn2JC9pfaYlOSPJDa3G/7WG9/SEdv3QygdXrOe/kCRJ0uTbKIMscC1wQDseBLZOsllruwZ4e1UNAnsBv59kr6r6IHAXcFBVHZRke+AdwKFVtQ8wBLy1b46fV9U+VfX59nrLqpoNvAE4p7WdDtxUVXsBfwV8ejX1zgReCLyEXoAFmAtcW1Wzq+rM4RdU1a+BdwFfaH2+AHwWOKp1ORRYXFU/a68HgP2Aw4Czk2wOvA5YUVXPAZ4DvD7J741UYFXNr6rBqhqctuWM1dyGJElSd2ysQXYhsG+SbYGHgW/TC7QH0Au5f9pWU28Cdgd2G2GM57X2BUkWAccCO/ed/8Kw/ucDVNU1wLZJtqMXTj/T2q8EntRqGu6iqnq0qm4BdhzD/a5yDnBMO34t8Mm+c//c5vge8H1gV+BFwDHt/r4DPAmYtR7zS5IkdcZG+Sdqq+o3Se4AjgO+BSwBDgKeDvwKOBV4TlX9Msm5wOYjDBPgiqp69WqmeWD4tGt5vSYPD5t3TKrqh0l+kuRgequvR/WfHqG+AG+qqsvGOqckSVJXbawrstBbeT2V3laCa4ET6a3AbksvhK5IsiPwR33X3Ads046vA16Q5OkASbZK8ow1zHdk6/dCeh/Xr2jzHtXa5wB3V9W961h/fy2j6fMJelsMLqiqlX3tr2z7ZncBngbcBlwGnNS2XZDkGUm2Wsf6JEmSOm1jD7IzgW9X1U+Ah+jtOV1ML9DeCnwOWNB3zXzgq0m+0faWHgecn2QJve0Ju65hvoeS3AScTW/vKcBp9LY4LKG39/XYUdS/BFiZZPFID3s13wB2W/WwV2u7GNiax24rAPgP4HrgK8CJVfUQvdB7C3BjkpuBj7GRrrJLkiSNt1SN5hP0TVOSq4BTq2poI6hlEDizqg7oazsXuKSqLhyPOabPnFUzjz1rPIaSprTl8w6b7BIkaZOXZGF7yP+3uHq3EWl/ZOEkHrs3VpIkSSMwyAJVNWcix0/yh8D7hjXfUVUvH1bHPP7767v624+buOokSZK6ya0FU9Dg4GANDU36LgpJkqS1WtPWgo35YS9JkiRptQyykiRJ6iSDrCRJkjrJh72moKV3rmBg7qWTXYa0UfOrtSRp4+eKrCRJkjrJICtJkqROMshKkiSpkwyykiRJ6iSDrCRJkjrJICtJkqROMshOkCRzklwyymveneTQtfQ5LcmpI7Rvl+QNo61TkiSpqwyyG5GqeldVfW2Ml28HGGQlSdKUYZAFkrwzyW1Jvpnk/CSnJrkqyQeSLEpyc5L9Wt/fb22LktyUZJs1DL11kguT3JrkvCRpY+yb5OokC5NclmRmaz83yRHt+MXtuoVJPjhsdXe3Vt/3k5zc2uYBu7S6zhjhHk9IMpRkaOWDK8bjbZMkSZpUU/4veyV5DvAKYG9gM+BGYGE7vWVVzU5yIHAOsAdwKvDGqlqQZGvgoTUM/2xgd+AuYAHwgiTfAT4EvKyqfpbkSOC9wGv7atoc+BhwYFXdkeT8YePuChwEbAPcluSjwFxgj6qaPVIhVTUfmA8wfeasWoe3RpIkaaM25YMs8ALgX6rqIeChJP/ad+58gKq6Jsm2SbajF0j/Icl5wJeq6kdrGPv6VeeTLAIGgHvoBeIr2gLtNODHw67bFfh+Vd3RV8cJfecvraqHgYeT/BTYcbQ3LUmS1HUG2TUbvnJZVTUvyaXAi4EFSf6wqm5dzfUP9x2vpPd+B1hWVc9fj7pGGleSJGlKcY9sb4X1j5Ns3rYKvKTv3JEASV4IrKiqFUl2qaqlVfU+4AZ6q6ejcRuwQ5Lnt7E3S7L7CH2elmSgv461uI/eVgNJkqQpYcqv5FXVDUkuBpYAPwGWAquehnooyU309s6u2sP6liQHAY8Cy4CvjHK+X7cHuj6YZAa9f4Oz2lir+vyqfZXWV5M8QC8wr23cnydZkORm4CtV9bbR1CVJktQ1qfK5nyRbV9X9SbYErqG3H/UfgFOramiSawrwj8D3qurM8Rh7+sxZNfPYs8ZjKGmTtXzeYZNdgiQJSLKwqgZHOufWgp757WGsG4EvVtWNk10Q8PpW0zJgBr1vMZAkSVIz5bcWAFTVa0Zom7Mu1ybZE/jMsOaHq+q561nTmcC4rMBKkiRtitxaMAUNDg7W0NCk7JiQJEkaFbcWSJIkaZNjkJUkSVInGWQlSZLUST7sNQUtvXMFA3Mvnewy1GF+NZUkaWPgiqwkSZI6ySArSZKkTjLISpIkqZMMspIkSeokg6wkSZI6ySC7gSQZSHLzOvR5Td/rwSQfnPjqJEmSuscgu3EZAP4ryFbVUFWdPHnlSJIkbbwMsk1bDb01yXlJvpvkwiRbJjkkyU1JliY5J8n01n95kr9v7dcneXprPzfJEX3j3r+aua5NcmP72b+dmgcckGRRklOSzElySbvmiUkuSrIkyXVJ9mrtp7W6rkry/SQGX0mSNCUYZB/rmcBHqupZwL3AW4FzgSOrak96f0DipL7+K1r7h4GzRjHPT4E/qKp9gCOBVdsH5gLXVtXsqjpz2DWnAzdV1V7AXwGf7ju3K/CHwH7AXyfZbPiESU5IMpRkaOWDK0ZRqiRJ0sbJIPtYP6yqBe34s8AhwB1VdXtr+xRwYF//8/t+P38U82wGfDzJUuACYLd1uOaFwGcAqupK4ElJtm3nLq2qh6vqbnohecfhF1fV/KoarKrBaVvOGEWpkiRJGyf/RO1j1bDX9wBPWsf+q44fof0HIcnjgCeMcN0pwE+AvVvfh8ZSbJ+H+45X4r+rJEmaAlyRfaynJlm1svoaYAgYWLX/FTgauLqv/5F9v7/djpcD+7bjl9JbfR1uBvDjqnq0jTmttd8HbLOa2q4FjgJIMge4u6ruXae7kiRJ2gS5cvdYtwFvTHIOcAtwMnAdcEGSxwM3AGf39f+dJEvorYi+urV9HPiXJIuBrwIPjDDPR4AvJjlmWJ8lwMp27bnATX3XnAac0+Z7EDh2/W5VkiSp21I1/NP0qSnJAHBJVe2xjv2XA4NtX2qnTJ85q2YeO5pn06THWj7vsMkuQZI0RSRZWFWDI51za4EkSZI6ya0FTVUtB9ZpNbb1H5iwYiRJkrRWBtkpaM+dZjDkR8OSJKnj3FogSZKkTjLISpIkqZMMspIkSeokg6wkSZI6yYe9pqCld65gYO6lk12GJpDf8ypJmgpckZUkSVInGWQlSZLUSQZZSZIkdZJBVpIkSZ1kkJUkSVInTViQTTKQ5DXjON7hSXbre/3uJIeO4/hzkuw/XuONsYarkgxOZg2SJEldMZErsgPAiEE2yVi+9utw4L+CbFW9q6q+NrbSRjQHmNQgK0mSpHU36iCb5M+SXJ9kUZKPJXlukiVJNk+yVZJlSfYA5gEHtH6nJDkuycVJrgS+nmTrJF9PcmOSpUle1jfHMW3MxUk+01ZKXwqc0cbbJcm5SY5o/Q9JclMb55wk01v78iSn982x62ruaQA4ETiljX9AkjuSbNbOb7vqdVs1/UDrd3OS/Vqfrdrc17daXjbSXK3vtCTvb9cvSfKmEfp8NMlQez9P72ufl+SWdt37W9sr21iLk1yzmjlPaOMNrXxwxRr+hSVJkrphVCujSZ4FHAm8oKp+k+QjwDOBi4H3AFsAn62qm5PMBU6tqpe0a48D9gH2qqpftFXZl1fVvUm2B65LcjG9Vdd3APtX1d1Jntj6XwxcUlUXtvFW1bQ5cC5wSFXdnuTTwEnAWa3su6tqnyRvAE4F/nz4fVXV8iRnA/dX1apweBVwGHAR8CrgS+2eAbasqtlJDgTOAfYA3g5cWVWvTbIdcH2Sr1XVAyO8lSfQW7GeXVWPJHniCH3e3u57Gr3gvxdwJ/ByYNeqqjYPwLuAP6yqO/vaht/jfGA+wPSZs2qkPpIkSV0y2hXZQ4B9gRuSLGqvnwa8G/gDYBD4+zVcf0VV/aIdB/jbJEuArwE7ATsCBwMXVNXdAH39V+eZwB1VdXt7/SngwL7zX2q/F9ILj+vqE8Dx7fh44JN9585vtV0DbNvC44uAue19uQrYHHjqasY+FPhYVT3SxhnpHv80yY3ATcDu9AL+CuAh4J+S/AnwYOu7ADg3yeuBaaO4R0mSpM4a7V7VAJ+qqr98TGMyE9ga2IxegBtpFZJh7UcBOwD7tpXO5e3a8fZw+72SUdxvVS1oD6zNAaZV1c39p4d3p/fevKKqblufYgGS/B691ePnVNUvk5wLbN5Wb/ej9x+II4C/AA6uqhOTPJfeCvLCJPtW1c/Xtw5JkqSN2WhXZL8OHJHkdwGSPDHJzsDHgHcC5wHva33vA7ZZw1gzgJ+2EHsQsHNrvxJ4ZZInrZpjLePdBgwkeXp7fTRw9Sjva3Xjfxr4HI9djYXe9gqSvBBYUVUrgMuAN6XtPUjy7DXMdQXwv9r2iv57XGVbeqF/RZIdgT9q/bYGZlTVvwGnAHu39l2q6jtV9S7gZ8BT1vmuJUmSOmpUK7JVdUuSdwCXJ3kc8BvgX4DfVNXn2n7ObyU5GLgWWJlkMb09rL8cNtx5wL8mWQoMAbe2OZYleS9wdZKV9D5aPw74PPDxJCfTW41cVdNDSY4HLmjB8Abg7NHcV/OvwIXtIa03VdW1rcb30LYS9HkoyU30VqBf29r+ht6+3CXtvbkDeMlq5voE8IzW9zfAx4EP993T4jb+rcAP6W0dgF7Q/pe2LzjAW1v7GUlmtbavA4vHcP+SJEmdkiqf+1md9q0IL6uqo/varqL3ENvQpBW2nqbPnFUzjz1r7R3VWcvnHTbZJUiSNC6SLKyqEb9nfyzf5zolJPkQvY/0XzzZtUiSJOm3Tbkg27YhvHlY84KqemN/Q1X91ne7tvY5o5jrD/nvPcOr3FFVL1/XMSRJkjQytxZMQYODgzU01NmdEZIkaQpZ09aCifwTtZIkSdKEMchKkiSpkwyykiRJ6qQp97CXYOmdKxiYe+lkl6Fx5lduSZKmGldkJUmS1EkGWUmSJHWSQVaSJEmdZJCVJElSJxlkJUmS1EkG2QmS5OQk301y3nqOM5Dk5vGqS5IkaVPh129NnDcAh1bVjzbkpEkeX1WPbMg5JUmSJoMrshMgydnA04CvJFmR5NS+cze3VdaBtmL78STLklyeZIvWZ98ki5MsBt7Yd+20JGckuSHJkiT/q7XPSXJtkouBWzbs3UqSJE0Og+wEqKoTgbuAg4Az19B1FvCPVQ1UnBgAACAASURBVLU7cA/witb+SeBNVbX3sP6vA1ZU1XOA5wCvT/J77dw+wJur6hkjTZTkhCRDSYZWPrhiTPclSZK0MTHITq47qmpRO14IDCTZDtiuqq5p7Z/p6/8i4Jgki4DvAE+iF4YBrq+qO1Y3UVXNr6rBqhqctuWM8b0LSZKkSeAe2Yn3CI/9D8PmfccP9x2vBLZYy1iht1J72WMakznAA+tRoyRJUue4IjvxltP72J8k+wC/t6bOVXUPcE+SF7amo/pOXwaclGSzNt4zkmw17hVLkiR1gCuyE++L9LYDLKO3HeD2dbjmeOCcJAVc3tf+CWAAuDFJgJ8Bh49vuZIkSd2QqprsGrSBTZ85q2Yee9Zkl6FxtnzeYZNdgiRJ4y7JwqoaHOmcWwskSZLUSQZZSZIkdZJBVpIkSZ3kw15T0J47zWDI/ZSSJKnjXJGVJElSJxlkJUmS1EkGWUmSJHWSe2SnoKV3rmBg7qWTXYbGyO+LlSSpxxVZSZIkdZJBVpIkSZ1kkJUkSVInGWQlSZLUSQZZSZIkdZJBVpIkSZ00pYNskncnOXQCx1+eZPt2/K219P2rtZz/tyTbreH84Ul2G1ulkiRJ3bPJBNn0jOp+qupdVfW1iapp2Fz7r6XLiEF21X1V1Yur6p41XH84YJCVJElTRqeDbJKBJLcl+TRwM/DOJDckWZLk9L5+72z9vpnk/CSntvZzkxzRjg9JclOSpUnOSTK9tS9PcnqSG9u5XddQz5OSXJ5kWZJPAOk7d3/7PTPJNUkWJbk5yQFJ5gFbtLbzRrivpwxb3T2m3ePiJJ9Jsj/wUuCMNsYuI9R2QpKhJEMrH1yxvm+9JEnSpOt0kG1mAR8BTgF2AvYDZgP7JjkwyXOAVwB7A38EDA4fIMnmwLnAkVW1J72/eHZSX5e7q2of4KPAqWuo5a+Bb1bV7sCXgaeO0Oc1wGVVNbvVtKiq5gK/qqrZVXVU/31V1e5V9YO+WncH3gEcXFV7A2+uqm8BFwNva2P8+/BJq2p+VQ1W1eC0LWes4RYkSZK6YVMIsj+oquuAF7Wfm4AbgV3phcEXAP9SVQ9V1X3Av44wxjOBO6rq9vb6U8CBfee/1H4vBAbWUMuBwGcBqupS4Jcj9LkBOD7JacCeraY13ddwBwMXVNXdbZ5frKEeSZKkTdamEGQfaL8D/F1bkZxdVU+vqn8apzkebr9X0lutHbOquoZe4L0TODfJMavp+sBq2iVJksSmEWRXuQx4bZKtAZLslOR3gQXAHyfZvJ17yQjX3gYMJHl6e300cPUYariG3tYBkvwR8DvDOyTZGfhJVX0c+ASwTzv1mySbrcMcVwKvTPKkNt4TW/t9wDZjqFmSJKmTNpkgW1WXA58Dvp1kKXAhsE1V3UBv/+gS4CvAUmDFsGsfAo4HLmjXPgqcPYYyTgcOTLIM+BPgP0boMwdYnOQm4EjgA619PrAkyXlruc9lwHuBq5MsBv6hnfo88Lb2wNpvPewlSZK0qUlVTXYNEy7J1lV1f5It6a2anlBVN052XZNl+sxZNfPYsya7DI3R8nmHTXYJkiRtMEkWVtVvPawP67nfs0Pmtz8WsDnwqakcYiVJkjYVU2JFdrwlOR5487DmBVX1xsmoZ7QGBwdraGhossuQJElaK1dkx1lVfRL45GTXIUmSNJVtMg97SZIkaWoxyEqSJKmTDLKSJEnqJPfITkFL71zBwNxLJ7uMKc2v0JIkaf25IitJkqROMshKkiSpkwyykiRJ6iSDrCRJkjrJIDtGSb41xusOb38ud239Tktyajs+N8kRY5lPkiRpU2WQHaOq2n+Mlx4OrDXIro8kfhuFJEna5BlkxyjJ/e33nCRXJbkwya1JzkuSdm5ekluSLEny/iT7Ay8FzkiyKMkuSV6f5IYki5N8McmWa5l33yRXJ1mY5LIkM1v7VUnOSjIEvHmCb1+SJGnSuXI3Pp4N7A7cBSwAXpDku8DLgV2rqpJsV1X3JLkYuKSqLgRIck9Vfbwdvwd4HfChkSZJslk797Kq+lmSI4H3Aq9tXZ5QVYOrufYE4ASAadvuMC43LUmSNJkMsuPj+qr6EUCSRcAAcB3wEPBPSS4BLlnNtXu0ALsdsDVw2RrmeSawB3BFW/SdBvy47/wXVndhVc0H5gNMnzmr1n5LkiRJGzeD7Ph4uO94JfD4qnokyX7AIcARwF8AB49w7bnA4VW1OMlxwJw1zBNgWVU9fzXnHxhl3ZIkSZ3lHtkJkmRrYEZV/RtwCrB3O3UfsE1f122AH7dtA0etZdjbgB2SPL/NsVmS3ce3ckmSpG4wyE6cbYBLkiwBvgm8tbV/HnhbkpuS7AK8E/gOvb21t65pwKr6Nb3V3fclWQwsAsb67QmSJEmdliq3S04102fOqpnHnjXZZUxpy+cdNtklSJLUCUkWru5hdldkJUmS1EkGWUmSJHWSQVaSJEmd5NdvTUF77jSDIfdoSpKkjnNFVpIkSZ1kkJUkSVInGWQlSZLUSe6RnYKW3rmCgbmXTnYZU5rfIytJ0vpzRVaSJEmdZJCVJElSJxlkJUmS1EkGWUmSJHWSQVaSJEmdZJCVJElSJ21yQTbJQJLXjON4hyfZre/1u5McOo7jz0my/3iNJ0mSNFVsckEWGABGDLJJxvK9uYcD/xVkq+pdVfW1sZU2ojmAQVaSJGmUOhNkk/xZkuuTLErysSTPTbIkyeZJtkqyLMkewDzggNbvlCTHJbk4yZXA15NsneTrSW5MsjTJy/rmOKaNuTjJZ9pK6UuBM9p4uyQ5N8kRrf8hSW5q45yTZHprX57k9L45dl3NPQ0AJwKntPEPSHJHks3a+W1XvU5yVZIPtH43J9mv9dmqzX19q+Vlq5nrhCRDSYZWPrhinP5VJEmSJk8n/rJXkmcBRwIvqKrfJPkI8EzgYuA9wBbAZ6vq5iRzgVOr6iXt2uOAfYC9quoXbVX25VV1b5LtgeuSXExv1fUdwP5VdXeSJ7b+FwOXVNWFbbxVNW0OnAscUlW3J/k0cBJwViv77qraJ8kbgFOBPx9+X1W1PMnZwP1V9f427lXAYcBFwKuAL7V7BtiyqmYnORA4B9gDeDtwZVW9Nsl2wPVJvlZVDwybaz4wH2D6zFk1ln8HSZKkjUlXVmQPAfYFbkiyqL1+GvBu4A+AQeDv13D9FVX1i3Yc4G+TLAG+BuwE7AgcDFxQVXcD9PVfnWcCd1TV7e31p4AD+85/qf1eSG+7w7r6BHB8Oz4e+GTfufNbbdcA27bg+iJgbntfrgI2B546ivkkSZI6qRMrsvTC56eq6i8f05jMBLYGNqMX4B4Y4VqGtR8F7ADs21Y6l7drx9vD7fdKRvE+V9WC9sDaHGBaVd3cf3p4d3rvzSuq6rb1KVaSJKlrurIi+3XgiCS/C5DkiUl2Bj4GvBM4D3hf63sfsM0axpoB/LSF2IOAnVv7lcArkzxp1RxrGe82YCDJ09vro4Grx3BvI43/aeBzPHY1FnrbK0jyQmBFVa0ALgPelLb3IMmzx1CDJElS53QiyFbVLfT2r17etgRcARwL/KaqPkfvAa/nJDkYWAKsbA9snTLCcOcBg0mWAscAt7Y5lgHvBa5Oshj4h9b/88Db2oNUu/TV9BC9j/4vaGM9Cpw9htv7V+Dlqx726qvxd2hbCfo8lOSmNs/rWtvf0FuRXpJkWXstSZK0yUuVz/1sbNq3Irysqo7ua7uK3kNsQ+s7/vSZs2rmsWetvaMmzPJ5h012CZIkdUKShVU1ONK5ruyRnTKSfAj4I+DFk12LJEnSxswgu4EkOR5487DmBVX1xv6GqnrTSNdX1ZwJKk2SJKmT3FowBQ0ODtbQ0HrvUJAkSZpwa9pa0ImHvSRJkqThDLKSJEnqJIOsJEmSOsmHvaagpXeuYGDupZNdxpTjV25JkjS+XJGVJElSJxlkJUmS1EkGWUmSJHWSQVaSJEmdZJCdQEmOS/LktfR5S5ItN1RNkiRJmwqD7MQ6DlhjkAXeAhhkJUmSRqnTQTbJQJJbk5yX5LtJLkyyZZJ3Jbkhyc1J5qdnlyQ39l07a9XrJMuT/F2SRUmGkuyT5LIk/57kxL5r3tbGXZLk9L4avpvk40mWJbk8yRZJjgAGgfPauFuMUP/J9ILuN5J8I8lrk5zVd/71Sc5c3X22PvsmuTrJwlbzzIl6vyVJkjYmnQ6yzTOBj1TVs4B7gTcAH66q51TVHsAWwEuq6t+BFUlmt+uOBz7ZN85/VNVs4FrgXOAI4HnAqsD6ImAWsB8wG9g3yYHt2lnAP1bV7sA9wCuq6kJgCDiqqmZX1a+GF15VHwTuAg6qqoOAfwb+OMlmfTWes7r7bP0+BBxRVfu2vu8d/VsoSZLUPZtCkP1hVS1ox58FXggclOQ7SZYCBwO7t/OfAI5PMg04Evhc3zgXt99Lge9U1X1V9TPg4STbAS9qPzcBNwK70guwAHdU1aJ2vBAYGMuNVNX9wJXAS5LsCmxWVUvXcJ/PBPYArkiyCHgH8D9HGjvJCW21eWjlgyvGUp4kSdJGZVP4y141wuuPAINV9cMkpwGbt3NfBP6aXlhcWFU/77vu4fb70b7jVa8fDwT4u6r6WP9kSQaG9V9JbxV4rD4B/BVwK49dMR7pPgMsq6rnr23QqpoPzAeYPnPW8LEkSZI6Z1NYkX1qklVB7jXAN9vx3Um2prdFAICqegi4DPgojw2J6+Iy4LVtTJLslOR313LNfcA2o+lTVd8BnkLvXs7v6zfSfd4G7LCqPclmSXZHkiRpCtgUguxtwBuTfBf4HXoh9ePAzfTC5w3D+p9Hb5X18tFMUlWX09uK8O22ZeFC1h5SzwXOXt3DXs184KtJvtHX9s/Agqr6ZV/bb91nVf2aXlB/X5LFwCJg/9HclyRJUlelqrufMreP9S9pD3Wt6zWnAjOq6p0TVdf6SnIJcGZVfb29HmCU97km02fOqpnHnrX2jhpXy+cdNtklSJLUOUkWVtXgSOc2hT2y6yzJl4Fd6D0AttFpD5VdDyxeFWIlSZI0sk4H2apaTu+p/XXt//KJq2bNWoj+vWHN/6eqLlv1oqruAZ4x/NrR3qckSdJU0Okg2yWTGaIlSZI2RQbZKWjPnWYw5H5NSZLUcZvCtxZIkiRpCjLISpIkqZMMspIkSeokg6wkSZI6yYe9pqCld65gYO6lk11Gp/nHDSRJmnyuyEqSJKmTDLKSJEnqJIOsJEmSOskgK0mSpE4yyEqSJKmTDLIdkMRvl5AkSRrGIDsBkmyV5NIki5PcnOTIJIckuSnJ0iTnJJne+i5Psn07HkxyVTs+LclnkiwAPpNkxyRfbmMuTrJ/6/dnSa5PsijJx5JMm6z7liRJ2pAMshPj/wPuqqq9q2oP4KvAucCRVbUnve/vPWkdxtkNOLSqXg18ELi6qvYG9gGWJXkWcCTwgqqaDawEjhppoCQnJBlKMrTywRXreXuSJEmTzyA7MZYCf5DkfUkOAAaAO6rq9nb+U8CB6zDOxVX1q3Z8MPBRgKpaWVUrgEOAfYEbkixqr5820kBVNb+qBqtqcNqWM8Z6X5IkSRsN915OgKq6Pck+wIuB9wBXrqH7I/z3fyg2H3bugbVMFeBTVfWXYypUkiSpw1yRnQBJngw8WFWfBc4Ang8MJHl663I0cHU7Xk5vVRXgFWsY9uu07QhJpiWZ0dqOSPK7rf2JSXYez3uRJEnaWBlkJ8aewPXt4/6/Bt4BHA9ckGQp8Chwdut7OvCBJEP09riuzpuBg9r1C4HdquqWNvblSZYAVwAzJ+KGJEmSNjapqsmuQRvY9JmzauaxZ012GZ22fN5hk12CJElTQpKFVTU40jlXZCVJktRJBllJkiR1kkFWkiRJneTXb01Be+40gyH3eEqSpI5zRVaSJEmdZJCVJElSJxlkJUmS1EnukZ2Clt65goG5l052GZ3jd8dKkrRxcUVWkiRJnWSQlSRJUicZZCVJktRJBllJkiR1kkFWkiRJnWSQlSRJUidtMkE2yeFJdltLn+OSPHktfc5NcsQ413ZikmNGec2Tk1w4nnVIkiRtSjaZIAscDqwxyALHAWsMshOhqs6uqk+P8pq7qmpcA7UkSdKmZFyCbJKLkixMsizJCa3t/iRntLavJdkvyVVJvp/kpa3P5kk+mWRpkpuSHNTaj0vy4b7xL0kyp2/c9yZZnOS6JDsm2R94KXBGkkVJdhmhxiOAQeC81meLJPOS3JJkSZL393U/MMm3Wq2rDZNJ5iS5Osm/tL7zkhyV5Pp2T7u0fqclObUdn9w35+db2++3mha192GbJANJbu57P76U5KtJvpfk7/tqeF2S29ucH+9/34bVekKSoSRDKx9csfZ/VEmSpI3ceK3Ivraq9qUXFE9O8iRgK+DKqtoduA94D/AHwMuBd7fr3ghUVe0JvBr4VJLN1zLXVsB1VbU3cA3w+qr6FnAx8Laqml1V/z78oqq6EBgCjqqq2cCWrZbdq2qvVt8qM4EXAi8B5q2lnr2BE4FnAUcDz6iq/YBPAG8aof9c4NltzhNb26nAG1tdBwC/GuG62cCRwJ7AkUme0rZJvBN4HvACYNfVFVlV86tqsKoGp205Yy23JEmStPEbryB7cpLFwHXAU4BZwK+Br7bzS4Grq+o37Xigtb8Q+CxAVd0K/AB4xlrm+jVwSTte2DfWaK0AHgL+KcmfAA/2nbuoqh6tqlv4/9u793i7qvre+58vF4MQDCjqK1A0ikGKBAJs8K5gOXjBCqdiEbUF9UhVqrU++DTnWD3ipUI5PuJd0z4QrdcDolI4BVFAaeSSHSAXbl4grQd5+niNKJpK+J0/1ohdbnb23kl29tpz78/79VqvNdeYY4z5myNzJb+MNeZa8Ohx+llRVfdU1Qbge8BXW3n/efZbTW9W+BXA/a1sOfD/JHkjsEdV3T9Ku69X1fqq+jVwK/BY4Eh64/qTNrYXjBOrJEnSjLHNiWz7yP8Y4KltlvQmYBfgN1VVrdoDwAaAqnoA2Gmcbu8fEVv/LG1/vxsn0NeoWrJ4JHAhvZnXy/p2b+jbzjhd9dd9oO/15s7zOOAjwGHAiiQ7VdVZwH8BHgosTzLazGr/cbb6vCVJkmaKyZiRnQf8tKruawnYU7ag7TXAywGS7A88BrgDWAcsTrJDkn3pJZzjuRfYfaJ1kswF5lXV/wL+kt4Sge0qyQ7AvlV1FfBX9MZubpL9qmpNVZ0NrGCMJQIjrACenWTPJDsBL94ugUuSJE1DkzGrdxnw2iS30UtCr9uCth8FPpZkDb1Z2FOrakOS5cBd9D5Cvw24cQJ9fR74u/bx/ImjrZMFlgEfT/Ir4PnAV9qa3ABv3oK4t9aOwKeTzGvH/GBV/SzJu9qNbg8AtwD/RG+d7piq6u4kfwPcAPwEuJ3ekglJkqQZL//xKb26KMncqvpFm5H9EnBeVX1prDZz5i+s+aecOzUBziDrzjpu0CFIkjTrJFlZVUOj7ZtJ3yM7W70jyc3AWnqz2F8ecDySJElTYkbeMJTkI/S+jqrfB6rq/K3sbxHwDyOKN1TVk7emv8lUVWcMOgZJkqRBcGnBLDQ0NFTDw8ODDkOSJGlcLi2QJEnSjGMiK0mSpE4ykZUkSVInzcibvTS2NXevZ8GSSwcdxsD5dVqSJHWbM7KSJEnqJBNZSZIkdZKJrCRJkjrJRFaSJEmdZCIrSZKkTjKRlSRJUieZyE5QkncmOWbQcQAkOTXJhwcdhyRJ0iDNyu+RTRIgVfXARNtU1du3Y0iSJEnaQrNmRjbJgiR3JPkUsBZ4W5IVSVYnObOv3ttavX9O8rkkZ7TyZUlObNt/kOSmJGuSnJdkTitfl+TMJDe2fQeMEc/cJOe3equTvLiVn9zK1iY5u6/+K5N8O8kNwNP7yh+Z5IvtXFYkefoohyPJaUmGkwxvvG/9No2lJEnSdDBrEtlmIfBR4C+BfYAjgcXA4UmeleQI4MXAIcDzgaGRHSTZBVgGnFRVi+jNar+ur8qPquow4GPAGWPE8jZgfVUtqqqDgSuT7A2cDTynxXVEkhOSzAfOpJfAPgM4sK+fDwDvr6pNsf/9aAerqqVVNVRVQzvuOm+MsCRJkrphti0t+Jequi7J/wCOBW5q5XPpJbm7A1+pql8Dv07yj6P08UTgrqr6dnv9SeB04Nz2+qL2vBL4ozFiOQZ46aYXVfXTJM8Crq6qHwIk+QzwrFalv/wLwP59/RzYWy0BwMOSzK2qX4xxbEmSpM6bbYnsL9tzgPdW1Sf6dyZ50yQcY0N73sjUjO8OwFNa8i1JkjRrzLalBZtcDrwqyVyAJPskeRSwHPjDJLu0fS8cpe0dwIIkT2iv/wT4xlbEcAW9mVxaDHsCNwDPTrJXkh2Bk1vf17fyRyTZGXhJXz9fBd7Q18/irYhFkiSpc2ZlIltVXwU+C1ybZA1wIbB7Va0ALgZWA/8ErAHWj2j7a+CVwAWt7QPAx7cijHcDe7abulYBR1fVPcAS4CpgFbCyqr7Syt8BXEsv2b6tr583AkPthrFbgdduRSySJEmdk6oadAzTyqb1pUl2Bb4JnFZVNw46rsk0Z/7Cmn/KueNXnOHWnXXcoEOQJEnjSLKyqh50Az7MvjWyE7E0yYHALsAnZ1oSK0mSNFM4I7udJXkl8BcjipdX1emj1Z8KQ0NDNTw8PKjDS5IkTZgzsgNUVecD5w86DkmSpJlmVt7sJUmSpO4zkZUkSVInmchKkiSpk1wjOwutuXs9C5ZcOugwJp1fpyVJ0uzijKwkSZI6yURWkiRJnWQiK0mSpE4ykZUkSVInmchKkiSpk0xkgSRvSrLroOMYKcnVSYba9v9Kskd7vH7QsUmSJA2aiWzPm4AtSmST7LidYhlVVb2gqn4G7AGYyEqSpFlvRiWySd6S5I1t+/1Jrmzbz0nymSQfSzKc5JYkZ7Z9bwT2Bq5KclUrOzbJtUluTHJBkrmtfF2Ss5PcCLxkMzEcnmRVe5yTZG0rPzXJh/vqXZLkqLb9oLhG6Xddkr2As4D9ktzc+v9UkhP66n0myfGjtD+tHWN4433rt3RoJUmSpp0ZlcgC1wDPbNtDwNwkO7eybwJvraoh4GDg2UkOrqoPAj8Ajq6qo1uy+NfAMVV1GDAMvLnvGD+uqsOq6vObieF84A1VdcgWxP2guMaouwT4XlUtrqq3AP8vcCpAknnA04AH/dpBVS2tqqGqGtpx13lbEJokSdL0NNMS2ZXA4UkeBmwArqWX0D6TXpL7x2029SbgScCBo/TxlFa+PMnNwCnAY/v2f2FzB0+yB7BHVX2zFf3DBOOeSFyjqqpvAAuTPBI4GfhiVd0/0faSJEldNaN+oraqfpPkLnozlN8CVgNHA08AfgWcARxRVT9NsgzYZZRuAlxRVSdv5jC/3Mrw7ud3/+OwC0CSx00wrrF8CngF8FLglVsZnyRJUqfMtBlZ6M28nkFvKcE1wGvpzXQ+jF4Suj7Jo4Hn97W5F9i9bV8HPD3JEwCS7JZk/4kcuN2M9bMkz2hFL+/bvQ5YnGSHJPsCR7byseIaTX+smyyjd8MaVXXrRGKVJEnqupmayM4Hrq2qfwN+DVxTVavoJbS3A58Flve1WQpcluSqqvohvRndzyVZTW95wgFbcPxXAh9pyxLSV74cuAu4FfggcCPAOHE9SFX9mN6yh7VJzmll/wbcRm99riRJ0qyQqhp0DDNWkgXAJVV10HY+zq7AGuCwqhr3KwnmzF9Y8085d3uGNBDrzjpu0CFIkqRJlmRluyn+QWbijOyskuQYerOxH5pIEitJkjRTzKibvaZSko8ATx9R/IGq+u3H+1W1Dtius7FV9TV+91sVJEmSZgWXFsxCQ0NDNTw8POgwJEmSxuXSAkmSJM04JrKSJEnqJBNZSZIkdZI3e81Ca+5ez4Illw46jEnn129JkjS7OCMrSZKkTjKRlSRJUieZyEqSJKmTTGQlSZLUSSaykiRJ6iQTWUmSJHXSpCWySU5N8uHJ6k9bJsmCJC8bdBySJElTxRnZmWMBYCIrSZJmjXET2SS7Jbk0yaoka5OclOSIJN9qZTck2b1V3zvJZUm+k+Rv+/o4Nsm1SW5MckGSua18XZL3Jrk5yXCSw5JcnuR7SV7b1/4tSVYkWZ3kzDFiXZDk9iTLknw7yWeSHJNkeYvpyFbvyBbPTe08ntjKT01y0WbO4WMtxlv6Y0jygnbMlUk+mOSSvnE7r43PTUmO7zvGl5Nc0c7/z5O8udW5LsnDW739Whwrk1yT5IBWvqwd51tJ7kxyYgvlLOCZbSz/cpSxOa3FP7zxvvXj/bFLkiRNexOZkX0e8IOqOqSqDgIuA74A/EVVHQIcA/yq1V0MnAQsAk5Ksm+SvYC/Bo6pqsOAYeDNff3/a1UtBq4BlgEnAk8BzoReEgwsBI5s/R+e5FljxPsE4H3AAe3xMuAZwBnAf2t1bgeeWVWHAm8H/qav/YPOoZW/taqGgIOBZyc5OMkuwCeA51fV4cAj+/p5K3BlVR0JHA2ck2S3tu8g4I+AI4D3APe1WK4F/rTVWQq8ofV7BvDRvr7nt3N6Ib0EFmAJcE1VLa6q948clKpaWlVDVTW0467zxhg+SZKkbpjIT9SuAd6X5GzgEuBnwD1VtQKgqn4OkATg61W1vr2+FXgssAdwILC81XkIvYRtk4v7jjO3qu4F7k2yIckewLHtcVOrN5deYvvNzcR7V1WtaTHc0mKqJGvoffwOMA/4ZJKFQAE797Uf7Ry+D/xxktPamM1v57QDcGdV3dXafg44rW0fC7woyRnt9S7AY9r2VX3nuR74x74xOLjNWD8NuKCNGcCcvhi/XFUPALcmefRmxkGSJGlGGzeRrapvJzkMeAHwbuDKMapv6Nve2PoPcEVVnTxOmwdGtH+gr/17q+oT48U6Sgz9fW7qD+Bd9JLJ/5xkAXD1WOeQ5HH0ZkWPqKqfJllGLzEdS4AXHg9gEwAAGqhJREFUV9Udv1OYPHkCMe4A/KzNVI+mv302U0eSJGlGm8ga2b3pffT9aeAc4MnA/CRHtP27JxkrIb4OeHqSJ7T6uyXZfwtivBx4Vd+62n2SPGoL2o9mHnB32z51AvUfBvwSWN9mQJ/fyu8AHt+SYegtSdjkcuANaVOqSQ6daHBtlvuuJC9pbZPkkHGa3QvsPk4dSZKkGWMia2QXATckuRn47/TWlJ4EfCjJKuAKxpidrKof0ksWP5dkNb1lBQdMNMCq+irwWeDatjzgQrY9Yftb4L1JbmJis9Kr6C1tuL3FsryV/wp4PXBZkpX0kslNd1K9i96ShdVticO7tjDGlwOvbmN8C3D8OPVXAxvTuwHvQTd7SZIkzTSpqkHH0GlJ5lbVL9rM60eA74x2s9V0Mmf+wpp/yrmDDmPSrTvruEGHIEmSJlmSle2G+wfxe2S33WvabPUt9JYsTHQtryRJkrbBRL61YNpJ8gjg66Ps+oOq+vFUxtJmX6f1DKwkSdJM1MlEtiWrm7ujX+NYtM88hv0YXpIkdZxLCyRJktRJJrKSJEnqJBNZSZIkdVIn18hq26y5ez0Lllw66DDG5ddpSZKksTgjK0mSpE4ykZUkSVInmchKkiSpk0xkJUmS1EkmspIkSeokE9kplmRBkrXb2MeLkiyZrJgkSZK6yK/f6qCquhi4eNBxSJIkDZIzsoOxU5LPJLktyYVJdk3y9iQrkqxNsjRJAJK8McmtSVYn+XwrOzXJh9v2siQfTPKtJHcmOXGQJyZJkjRVTGQH44nAR6vq94GfA68HPlxVR1TVQcBDgRe2ukuAQ6vqYOC1m+lvPvCM1uas0SokOS3JcJLhjfetn8RTkSRJGgwT2cH4flUtb9ufppeEHp3k+iRrgOcAT2r7VwOfSfIK4P7N9Pflqnqgqm4FHj1ahapaWlVDVTW0467zJu9MJEmSBsREdjBqlNcfBU6sqkXA3wG7tH3HAR8BDgNWJBltXfOGvu1McqySJEnTkonsYDwmyVPb9suAf27bP0oyFzgRIMkOwL5VdRXwV8A8YO5UBytJkjQd+a0Fg3EHcHqS84BbgY8BewJrgf8PWNHq7Qh8Osk8ejOtH6yqn7X7wCRJkma1VI38lFsz3Zz5C2v+KecOOoxxrTvruEGHIEmSBizJyqoaGm2fSwskSZLUSSaykiRJ6iTXyM5Ci/aZx7Af20uSpI5zRlaSJEmdZCIrSZKkTjKRlSRJUieZyEqSJKmTvNlrFlpz93oWLLl00GGMyu+OlSRJE+WMrCRJkjrJRFaSJEmdZCIrSZKkTjKRlSRJUieZyEqSJKmTOpHIJjkhyYF9r69OMrQN/e2d5MK2vTjJCyYjznGOuS7JXtv7OJIkSbNFJxJZ4ATgwHFrTUCSnarqB1V1YitaDGz3RFaSJEmTa2CJbJIvJ1mZ5JYkp7WyX/TtPzHJsiRPA14EnJPk5iT7tSovSXJDkm8neWZrs0uS85OsSXJTkqNb+alJLk5yJfD1JAuSrE3yEOCdwEmt75OSfCfJI1u7HZJ8d9PrUc7hJa2fVUm+2cp2TPI/WvnqJG/oa/KGJDe2+A5o9XdLcl47l5uSHN8X85eTXNFmc/88yZtbneuSPLzV2y/JZW0sr9nU7yixnpZkOMnwxvvWb9WfmSRJ0nQyyB9EeFVV/STJQ4EVSb44WqWq+laSi4FLqmrTcgCAnarqyLYs4L8DxwCn95rUopbQfTXJ/q2rw4CD2zEXtL7/PcnbgaGq+vPW9wHAy4FzW5+rquqHmzmHtwPPraq7k+zRyk4DFgCLq+r+TQln86OqOizJ64EzgP8CvBW4sqpe1fq4IcnXWv2DgEOBXYDvAn9VVYcmeT/wpy3GpcBrq+o7SZ4MfBR4zijjuLTVZc78hbWZ85EkSeqMQS4teGOSVcB1wL7Awi1sf1F7XkkvcQR4BvBpgKq6HfgXYFMie0VV/WQC/Z5HL0kEeBVw/hh1lwPLkrwG2LGVHQN8oqrub3H0H3O0mI8FliS5GbiaXtL6mLbvqqq6tyXS64F/bOVrgAVJ5gJPAy5o7T8BzJ/AOUqSJHXeQGZkkxxFL+F7alXdl+Rqeglc/0zhLuN0s6E9b2Ri5/HLicRWVd9P8m9JngMcSW92dnN1X9tmQY8DViY5fCtiDvDiqrqjv2Lrd0Nf0QN9rx9o7XcAflZVi8c/M0mSpJllUDOy84CftiT2AOAprfzfkvx+kh2A/9xX/15g9wn0ew0t8WxLCh4D3DFmi9H7/nt6M7sXVNXGzTVMsl9VXV9Vbwd+SG9m+Qrgz5Ls1Oo8fHPtm8vprZ1Nq3/oOPV/q6p+DtyV5CWtbZIcMtH2kiRJXTaoRPYyYKcktwFn0VteALAEuAT4FnBPX/3PA29pNzrtx+Z9FNghyRrgC8CpVbVhjPoAVwEHbrrZq5VdDMxl7GUF0LsBbU2StS3mVfSS4H8FVrelEy8bp493ATu3+re011vi5cCr27FuAY7fwvaSJEmdlCrv+xkpve+ofX9VPXPQsWwPc+YvrPmnnDvoMEa17qzjBh2CJEmaRpKsrKpRfz9gkN9aMC0lWQK8jjHWxkqSJGnwTGRHqKqz6C13+K0kbwVeMqLqBVX1nikLTJIkSb/DpQWz0NDQUA0PDw86DEmSpHGNtbSgKz9RK0mSJP0OE1lJkiR1komsJEmSOsmbvWahNXevZ8GSSwcdxm/5lVuSJGlrOCMrSZKkTjKRlSRJUieZyEqSJKmTTGQlSZLUSSaykiRJ6iQTWUmSJHWSiawkSZI6yUR2EiXZLcmlSVYlWZvkpCSHJ/lGkpVJLk8yv9V9TZIVre4Xk+zayl/S2q5K8s1WtkuS85OsSXJTkqNb+alJLkpyWZLvJPnbwZ29JEnS1DKRnVzPA35QVYdU1UHAZcCHgBOr6nDgPOA9re5FVXVEVR0C3Aa8upW/HXhuK39RKzsdqKpaBJwMfDLJLm3fYuAkYBFwUpJ9RwssyWlJhpMMb7xv/WSesyRJ0kD4y16Taw3wviRnA5cAPwUOAq5IArAjcE+re1CSdwN7AHOBy1v5cmBZkv8JXNTKnkEvIaaqbk/yL8D+bd/Xq2o9QJJbgccC3x8ZWFUtBZYCzJm/sCbrhCVJkgbFRHYSVdW3kxwGvAB4N3AlcEtVPXWU6suAE6pqVZJTgaNaH69N8mTgOGBlksPHOeyGvu2N+GcqSZJmCZcWTKIkewP3VdWngXOAJwOPTPLUtn/nJE9q1XcH7kmyM/Dyvj72q6rrq+rtwA+BfYFrNtVJsj/wGOCOKTotSZKkacnZu8m1CDgnyQPAb4DXAfcDH0wyj954nwvcArwNuJ5esno9vcSW1n4hEODrwCrgduBjSda0/k6tqg1tuYIkSdKslCqXS842c+YvrPmnnDvoMH5r3VnHDToESZI0TSVZWVVDo+1zaYEkSZI6yURWkiRJnWQiK0mSpE7yZq9ZaNE+8xh2XaokSeo4Z2QlSZLUSSaykiRJ6iQTWUmSJHWSa2RnoTV3r2fBkku3W/9+L6wkSZoKzshKkiSpk0xkJUmS1EkmspIkSeokE1lJkiR1komsJEmSOslEtuOSfKs9L0jyskHHI0mSNFVMZKehJBP+WrSqelrbXACYyEqSpFnDRHY7aTOktyf5TJLbklyYZNckhyf5RpKVSS5PMr/VvzrJuUmGgb9or9+fZLi1PyLJRUm+k+Tdfcf5Rds8C3hmkpuT/OUATlmSJGlKmchuX08EPlpVvw/8HDgd+BBwYlUdDpwHvKev/kOqaqiq3tde/3tVDQEfB77S2h8EnJrkESOOtQS4pqoWV9X7RwaS5LSWFA9vvG/9ZJ6jJEnSQPjLXtvX96tqedv+NPDf6CWiVyQB2BG4p6/+F0a0v7g9rwFuqap7AJLcCewL/HiigVTVUmApwJz5C2vLTkOSJGn6MZHdvkYmjPfSS0ifupn6vxzxekN7fqBve9Nr/+wkSdKs5tKC7esxSTYlrS8DrgMeuaksyc5JnjRJx7oX2H2S+pIkSZr2TGS3rzuA05PcBuxJWx8LnJ1kFXAz8LQx2m+J1cDGJKu82UuSJM0Gfjy9fd1fVa8YUXYz8KyRFavqqM29rqqrgas3s29ue/4N8JxtjliSJKkjnJGVJElSJzkju51U1Tp631AgSZKk7cBEdhZatM88hs86btBhSJIkbROXFkiSJKmTTGQlSZLUSSaykiRJ6iQTWUmSJHWSN3vNQmvuXs+CJZduVdt13iQmSZKmCWdkJUmS1EkmspIkSeokE1lJkiR1komsJEmSOslEVpIkSZ00YxPZJO9IcsYo5QuSrG3bQ0k+OPXRPViSPZK8vu/13kkuHGRMkiRJ09mMTWQnoqqGq+qNg46j2QP4bSJbVT+oqhMHGI8kSdK01plEts2k3p7kM0luS3Jhkl2TrEuyV6szlOTqvmaHJLk2yXeSvGaUPo9Kcknbnpvk/CRrkqxO8uIxYvlFknOS3JLka0mOTHJ1kjuTvKjV2bHVWdH6+7O+9m/pKz+zFZ8F7Jfk5tauf+b41CQXJbmsncvf9vX16iTfTnJDkr9L8uHNxHxakuEkwxvvWz/RYZckSZq2uvaDCE8EXl1Vy5OcR98M5mYcDDwF2A24KclYvwLwNmB9VS0CSLLnGHV3A66sqrck+RLwbuA/AQcCnwQuBl7d+jsiyRxgeZKvAgvb40ggwMVJngUsAQ6qqsXt+AtGHHMxcCiwAbgjyYeAjS3uw4B7gSuBVaMFXFVLgaUAc+YvrDHOTZIkqRO6lsh+v6qWt+1PA+MtC/hKVf0K+FWSq+gljzdvpu4xwEs3vaiqn47R778Dl7XtNcCGqvpNkjXAglZ+LHBwkk3LA+bRS2CPbY+bWvncVv6v45zL16tqPUCSW4HHAnsB36iqn7TyC4D9x+lHkiRpRuhaIjtyJrGA+/mPJRK7TKD+ZPhNVW3q6wF6s6RU1QNJNo1pgDdU1eX9DZM8F3hvVX1iRPmCcY65oW97I937s5MkSZpUnVkj2zwmyVPb9suAfwbWAYe3spHrWo9PskuSRwBHASvG6PsK4PRNL8ZZWjARlwOvS7Jz62//JLu18lclmdvK90nyKHpLA3bfwmOsAJ6dZM+WQG92Xa8kSdJM07VE9g7g9CS3AXsCHwPOBD6QZJjeTGW/1cBVwHXAu6rqB2P0/W5gzyRrk6wCjt7GWP8euBW4sd209Qlgp6r6KvBZ4Nq2FOFCYPeq+jG9dbRrk5wzkQNU1d3A3wA3AMvpJfXeySVJkmaF/Mcn5NNb++j9kqo6aMChTCtJ5lbVL9qM7JeA86rqS2O1mTN/Yc0/5dytOt66s47bqnaSJElbI8nKqhoabV/XZmT1YO9IcjOwFrgL+PKA45EkSZoSnblhqKrWAVM6G5vkemDOiOI/qao1UxnHWKrqQb9eJkmSNBt0JpEdhKp68qBj2B4W7TOPYZcISJKkjnNpgSRJkjrJRFaSJEmdZCIrSZKkTjKRlSRJUieZyEqSJKmTTGQlSZLUSSaykiRJ6iQTWUmSJHWSiawkSZI6yURWkiRJnWQiK0mSpE4ykZUkSVInmchKkiSpk0xkJUmS1EkmspIkSeokE1lJkiR1komsJEmSOslEVpIkSZ1kIitJkqROMpGVJElSJ5nISpIkqZNMZCVJktRJJrKSJEnqJBNZSZIkdVKqatAxaIoluRe4Y9BxdMRewI8GHUSHOF4T51htGcdr4hyrLeN4TdygxuqxVfXI0XbsNNWRaFq4o6qGBh1EFyQZdqwmzvGaOMdqyzheE+dYbRnHa+Km41i5tECSJEmdZCIrSZKkTjKRnZ2WDjqADnGstozjNXGO1ZZxvCbOsdoyjtfETbux8mYvSZIkdZIzspIkSeokE1lJkiR1kolsxyV5XpI7knw3yZJR9s9J8oW2//okC/r2/ddWfkeS5060zy7b2vFK8p+SrEyypj0/p6/N1a3Pm9vjUVN3RtvPNozVgiS/6huPj/e1ObyN4XeTfDBJpu6Mtq9tGK+X943VzUkeSLK47Zut19azktyY5P4kJ47Yd0qS77THKX3ls/naGnW8kixOcm2SW5KsTnJS375lSe7qu7YWT9X5bE/beG1t7BuPi/vKH9fes99t7+GHTMW5TIVtuLaOHvH31q+TnND2Te21VVU+OvoAdgS+BzweeAiwCjhwRJ3XAx9v2y8FvtC2D2z15wCPa/3sOJE+u/rYxvE6FNi7bR8E3N3X5mpgaNDnN43GagGwdjP93gA8BQjwT8DzB32ugx6vEXUWAd/z2mIBcDDwKeDEvvKHA3e25z3b9p5eW5sdr/2BhW17b+AeYI/2ell/3Znw2Jaxavt+sZl+/yfw0rb9ceB1gz7X6TBefXUeDvwE2HUQ15Yzst12JPDdqrqzqv4d+Dxw/Ig6xwOfbNsXAn/QZiqOBz5fVRuq6i7gu62/ifTZVVs9XlV1U1X9oJXfAjw0yZwpiXowtuXaGlWS+cDDquq66v1t9ynghMkPfSAma7xObm1nsnHHqqrWVdVq4IERbZ8LXFFVP6mqnwJXAM+b7dfW5sarqr5dVd9p2z8A/n9g1F9HmiG25doaVXuPPofeexZ67+FZf22NcCLwT1V13/YLdfNMZLttH+D7fa//dysbtU5V3Q+sBx4xRtuJ9NlV2zJe/V4M3FhVG/rKzm8fobxthnykua1j9bgkNyX5RpJn9tX/3+P02VWTdW2dBHxuRNlsvLa2tO1sv7bGleRIerNu3+srfk9bcvD+GfIf820dq12SDCe5btPH5PTeoz9r79mt6XM6m6x/71/Kg//emrJry0RW2gJJngScDfxZX/HLq2oR8Mz2+JNBxDaN3AM8pqoOBd4MfDbJwwYc07SX5MnAfVW1tq/Ya0vbrM1Y/wPwyqraNLP2X4EDgCPofTT8VwMKbzp5bPV+fvVlwLlJ9ht0QNNdu7YWAZf3FU/ptWUi2213A/v2vf69VjZqnSQ7AfOAH4/RdiJ9dtW2jBdJfg/4EvCnVfXbWY2qurs93wt8lt7HNV231WPVlqv8GKCqVtKbAdq/1f+9cfrsqm26tpoHzWrM4mtrS9vO9mtrs9p/Ii8F3lpV120qr6p7qmcDcD5eW/3vtzvprU8/lN57dI/2nt3iPqe5yfj3/o+BL1XVbzYVTPW1ZSLbbSuAhe2OyofQ+4fw4hF1LgY23dl7InBlW0N2MfDS9O6kfhywkN7NEhPps6u2eryS7EHvH4MlVbV8U+UkOyXZq23vDLwQWEv3bctYPTLJjgBJHk/v2rqzqu4Bfp7kKe0j8j8FvjIVJzMFtuW9SJId6P2D8Nv1sbP82tqcy4Fjk+yZZE/gWOByr63RtfpfAj5VVReO2De/PYfems9ZfW21a2pO294LeDpwa3uPXkXvPQu99/Csv7b6nMyI/4BP+bU1VXeV+dg+D+AFwLfpzXq9tZW9E3hR294FuIDezVw3AI/va/vW1u4O+u7wHa3PmfLY2vEC/hr4JXBz3+NRwG7ASmA1vZvAPgDsOOjzHPBYvbiNxc3AjcAf9vU5RO8vte8BH6b9uuBMeGzje/Eo4LoR/c3ma+sIeuv1fklvRuyWvravamP4XXoflXttbWa8gFcAvxnx99bitu9KYE0bs08Dcwd9ngMeq6e18VjVnl/d1+fj23v2u+09PGfQ5zno8Wr7FtCbwd1hRJ9Tem35E7WSJEnqJJcWSJIkqZNMZCVJktRJJrKSJEnqJBNZSZIkdZKJrCRJkjrJRFaSpoEkG9tP0a5N8o/tu4vHqv+OJGeMU+eEJAf2vX5nkmMmIdZlSU4cv+bkSfKmJLtO5TElTX8mspI0PfyqqhZX1UHAT4DTJ6HPE4DfJrJV9faq+tok9Dul2g9svAkwkZX0O0xkJWn6uRbYByDJfkkuS7IyyTVJDhhZOclrkqxIsirJF5PsmuRpwIuAc9pM736bZlKTPC/JBX3tj0pySds+Nsm1SW5MckGSuWMFmmRdkve2YwwnOSzJ5Um+l+S1ff1/M8mlSe5I8vH2a2YkOTnJmjYTfXZfv79I8r4kq+j9eMvewFVJrmr7P9aOd0uSM0fEc2aLf82m8UoyN8n5rWx1khdvzflKml5MZCVpGmmzj3/Af/xU5FLgDVV1OHAG8NFRml1UVUdU1SHAbfR+lehbrY+3tJne7/XV/xrw5CS7tdcnAZ9vP83518AxVXUYMAy8eQJh/2tVLQauAZbR+znPpwBn9tU5EngDvRni/YA/SrI3cDbwHGAxcESSE1r93YDrq+qQqnon8APg6Ko6uu1/a1UNAQcDz05ycN+xftTi/1gbM4C3AeuralFVHQxcuQ3nK2ma2GnQAUiSAHhokpvpzcTeBlzRZgefBlzQ+9lyAOaM0vagJO8G9gDmApePdaCquj/JZcAfJrkQOA74v4Fn00s0l7fjPYTe7PB4NiXda+j9HOW9wL1JNvSt9b2hqu4ESPI54Bn0fj716qr6YSv/DPAs4MvARuCLYxzzj5OcRu/fsfkt7tVt30XteSXwR237GHq/Jb9pDH6a5IVbeb6SpgkTWUmaHn5VVYvbDU2X01sjuwz4WZvtHMsy4ISqWpXkVOCoCRzv88Cf01uPO1xV96aXzV1RVSdvYewb2vMDfdubXm/6d2bk76GP9/vov66qjaPtSPI4ejOtR7SEdBmwyyjxbGTsf+e29nwlTRMuLZCkaaSq7gPeCPxfwH3AXUleApCeQ0ZptjtwT5KdgZf3ld/b9o3mG8BhwGvoJbUA1wFPT/KEdrzdkuy/jae0yZFJHtfWxp4E/DNwA71lAXu1JRUnt7hG038uDwN+CaxP8mjg+RM4/hX03UCXZE+27/lKmgImspI0zVTVTfQ+Jj+ZXmL66nbT0y3A8aM0eRtwPbAcuL2v/PPAW5LclGS/EcfYCFxCLwm8pJX9EDgV+FyS1fQ+Zn/QzWVbaQXwYXrLJu4CvlRV9wBLgKuAVcDKqvrKZtovBS5LclVVrQJuoneun6V33uN5N7Bnu6lsFb31ttvzfCVNgVSN9+mOJElbL8lRwBlV9cJBxyJpZnFGVpIkSZ3kjKwkSZI6yRlZSZIkdZKJrCRJkjrJRFaSJEmdZCIrSZKkTjKRlSRJUif9H2kFF51LRX6DAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 720x720 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "KHfeovTXdqKo"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### **Funtional need repair model**"
      ],
      "metadata": {
        "id": "CiOTKtEhFUhC"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Create training and test sets: test size 0.2\n",
        "X_train, X_test, y_train, y_test= train_test_split(X, y_repair, test_size = 0.2, random_state=42, stratify = y_repair)\n",
        "\n",
        "param_grid = { 'n_estimators': [  125, 150, 175],\n",
        "             'max_features':['sqrt'],\n",
        "             'max_depth' : [12,15, 17 ,20, 25, 30],\n",
        "             'min_samples_split':[18,9,10]}\n",
        "\n",
        "\n",
        "\n",
        "#Create a based model\n",
        "rf =     RandomForestClassifier( random_state=42, warm_start = True)  \n",
        "\n",
        "grid_search_rf = GridSearchCV(estimator = rf, param_grid = param_grid, cv = 5,  verbose = 2,  scoring = 'accuracy')\n",
        "\n",
        "#fitmodel\n",
        "grid_search_rf .fit(X_train, y_train)\n",
        "\n",
        "#print best parameters \n",
        "print('Best Score:', grid_search_rf .best_score_)\n",
        "print('Parameters:', grid_search_rf .best_params_)\n",
        "print('Best Model:', grid_search_rf .best_estimator_)\n"
      ],
      "metadata": {
        "id": "hmox9Eq5FYiJ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#score model\n",
        "best_model =  RandomForestClassifier(max_depth=20, max_features='sqrt', min_samples_split=9,  n_estimators=125, random_state=42, warm_start=True)\n",
        "\n",
        "X_train, X_test, y_train, y_test= train_test_split(X, y_repair, test_size = 0.2, random_state=42 , stratify=y_repair)\n",
        "\n",
        "#fit the best model\n",
        "best_model.fit(X_train, y_train)\n",
        "y_pred = best_model.predict(X_test)\n",
        "y_pred_train = best_model.predict(X_train)\n",
        "\n",
        "\n",
        "#print best model scores on test data\n",
        "print(\"Accuracy score train: {}\".format(accuracy_score(y_train, y_pred_train)))\n",
        "print(\"Accuracy score test: {}\".format(accuracy_score(y_test, y_pred)))\n",
        "print()\n",
        "\n",
        "#print(confusion_matrix(y_test, y_pred))\n",
        "print(classification_report(y_test, y_pred))\n",
        "\n",
        "feature_importances(best_model)\n",
        "\n",
        "confusion_matrix_plotter(y_test, y_pred)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "Fc_bGkXpFWqw",
        "outputId": "30d1cfac-2563-4f59-d45d-09134af69a1f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Accuracy score train: 0.9510511584839748\n",
            "Accuracy score test: 0.937962962962963\n",
            "\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.94      0.99      0.97     11017\n",
            "           1       0.75      0.22      0.34       863\n",
            "\n",
            "    accuracy                           0.94     11880\n",
            "   macro avg       0.84      0.61      0.65     11880\n",
            "weighted avg       0.93      0.94      0.92     11880\n",
            "\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x720 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAdkAAAE+CAYAAADBBmo4AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3dd5wV5d338c9vBayoIIgo3rEE+2PBEtRoiL1F1Bi78kQj0RiixttoTCLRaDS3xoYlomIv2PW2ocGKT1QwGgugooKAKFVAirTr+WMH3AUWlyMz6+583r7mtedcc5051/DC/fKbuWYmUkpIkqRlr6qhByBJUlNlyEqSlBNDVpKknBiykiTlxJCVJCknzRp6ADWtuM2vneqsRm/SwKsbegjSMrFCMyKP7Vb6u37GG1fnMp48WclKkpST71QlK0kqgShPfWfISpKKFY3uqG/FDFlJUrGsZCVJyomVrCRJObGSlSQpJ1aykiTlxEpWkqScWMlKkpSTElWy5dlTSZIKZiUrSSqWh4slScpJiQ4XG7KSpGJZyUqSlBMrWUmScmLISpKUkyoPF0uSlA8rWUmScuLEJ0mSclKiSrY8eypJ+m6IqGz5xs1Gn4gYGxHv1GhrHRHPRMQH2c9WWXtExFURMSwi3oqITjU+0y3r/0FEdKvRvm1EvJ195qqIbx6UIStJKlZUVbZ8s1uAfRZqOxvon1LqCPTP3gPsC3TMlu7AdVAdykBP4AfADkDP+cGc9TmxxucW/q5FGLKSpGLlVMmmlF4EJi7U3BW4NXt9K3BQjfbbUrVXgNUjoj2wN/BMSmliSmkS8AywT7Zu1ZTSKymlBNxWY1t1MmQlScWqsJKNiO4RMajG0r0e39YupTQme/0Z0C57vQ4wska/UVnbktpHLaZ9iZz4JEkqVoWzi1NKvYHelX5tSilFRKr085WwkpUkFSu/c7KL83l2qJfs59isfTSwbo1+HbK2JbV3WEz7EhmykqRi5XROtg6PAvNnCHcDHqnRflw2y7gzMDk7rNwP2CsiWmUTnvYC+mXrpkRE52xW8XE1tlUnDxdLkpqEiLgb6AK0iYhRVM8Svhi4NyJOAEYAh2XdnwD2A4YB04GfA6SUJkbEX4CBWb/zU0rzJ1P9iuoZzCsCT2bLEhmykqRi5XQzipTSkXWs2n0xfRNwSh3b6QP0WUz7IGCLpRmTIStJKlaJ7vhkyEqSiuW9iyVJyomVrCRJObGSlSQpJ1aykiTlxEpWkqR81OMJcU2GIStJKpQhK0lSXsqTsYasJKlYVrKSJOXEkJUkKSeGrCRJOTFkJUnKS3ky1pCVJBWrTJVsee5tJUlSwaxkJUmFKlMla8hKkgplyEqSlBNDVpKkvJQnYw1ZSVKxrGQlScqJIStJUk4MWUmS8lKejDVkJUnFspKVJCknhqwkSTkxZCVJyokhK0lSXsqTsYasJKlYVrKSJOXEkJUkKSeGrCRJeSlPxhqy32U7d9qQ047bg06brsvaa67Oiefezh3/+2qtPn/45X6c8NOdWb3ligx8ZwSnXdSXIR99tmD91pt04IJTD2Lbzf+LuXMTD/d/k7P+/gDTZsxa0GfGG1cv8t09LryHG+8fAMAu23akxzE/ZrvNv8dqq6zIhyPHcfVdz3HbI6/ktOdStXHjxnLlZX9nwEsvMG3aNDp0WJc/nPtnttt+BwD+dM7ZPPrIQ7U+83+23Io77r63IYYrLSLXkI2IfYArgeWAG1NKF+f5fU3NKistz+Bhn3LXY69y4/nHLbL+jP+7B6ceuxvde97B+8M/55zu+/L4P3qw5UHn8+X0r2jfdjUe/0cPHnzmDU6/+F5WXXkFLjnzp9xw/rEcdeZNtbZ18vl38uSL7yx4P/nLmQted95qfd794FMuv+WfjBk/mT132pRr/ngkX301h75PDcrvD0ClNmXKFLodcyTbdNqWXtf2plXrVoweNYrWrdeo1a/zjjtx4UX/s+B98+bNix6qlpKHi5eBiFgOuAbYExgFDIyIR1NKg/P6zqam34DB9BtQ/cfV+7xjF1l/ylE/5tKbn+Hh/m8C8Itzb+eT/hdx+L7bcdMDL7PvLlswb17i1Iv6Mm9eAqDHhX0ZdN85bLBuGz4aOX7BtiZPncHnE6YudhyX9Hm61vsb7hvAj7bbiIP22NqQVW5u6XMjbdu0rRWgHTqsu0i/5i1a0KZt2yKHpm+pTCFbleO2dwCGpZQ+SinNAu4Buub4faWy3jpr0L7tavT/15AFbTO/ms2Afw+j81YbALB8i2bMnjN3QcACzPiq+jDxTltvWGt7l555KCOfvZgBd5zJLw794Tf+T9By5RWYNGX6stodaRHP9f8nW2y5FWeecRpddtmRww7pyt133kFKqVa/N//9Ol122ZGf7Lc35537RyZMmNBAI1Z9RURFS2OUZ8iuA4ys8X5U1qZlYK02qwIwdmLt6nPshKm0W6N63fOvvUeb1Vfhv3++J82bLcfqLVfkgt9U/ztnrbarLfjMedc+xrFn9WH/k3pxX7/Xufi3B/O7E/aq87v33WULfrzDxvR54OVlvVvSAqNGjeTee+6iQ4d1ua73TRx17HFcefnfueeuOxf02emHu3DBX//GDTfdwhlnnsU7b7/Ficd3Y9asWUvYshpamUK2wSc+RUR3oDtAsw5daNZm8wYeUdMx5KPPOPHc27n4jEP48yk/Yc7cuVx79wt8Nn4Kad68Bf0uvuGpBa/fen80y1VV8bsT9uZvN/ZbZJs7brUBt/y1G2dcch+D3h1RyH6onObNS2y+xRacevoZAGy66WZ8MmIEfe+5kyOPPgaAfffbf0H/jhttzGabb86+e+7Giy88zx571v0PRTWwxpmXFckzZEcDNU+gdMjaakkp9QZ6A6y4za/Twuu1eJ+NnwLAmq1bMvKzSQva11yjJZ9PmLLgfd+nBtH3qUGs2bol02Z8RUrwm2N24+NRdR9Se+2d4azWckXWbN2yVqW809Yb8FCvk/nLdY9zw30Dctgr6Wtt27Zlgw1rn9ZYf4MNGHPHmDo/s+aa7VizXTs+GTE859Hp22isVWkl8jxcPBDoGBHrR0QL4Ajg0Ry/r1SGj57AmHGT2a3zJgvalm/RjJ232ZBX/vPRIv3HTpzKtBmzOHTvTsycNZv+rwytc9tbbdSBGTNn8cXUGQvadu60IQ9f/SsuvP5Jrr7r+WW5K9Jibb1NJ4Z//HGtthHDh7N2+7Xr/MykSRMZ+/lY2rZdM+/h6VvwcPEykFKaExG/BvpRfQlPn5TSu3l9X1O08oot2HDd6lmTVRGs274VW260DpOmTGfkZ5O45q7nOPP4vXl/+Od8MGIsZ5+4D9NmzKLvk1/P+D3p8F159a2PmTrtK3bvvAl/Pe0g/tTrESZ/WR2g++26Be3WWJVX3/qYGV/N5kfbd+RPJ+9PnwdfZtbsOUD1dbIP9TqJ3ve+RN8nB9JujZYAzJ2XGD/py4L/VFQWxxzXjW7HHMkN11/H3vvsx9Chg7n7ztvpcepvAZg+bRrXXXs1e+y5F23atuXT0aO56orLaL1Ga3bbY48GHr2WJK+8jIjTgV8ACXgb+DnQnuqJt2sArwPHppRmRcTywG3AtsAE4PCU0vBsO78HTgDmAr9JKS167qy+Y1p4pl5D8nBxbbts25Gnbzx1kfbbH32F7j3vAL6+GUWrVVdi4DvDOe2iexn84deH0278y7Hs88MtWGWlFrw3/HOuuK0/dz8+cMH6PXfalPN7HMiG67alqir4eNQEbnno//GPe19k7tzq87a9zzuGYw/svMg4Rnw6gU3277msd7vRmzRw0Zt7qDIvvvA8va64jOHDP2at9mtzxFFHc9TRxxIRzJw5k9N6nMLQoYOZOmUqbdu2ZfsdfsApPU5lrfbtG3roTcIKzfI5e9rxzKcq+l3/wSX71DmeiFgHGABsllKaERH3Ak8A+wEPppTuiYh/AP9JKV0XEb8CtkwpnRQRRwAHp5QOj4jNgLupvkJmbeCfwEYppbmVjNmQlZYxQ1ZNRV4hu9HvKgvZ9//nG0P2FWArYArwMNALuBNYKzu6uiPw55TS3hHRL3v9r4hoBnwGtAXOBkgpXZRtd0G/Ssac5zlZSZIWUek52YjoHhGDaizd528zpTQauBT4BBgDTKb68PAXKaU5Wbeal5IuuMw0Wz+Z6kPKy/Ty0wa/hEeSVC6VnpOteTXKotuMVlTf8Gh94AvgPmCfyr5p2TFkJUmFqqrK5Sj0HsDHKaVxABHxILAzsHpENMuq1ZqXks6/zHRUdrh4NaonQNXr8tP68nCxJKlQEZUt3+AToHNErBTV1/vsDgwGngMOzfp0Ax7JXj+avSdb/2yqnqT0KHBERCwfEesDHYHXKt1XK1lJUqHyuOY1pfRqRNwP/BuYA7xB9aHlx4F7IuKCrG3+I8huAm6PiGHARKrv5UBK6d1sZvLgbDunVDqzGAxZSVLB8rpONqXUE1j4usKPqL4cZ+G+M4Gf1bGdC4ELl8WYPFwsSVJOrGQlSYVqrLdIrIQhK0kqlCErSVJOSpSxhqwkqVhWspIk5aREGWvISpKKZSUrSVJOSpSxhqwkqVhWspIk5aREGWvISpKKZSUrSVJOSpSxhqwkqVhWspIk5aREGWvISpKKZSUrSVJOSpSxhqwkqVhlqmR9aLskSTmxkpUkFapMlawhK0kqVIky1pCVJBXLSlaSpJyUKGMNWUlSsaxkJUnKSYky1pCVJBWrqkQpa8hKkgpVoow1ZCVJxfKcrCRJOakqT8YaspKkYlnJSpKUkxJlrCErSSpWUJ6UNWQlSYXynKwkSTnxnKwkSTkpUcb60HZJkvJiJStJKpS3VZQkKSclylhDVpJULCc+SZKUkxJlrCErSSqW52QlScpJeSLWS3gkSQWLiIqWemx39Yi4PyKGRsSQiNgxIlpHxDMR8UH2s1XWNyLiqogYFhFvRUSnGtvplvX/ICK6fZt9NWQlSYWqisqWergSeCqltAmwFTAEOBvon1LqCPTP3gPsC3TMlu7AdQAR0RroCfwA2AHoOT+YK9rXSj8oSVIl8qhkI2I1YFfgJoCU0qyU0hdAV+DWrNutwEHZ667AbanaK8DqEdEe2Bt4JqU0MaU0CXgG2KfSfa3znGxE9AJSXetTSr+p9EslSeVV6byniOhOddU5X++UUu/s9frAOODmiNgKeB04FWiXUhqT9fkMaJe9XgcYWWNbo7K2utorsqSJT4Mq3agkSXWp9DrZLFB717G6GdAJ6JFSejUiruTrQ8PzP58ios7iMQ91hmxK6da61kmSVKmcHnU3ChiVUno1e38/1SH7eUS0TymNyQ4Hj83WjwbWrfH5DlnbaKDLQu3PVzqobzwnGxFtI+LSiHgiIp6dv1T6hZKkcsvjnGxK6TNgZERsnDXtDgwGHgXmzxDuBjySvX4UOC6bZdwZmJwdVu4H7BURrbIJT3tlbRWpz3WydwJ9gf2Bk7JBjqv0CyVJ5ZbjdbI9gDsjogXwEfBzqovJeyPiBGAEcFjW9wlgP2AYMD3rS0ppYkT8BRiY9Ts/pTSx0gHVJ2TXSCndFBGnppReAF6IiIHf+ClJkhYjrzs+pZTeBLZbzKrdF9M3AafUsZ0+QJ9lMab6hOzs7OeYiNgf+BRovSy+XJJUPiW6q2K9QvaC7PqjM4BewKrA6bmOSpKkJuAbQzal9Fj2cjLw43yHI0lq6nzUXQ0RcTOLuSlFSun4XEYkSWrSSpSx9Tpc/FiN1ysAB1N9XlaSpKXmo+5qSCk9UPN9RNwNDMhtRJKkJq1EGVvR82Q7Amsu64EAjBpwRR6blQo1LxV61zYpR/mkoedka4iIqdQ+J/sZcFZuI5IkNWllevxbfQ4XtyxiIJKkcihTJVufexf3r0+bJEn1keND279zlvQ82RWAlYA22U2S5+/iqnyLZ+tJksqtsQZmJZZ0uPiXwGnA2lQ//Hb+H8sU4OqcxyVJaqLKdLh4Sc+TvRK4MiJ6pJR6FTgmSVITVqZKtj6TvOZFxOrz32TP2PtVjmOSJDVhEZUtjVF9QvbElNIX89+klCYBJ+Y3JElSU1YVUdHSGNXnZhTLRURkz94jIpYDWuQ7LElSU+V1srU9BfSNiOuz978EnsxvSJKkpqyRFqUVqU/IngV0B07K3r8FrJXbiCRJTVpjPfRbiW+s2lNK84BXgeHADsBuwJB8hyVJaqrKNPFpSTej2Ag4MlvGA30BUko+uF2SpHpY0uHiocBLwAEppWEAEXF6IaOSJDVZXidb7RBgDPBcRNwQEbuT13OPJEmlUaZLeOoM2ZTSwymlI4BNgOeovsXimhFxXUTsVdQAJUlNS5nOydZn4tO0lNJdKaWfAB2AN/B5spKkCpXpKTxLdU1wSmlSSql3Smn3vAYkSWraosL/GqP6XCcrSdIy01ir0koYspKkQhmykiTlxOfJSpKUEytZSZJyUqJC1pCVJBWrsd5YohKGrCSpUB4uliQpJyUqZA1ZSVKxqhrpjSUqYchKkgplJStJUk7KdE52qe5dLEmS6s9KVpJUKC/hkSQpJyXKWENWklQsK1lJknJSoox14pMkqVhVFS71ERHLRcQbEfFY9n79iHg1IoZFRN+IaJG1L5+9H5atX6/GNn6ftb8XEXt/232VJKkwEVHRUk+nAkNqvP8bcHlK6fvAJOCErP0EYFLWfnnWj4jYDDgC2BzYB7g2IpardF8NWUlSoaLC5Ru3G9EB2B+4MXsfwG7A/VmXW4GDstdds/dk63fP+ncF7kkpfZVS+hgYBuxQ6b4aspKkQlVFVLRERPeIGFRj6b7Qpq8AfgfMy96vAXyRUpqTvR8FrJO9XgcYCZCtn5z1X9C+mM8sNSc+SZIKVem8p5RSb6D3YrcZcQAwNqX0ekR0qXRsy5ohK0kqVE6zi3cGDoyI/YAVgFWBK4HVI6JZVq12AEZn/UcD6wKjIqIZsBowoUb7fDU/s9Q8XCxJKlQeE59SSr9PKXVIKa1H9cSlZ1NKRwPPAYdm3boBj2SvH83ek61/NqWUsvYjstnH6wMdgdcq3VcrWUlSoQqu7s4C7omIC4A3gJuy9puA2yNiGDCR6mAmpfRuRNwLDAbmAKeklOZW+uVRHdzfDROmzfnuDEaq0IotKp7tL32nrNQ8nwO79775aUW/6w/beu1GdxsLK1lJUqEaXVJ+C4asJKlQS3FjiUbPkJUkFapMM27LtK+SJBXKSlaSVCgPF0uSlJPyRKwhK0kqWIkKWUNWklSsqhLVsoasJKlQVrKSJOUkrGQlScqHlawkSTnxnKwkSTmxkpUkKSeGrCRJOXHikyRJOakqT8YaspKkYlnJSpKUE8/JSpKUEytZNRrjx43jul6X8a8BLzF9+jTWXqcDZ55zLttsuz0AO3XafLGfO+RnR/Dfv/8Tc2bP5vprr+KVlwcwetRIVl5lZTpttwMn9zidtdqvXeSuqOReHzSQ227pw5DB7zJu7FjOu+CvHHjQIQvWTxg/nisvv5R//b+X+XLqVDptux2/O+ePfO976y3o88B9fXnqiccZOnQIX06dyuP9/sna63RogL3RknhOdhmIiD7AAcDYlNIWeX1PmU2dOoWTjj+GLbfuxKVXXcvqrVrz6ahRtGrVekGf/336+VqfGTr4Xc487RR232sfAGbOnMn7Q4fQ7YTudNx4E6Z9OZVel13Cb3/9S27r+xDNmvnvMBVj+vTpfP/7HTngwK6ce87ZtdallDj91FOoiiouv+oaVlllFe647RZO+sXxPPjIY6y40kpA9d/nzjvtTJfddufSv13UELsh1ZLnb9BbgKuB23L8jlK785Y+rNGmDef+5etfJgv/q32NNm1rvX/xhWf5r++tt6DSXaVlS6687sZafX73x54cfWhXRnz8ERt23Cin0Uu17bLrj9hl1x8B0PMP59Ra98mI4bz9n/9wz/0Ps/EmmwBwzp/+zB5dfsiTTzzOIYf+DICjj+0GwLvvvF3gyLW0ynS4uCqvDaeUXgQm5rV9wYvP92ezLbbkT2edwX6770K3Iw7h/nvuJKW02P7Tp0+jf78nOfDgQ5e43WlfTgOg5aqrLvMxS5WYNWsWAMsv32JBW1VVFS2at+DNN15vqGGpQhGVLY1RbiGr/H06ehQP3XcPa3fowOXX9OZnRx7Ldb0u54G+dy22/9NPPs7s2bPZ9ydd69zm7Nmz6HX5//DDXbuwZru18hq6tFTWW38D1mq/Nr2uvJzJk79g9uxZ3HzTDXz++WeMHzeuoYenpRQVLo1Rg4dsRHSPiEERMejWPjc09HAalXnz5rHRJptxco/T2XiTTTmg68H87IhjeODeexbb/9GH7meXLrvVOmdb05w5czjvj2fz5dSp/OHPF+Y5dGmpNG/enL9fcRWjRo6ky86d2XG7bRj02qvsvMuuRFWD/xrTUqqKqGhpjBp8VktKqTfQG2DCtDmLP86pxVqjTVvW32DDWm3fW38DPr/7jkX6vv/eEIYOfpeTfn3aYrc1Z84cep5zJh8O+4Bret/MaquvnsuYpUpttvkW9H3gYaZOncrs2bNp3bo1xx55GJtt7rzKxqZxxmVl/CdgI7bl1tvwyfCPa7WN/GQ4a7Vvv0jfRx68j7XX6cD2P9hxkXVzZs/m3LPP4MMP3ufq629eZLKU9F3SsmVLWrduzYgRwxn87jt0+fFuDT0kLa0SHS/OLWQj4m7gX8DGETEqIk7I67vK6vCjj+Odd97ilhuvZ9QnI3j2mX7cd8+d/PSwI2v1mzljBk8/+TgHdD2EWOiQy5w5c/jDWb/lnbff4ry/XkIETBg/jgnjx/HVzJlF7o5Kbvr0abw3dAjvDR1CSvMYM2YM7w0dwpgxnwLwTL+nGPjaK4waOZLnnu3PySceT5fddmfHnX+4YBvjx4/jvaFD+GTEcAA+/PBD3hs6hMmTv2iIXVIdosL/GqOoayZqQ/Bw8dJ7+aUXuP7qK/hkxHDardWenx5+FD874uhaYfrYIw/xtwt68uAT/6Rt2zVrfX7Mp6P56QF7LXbbf/jzBex/4MG5jr8pWrHFcg09hEZp0GuvcuLx3RZp/0nXgzj/wou5647buO3mPkyYMIE2bdtywIFd6X7SyTRv/vWM439c04vrr7tmkW0sfGML1c9KzfM5EfraR5Mr+l2/wwarNbqkNWSlZcyQVVORV8gOrDBkt2+EIdvgE58kSSXT6KKycoasJKlQjfX8aiUMWUlSoRrpJa8VMWQlSYUqUcYaspKkgpUoZQ1ZSVKhPCcrSVJOynRO1tsqSpKUEytZSVKhSlTIGrKSpIKVKGUNWUlSoco08clzspKkQkVUtix5m7FuRDwXEYMj4t2IODVrbx0Rz0TEB9nPVll7RMRVETEsIt6KiE41ttUt6/9BRCz61IqlYMhKkgqV0+Nk5wBnpJQ2AzoDp0TEZsDZQP+UUkegf/YeYF+gY7Z0B66D6lAGegI/AHYAes4P5koYspKkYuWQsimlMSmlf2evpwJDgHWArsCtWbdbgYOy112B21K1V4DVI6I9sDfwTEppYkppEvAMsE+lu2rISpIKVelD2yOie0QMqrF0X+z2I9YDtgFeBdqllMZkqz4D2mWv1wFG1vjYqKytrvaKOPFJklSoSm9GkVLqDfRe8rZjFeAB4LSU0pSo8WUppRQRhT633EpWklSonM7JEhHNqQ7YO1NKD2bNn2eHgcl+js3aRwPr1vh4h6ytrvaKGLKSpGLlkLJRXbLeBAxJKV1WY9WjwPwZwt2AR2q0H5fNMu4MTM4OK/cD9oqIVtmEp72ytop4uFiSVKicrpPdGTgWeDsi3szazgEuBu6NiBOAEcBh2bongP2AYcB04OcAKaWJEfEXYGDW7/yU0sRKBxUpFXp4eokmTJvz3RmMVKEVWyzX0EOQlomVmudzK//3Ppte0e/6jddaqdHdxcJKVpJUqEaXlN+CIStJKlaJUtaQlSQVqkz3LjZkJUmFKtND2w1ZSVKhSpSxXicrSVJerGQlScUqUSlryEqSCuXEJ0mScuLEJ0mSclKijDVkJUkFK1HKGrKSpEJ5TlaSpJx4TlaSpJyUKGMNWUlSsaxkJUnKTXlS1pCVJBXKSlaSpJyUKGMNWUlSsaxkJUnKidfJSpKUl/JkrCErSSpWiTLWh7ZLkpQXK1lJUqGc+CRJUk6c+CRJUl7Kk7GGrCSpWCXKWENWklQsz8lKkpQTz8lKkpSTMlWyXicrSVJOrGQlSYUqUyVryEqSCuU5WUmScmIlK0lSTkqUsYasJKlgJUpZQ1aSVCjPyUqSlBPPyUqSlJMSZaw3o5AkKS9WspKkYpWolDVkJUmFKtPEp0gpNfQYVKCI6J5S6t3Q45C+Df8eq7HwnGz5dG/oAUjLgH+P1SgYspIk5cSQlSQpJ4Zs+XgeS02Bf4/VKDjxSZKknFjJSpKUE0NWkqScGLKSJOXEkG3iImLjiNgxIppHxHINPR7p2/DvsBobJz41YRFxCPBXYHS2DAJuSSlNadCBSUspIjZKKb2fvV4upTS3occk1YeVbBMVEc2Bw4ETUkq7A48A6wJnRcSqDTo4aSlExAHAmxFxF0BKaa4VrRoLQ7ZpWxXomL1+CHgMaA4cFVGmxyarsYqIlYFfA6cBsyLiDjBo1XgYsk1USmk2cBlwSETsklKaBwwA3gR+2KCDk+oppTQNOB64C/hvYIWaQduQY5Pqw5Bt2l4CngaOjYhdU0pzU0p3AWsDWzXs0KT6SSl9mlL6MqU0HvglsOL8oI2IThGxScOOUKqbz5NtwlJKMyPiTiABv89+GX0FtAPGNOjgpAqklCZExC+BSyJiKLAc8OMGHpZUJ0O2iUspTYqIG4DBVFcBM4FjUkqfN+zIpMqklMZHxFvAvsCeKaVRDT0mqS5ewlMi2USRlJ2flRqliGgF3AuckVJ6q6HHIy2JISup0YmIFVJKMxt6HNI3MWQlScqJs4slScqJIStJUk4MWUmScmLISpKUE0NWpRURcyPizYh4JyLui4iVvsW2bomIQ7PXN0bEZkvo2yUidqrgO4ZHRJtKxyipeIasymxGSmnrlNIWwCzgpJorI6Kim7WklH6RUhq8hC5dgKUOWUmNjyErVXsJ+H5WZb4UERKmIAIAAAH6SURBVI8CgyNiuYi4JCIGRsRb2S39iGpXR8R7EfFPYM35G4qI5yNiu+z1PhHx74j4T0T0j4j1qA7z07MqepeIaBsRD2TfMTAids4+u0ZEPB0R70bEjYBPTpIaGW+rqNLLKtZ9gaeypk7AFimljyOiOzA5pbR9RCwPvBwRTwPbABsDm1F9L+jBQJ+FttsWuAHYNdtW65TSxIj4B/BlSunSrN9dwOUppQER8V9AP2BToCcwIKV0fkTsD5yQ6x+EpGXOkFWZrRgRb2avXwJuovow7msppY+z9r2ALeefbwVWo/oZvbsCd2ePW/s0Ip5dzPY7Ay/O31ZKaWId49gD2KzGI35XjYhVsu84JPvs4xExqcL9lNRADFmV2YyU0tY1G7Kgm1azCeiRUuq3UL/9luE4qoDOC98msEboSmqkPCcrLVk/4OSIaA4QERtFxMrAi8Dh2Tnb9iz+cWuvALtGxPrZZ1tn7VOBljX6PQ30mP8mIuYH/4vAUVnbvkCrZbZXkgphyEpLdiPV51v/HRHvANdTfQToIeCDbN1twL8W/mBKaRzQHXgwIv4D9M1W/S9w8PyJT8BvgO2yiVWD+XqW83lUh/S7VB82/iSnfZSUEx8QIElSTqxkJUnKiSErSVJODFlJknJiyEqSlBNDVpKknBiykiTlxJCVJCkn/x+wGAB/Q2/LsAAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 576x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "XgyRENmrd3SR"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "####**Stacking together**"
      ],
      "metadata": {
        "id": "RAckixhoGBdH"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#tuned models\n",
        "functional = RandomForestClassifier(max_depth=70, max_features='log2', min_samples_split=7,  n_estimators=180, random_state=42, warm_start=True)\n",
        "non_functional = RandomForestClassifier(max_depth=30, max_features='sqrt', min_samples_split=5, n_estimators=180, random_state=42, warm_start=True)\n",
        "repair = RandomForestClassifier(max_depth=20, max_features='sqrt', min_samples_split=9,    n_estimators=125, random_state=42, warm_start=True)\n"
      ],
      "metadata": {
        "id": "hjUZLkg0GACn"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#y cannot be categorical so we have to relabel it\n",
        "y_encoded = y.copy()\n",
        "y_encoded = y_encoded.replace(['non functional'], 0)\n",
        "y_encoded  = y_encoded.replace('functional needs repair', 1)\n",
        "y_encoded = y_encoded.replace('functional', 2)"
      ],
      "metadata": {
        "id": "N3erM4CwGgHH"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#create stacking classifier\n",
        "stack = StackingClassifier([('functional', functional), ('non_functional', non_functional), ('repair', repair)])\n",
        "\n",
        "# Create training and test sets: test size 0.2\n",
        "X_train, X_test, y_train, y_test= train_test_split(X, y_encoded, test_size = 0.2, random_state=42)\n",
        "\n",
        "# fitting the model\n",
        "stack.fit(X_train, y_train)\n",
        "\n",
        "# make predictions for test data\n",
        "y_pred = stack.predict(X_test)\n",
        "y_pred_train = stack.predict(X_train)\n",
        "\n",
        "\n",
        "#print best model scores on test data\n",
        "print(\"Accuracy score train: {}\".format(accuracy_score(y_train, y_pred_train)))\n",
        "print(\"Accuracy score test: {}\".format(accuracy_score(y_test, y_pred)))\n",
        "print()\n",
        "\n",
        "#print(confusion_matrix(y_test, y_pred))\n",
        "print(classification_report(y_test, y_pred))\n",
        "\n",
        "confusion_matrix_plotter(y_test, y_pred)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 561
        },
        "id": "G2eCk8hcGJ4Y",
        "outputId": "12a82821-425e-4bba-a007-0089c182218d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Accuracy score train: 0.9101412066752247\n",
            "Accuracy score test: 0.8087542087542088\n",
            "\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.84      0.78      0.81      4559\n",
            "           1       0.60      0.30      0.40       845\n",
            "           2       0.80      0.90      0.85      6476\n",
            "\n",
            "    accuracy                           0.81     11880\n",
            "   macro avg       0.75      0.66      0.68     11880\n",
            "weighted avg       0.80      0.81      0.80     11880\n",
            "\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "fBVvZ-H3VSFx"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}