travel_insurance_claims_prediction.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZUqiI9iQ4xIA"
      },
      "source": [
        "# **Travel Insurance Claims Prediction**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "Hi there! this notebook shows the analysis of Team Ceria to create a classification model of travel insurance claims using resampling methods, XGBoost, and Bayesian Optimization. The final result of the best XGBoost model will also be interpreted using feature importance and SHAP."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "id": "iD3OJa6C4xID",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [],
      "source": [
        "import pandas as pd\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import xgboost as xgb\n",
        "import shap\n",
        "\n",
        "from pandas_profiling import ProfileReport\n",
        "from sklearn.impute import KNNImputer\n",
        "from sklearn.model_selection import train_test_split, cross_val_score\n",
        "from sklearn.metrics import accuracy_score, confusion_matrix, classification_report, roc_auc_score\n",
        "from imblearn.over_sampling import *\n",
        "from imblearn.under_sampling import *\n",
        "from imblearn.combine import *\n",
        "from bayes_opt import BayesianOptimization\n",
        "from xgboost import plot_importance"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WKMeboR84xIE"
      },
      "source": [
        "### **Load Travel Insurance Dataset**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "id": "szioXL1y4xIF",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [],
      "source": [
        "# Original dataset was provided by MASA committee and uploaded to the team's Github repo\n",
        "travel_df = pd.read_csv('https://raw.githubusercontent.com/matthewfarant/masa_hackathon_2022/main/travel_insurance_dataset.csv?token=GHSAT0AAAAAABWYHOFP5HGVUC7CKTPSKBNUYXYJHAQ')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 250
        },
        "id": "KG04o64h4xIF",
        "outputId": "81b5e69d-f8e8-47e9-e5bd-3178d450459a",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [
        {
          "data": {
            "text/html": [
              "\n",
              "  <div id=\"df-095a6cac-e21a-40a9-aeaa-658b2ef3e03f\">\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>Agency</th>\n",
              "      <th>Agency Type</th>\n",
              "      <th>Distribution Channel</th>\n",
              "      <th>Product Name</th>\n",
              "      <th>Claim</th>\n",
              "      <th>Duration</th>\n",
              "      <th>Destination</th>\n",
              "      <th>Net Sales</th>\n",
              "      <th>Commision (in value)</th>\n",
              "      <th>Gender</th>\n",
              "      <th>Age</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>CBH</td>\n",
              "      <td>Travel Agency</td>\n",
              "      <td>Offline</td>\n",
              "      <td>Comprehensive Plan</td>\n",
              "      <td>No</td>\n",
              "      <td>186</td>\n",
              "      <td>MALAYSIA</td>\n",
              "      <td>-29.0</td>\n",
              "      <td>9.57</td>\n",
              "      <td>F</td>\n",
              "      <td>81</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>CBH</td>\n",
              "      <td>Travel Agency</td>\n",
              "      <td>Offline</td>\n",
              "      <td>Comprehensive Plan</td>\n",
              "      <td>No</td>\n",
              "      <td>186</td>\n",
              "      <td>MALAYSIA</td>\n",
              "      <td>-29.0</td>\n",
              "      <td>9.57</td>\n",
              "      <td>F</td>\n",
              "      <td>71</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>CWT</td>\n",
              "      <td>Travel Agency</td>\n",
              "      <td>Online</td>\n",
              "      <td>Rental Vehicle Excess Insurance</td>\n",
              "      <td>No</td>\n",
              "      <td>65</td>\n",
              "      <td>AUSTRALIA</td>\n",
              "      <td>-49.5</td>\n",
              "      <td>29.70</td>\n",
              "      <td>NaN</td>\n",
              "      <td>32</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>CWT</td>\n",
              "      <td>Travel Agency</td>\n",
              "      <td>Online</td>\n",
              "      <td>Rental Vehicle Excess Insurance</td>\n",
              "      <td>No</td>\n",
              "      <td>60</td>\n",
              "      <td>AUSTRALIA</td>\n",
              "      <td>-39.6</td>\n",
              "      <td>23.76</td>\n",
              "      <td>NaN</td>\n",
              "      <td>32</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>CWT</td>\n",
              "      <td>Travel Agency</td>\n",
              "      <td>Online</td>\n",
              "      <td>Rental Vehicle Excess Insurance</td>\n",
              "      <td>No</td>\n",
              "      <td>79</td>\n",
              "      <td>ITALY</td>\n",
              "      <td>-19.8</td>\n",
              "      <td>11.88</td>\n",
              "      <td>NaN</td>\n",
              "      <td>41</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-095a6cac-e21a-40a9-aeaa-658b2ef3e03f')\"\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-095a6cac-e21a-40a9-aeaa-658b2ef3e03f 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-095a6cac-e21a-40a9-aeaa-658b2ef3e03f');\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": [
              "  Agency    Agency Type Distribution Channel                     Product Name  \\\n",
              "0    CBH  Travel Agency              Offline               Comprehensive Plan   \n",
              "1    CBH  Travel Agency              Offline               Comprehensive Plan   \n",
              "2    CWT  Travel Agency               Online  Rental Vehicle Excess Insurance   \n",
              "3    CWT  Travel Agency               Online  Rental Vehicle Excess Insurance   \n",
              "4    CWT  Travel Agency               Online  Rental Vehicle Excess Insurance   \n",
              "\n",
              "  Claim  Duration Destination  Net Sales  Commision (in value) Gender  Age  \n",
              "0    No       186    MALAYSIA      -29.0                  9.57      F   81  \n",
              "1    No       186    MALAYSIA      -29.0                  9.57      F   71  \n",
              "2    No        65   AUSTRALIA      -49.5                 29.70    NaN   32  \n",
              "3    No        60   AUSTRALIA      -39.6                 23.76    NaN   32  \n",
              "4    No        79       ITALY      -19.8                 11.88    NaN   41  "
            ]
          },
          "execution_count": 7,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "travel_df.head()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UXdeEy2O4xIG"
      },
      "source": [
        "### **Auto-EDA using Pandas Profiling**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "The Exploratory Data Analysis (EDA) will be conducted automatically using the Pandas Profiling library. The HTML report generated by this library contains the detailed overview of all the features, missing & duplicated values, correlation matrix, and many more. If you want to see the report, you can **download the travel_insurance_eda.html file** in this repository."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "wZ11PWBG4xIG",
        "outputId": "7ffd646c-bebf-4196-c6df-fe3373573788",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "Summarize dataset: 100%|██████████| 41/41 [00:50<00:00,  1.23s/it, Completed]                                         \n",
            "Generate report structure: 100%|██████████| 1/1 [00:09<00:00,  9.23s/it]\n",
            "Render HTML: 100%|██████████| 1/1 [00:02<00:00,  2.50s/it]\n",
            "Export report to file: 100%|██████████| 1/1 [00:00<00:00, 168.37it/s]\n"
          ]
        }
      ],
      "source": [
        "prof = ProfileReport(travel_df)\n",
        "prof.to_file(output_file='travel_insurance_eda.html')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pwggrRTV4xIG"
      },
      "source": [
        "### **Feature Engineering**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "After examining the EDA result generated by Pandas Profiling, a feature engineering process that includes data cleaning, data splitting, and handling imbalanced data will be conducted to achieve a better model performance, measured by the ROC-AUC value."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 103,
      "metadata": {
        "id": "VO-NnnfN4xIH",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [],
      "source": [
        "def feature_engineering(df):\n",
        "    # Replace abnormal age & duration values with NaN. However, more context is needed for these abnormal values.\n",
        "    df.loc[df['Age']>100,'Age'] = np.NaN\n",
        "    df.loc[(df['Duration']>1000) | (df['Duration']<0),'Duration'] = np.NaN \n",
        "    # Dummy encoding of categorical features\n",
        "    df = pd.get_dummies(df,drop_first=True)\n",
        "    # KNN imputation for missing values\n",
        "    col = df.columns\n",
        "    imputer = KNNImputer(n_neighbors=2)\n",
        "    df = pd.DataFrame(imputer.fit_transform(df))\n",
        "    df.columns = col\n",
        "    return df\n",
        "\n",
        "cleaned_df = feature_engineering(travel_df)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ovteZI114xIH"
      },
      "source": [
        "### Data Splitting"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "In this study, an 80:20 split ratio will be used."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 105,
      "metadata": {
        "id": "9eX_afY74xIH",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [],
      "source": [
        "def data_splitting(df,target_col,test_size):\n",
        "    global X, y, X_train, X_test, y_train, y_test\n",
        "    X = df.loc[:, df.columns != target_col]\n",
        "    y = df.loc[:, target_col]\n",
        "    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = test_size, random_state = 0)\n",
        "\n",
        "data_splitting(cleaned_df,'Claim_Yes',0.2)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1xuJzogz4xIH"
      },
      "source": [
        "### Imbalanced Data: Claim vs. No Claim"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "The main problem in this study is the existence of imbalanced data in the target feature. The code below shows that the number of claims is only 1% of the total number of policies. In this study, several methods of oversampling, undersampling, and combination resampling will compared and later be used to tackle this problem."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 106,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hvk5dt1Y4xII",
        "outputId": "c1a74792-4cd0-41f7-b3f8-e3fcdbf63555",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "0.0    62399\n",
              "1.0      927\n",
              "Name: Claim_Yes, dtype: int64"
            ]
          },
          "execution_count": 106,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# 0 = No Claim\n",
        "# 1 = Claim\n",
        "cleaned_df['Claim_Yes'].value_counts()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dl8hjT7H4xII"
      },
      "source": [
        "The comparison is done using the XGBoost ROC-AUC value generated by cross validation score of 5-fold CV. Notice that the resampling process is done **only on training data**. A common mistake is to do the resampling before train-test splitting which generates an almost perfect model, but with misleading results. This phenomenon is usually called \"Data Leakage\"."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "id": "4XLs-gdP8cGW",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [],
      "source": [
        "model = xgb.XGBRFClassifier(random_state=100)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Km_uJHxh8RoA",
        "outputId": "46d44e2f-2b8a-42fb-e747-b139b26264d7",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "The XGB ROC-AUC Mean Score for  SMOTE(random_state=100)  is  0.9137785571469423\n",
            "The XGB ROC-AUC Mean Score for  BorderlineSMOTE(random_state=100)  is  0.9515687439601501\n",
            "The XGB ROC-AUC Mean Score for  SVMSMOTE(random_state=100)  is  0.9415471152057411\n"
          ]
        }
      ],
      "source": [
        "# some oversampling methods\n",
        "over_methods = [\n",
        "    SMOTE(random_state=100),\n",
        "    BorderlineSMOTE(random_state=100),\n",
        "    SVMSMOTE(random_state=100)\n",
        "]\n",
        "# cross validation (oversampling)\n",
        "for methods in over_methods:\n",
        "  X_train_res, y_train_res = methods.fit_resample(X_train, y_train.ravel())\n",
        "  score = cross_val_score(model, X_train_res, y_train_res, cv=5, scoring=\"roc_auc\", n_jobs=-1)\n",
        "  print(\"The XGB ROC-AUC Mean Score for \",methods,\" is \",np.mean(score))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "jfmraNt5ENbX",
        "outputId": "1da9fa23-2c96-4fe2-f1ff-9797b1f49406",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "The XGB ROC-AUC Mean Score for  RandomUnderSampler(random_state=100)  is  0.8783143306705398\n",
            "The XGB ROC-AUC Mean Score for  NearMiss()  is  0.8139634874525942\n",
            "The XGB ROC-AUC Mean Score for  EditedNearestNeighbours()  is  0.7767297528225076\n"
          ]
        }
      ],
      "source": [
        "# some undersampling methods\n",
        "under_methods = [\n",
        "    RandomUnderSampler(random_state=100),\n",
        "    NearMiss(),\n",
        "    EditedNearestNeighbours()\n",
        "]\n",
        "# cross validation (undersampling)\n",
        "for methods in under_methods:\n",
        "  X_train_res, y_train_res = methods.fit_resample(X_train, y_train.ravel())\n",
        "  score = cross_val_score(model, X_train_res, y_train_res, cv=5, scoring=\"roc_auc\", n_jobs=-1)\n",
        "  print(\"The XGB ROC-AUC Mean Score for \",methods,\" is \",np.mean(score))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "umOSPcC44XLv",
        "outputId": "6821d8df-d243-410b-81cf-cf318610512b",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "The XGB ROC-AUC Mean Score for  SMOTEENN(random_state=100)  is  0.9317809676032731\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "A worker stopped while some jobs were given to the executor. This can be caused by a too short worker timeout or by a memory leak.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "The XGB ROC-AUC Mean Score for  SMOTETomek(random_state=100)  is  0.9123865783176053\n"
          ]
        }
      ],
      "source": [
        "# some combination methods\n",
        "comb_methods = [\n",
        "    SMOTEENN(random_state=100),\n",
        "    SMOTETomek(random_state=100)\n",
        "]\n",
        "# cross validation (combination)\n",
        "for methods in comb_methods:\n",
        "  X_train_res, y_train_res = methods.fit_resample(X_train, y_train.ravel())\n",
        "  score = cross_val_score(model, X_train_res, y_train_res, cv=5, scoring=\"roc_auc\", n_jobs=-1)\n",
        "  print(\"The XGB ROC-AUC Mean Score for \",methods,\" is \",np.mean(score))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "From the codes above, we can see which one is the best method for each resampling categories. Next, we will use these three methods to model the claims using XGBoost."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 107,
      "metadata": {
        "id": "haqRUwkJ4xII",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [],
      "source": [
        "# (Based on above examination)\n",
        "\n",
        "# Best oversampling\n",
        "bsm = BorderlineSMOTE(random_state=100)\n",
        "X_train_res_bsm, y_train_res_bsm = bsm.fit_resample(X_train, y_train.ravel())\n",
        "\n",
        "# Best undersampling\n",
        "rus = RandomUnderSampler(random_state=100)\n",
        "X_train_res_rus, y_train_res_rus = rus.fit_resample(X_train, y_train.ravel())\n",
        "\n",
        "# Best combination\n",
        "com = SMOTEENN(random_state=100)\n",
        "X_train_res_com, y_train_res_com = com.fit_resample(X_train, y_train.ravel())"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "X4IvwOI44xIJ"
      },
      "source": [
        "### **Hyperparameter Tuning using Bayesian Optimization**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "Hyperparameter tuning is an essential process in XGBoost modelling due to the numerous parameters that XGBoost has. Contrary to GridSearchCV or Random Search method that people commonly use, Bayesian Optimization doesn't pick the parameters from the hyperparameter space randomly. Instead, it will form a surrogate function to replicate the black-box function of XGBoost ROC-AUC that we want to maximize and sequentially pick the appropriate parameters until the ROC-AUC value is maximized"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 68,
      "metadata": {
        "id": "LOWan2Ct4xIK",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [],
      "source": [
        "# Hyperparameter Space\n",
        "hps = {\n",
        "  'min_child_weight':(3, 20),\n",
        "  'gamma':(0, 10),\n",
        "  'subsample':(0.5, 1),\n",
        "  'colsample_bytree':(0.1, 1),\n",
        "  'max_depth': (2, 20),\n",
        "  'learning_rate': (0, 1)\n",
        "}\n",
        "\n",
        "# Base function that will be optimized\n",
        "def hyper_xgboost(max_depth, subsample, colsample_bytree,min_child_weight, gamma, learning_rate):\n",
        "    params = {\n",
        "    'objective': 'binary:logistic',\n",
        "    'eval_metric':'auc',\n",
        "    'nthread':-1\n",
        "     }\n",
        "    params['max_depth'] = int(round(max_depth))\n",
        "    params['subsample'] = max(min(subsample, 1), 0)\n",
        "    params['colsample_bytree'] = max(min(colsample_bytree, 1), 0)\n",
        "    params['min_child_weight'] = int(min_child_weight)\n",
        "    params['gamma'] = max(gamma, 0)\n",
        "    params['learning_rate'] = learning_rate\n",
        "\n",
        "    scores = xgb.cv(params, dtrain, num_boost_round=500,verbose_eval=False,   #XGBoost requires data to be transformed into DMatrix\n",
        "                    early_stopping_rounds=10, nfold=5)\n",
        "    return scores['test-auc-mean'].iloc[-1]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "BorderlineSMOTE"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 44,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dYRF5WT-4xIK",
        "outputId": "d7ddef3f-e81b-4161-8d05-a8d7f1131ef2",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "|   iter    |  target   | colsam... |   gamma   | learni... | max_depth | min_ch... | subsample |\n",
            "-------------------------------------------------------------------------------------------------\n",
            "| \u001b[0m 1       \u001b[0m | \u001b[0m 0.9818  \u001b[0m | \u001b[0m 0.4753  \u001b[0m | \u001b[0m 7.203   \u001b[0m | \u001b[0m 0.000114\u001b[0m | \u001b[0m 7.442   \u001b[0m | \u001b[0m 5.495   \u001b[0m | \u001b[0m 0.5462  \u001b[0m |\n",
            "| \u001b[95m 2       \u001b[0m | \u001b[95m 0.9973  \u001b[0m | \u001b[95m 0.2676  \u001b[0m | \u001b[95m 3.456   \u001b[0m | \u001b[95m 0.3968  \u001b[0m | \u001b[95m 11.7    \u001b[0m | \u001b[95m 10.13   \u001b[0m | \u001b[95m 0.8426  \u001b[0m |\n",
            "| \u001b[0m 3       \u001b[0m | \u001b[0m 0.997   \u001b[0m | \u001b[0m 0.2908  \u001b[0m | \u001b[0m 3.604   \u001b[0m | \u001b[0m 0.8477  \u001b[0m | \u001b[0m 11.16   \u001b[0m | \u001b[0m 10.07   \u001b[0m | \u001b[0m 0.6585  \u001b[0m |\n",
            "| \u001b[0m 4       \u001b[0m | \u001b[0m 0.9969  \u001b[0m | \u001b[0m 0.2957  \u001b[0m | \u001b[0m 0.2535  \u001b[0m | \u001b[0m 1.0     \u001b[0m | \u001b[0m 11.53   \u001b[0m | \u001b[0m 15.65   \u001b[0m | \u001b[0m 0.5     \u001b[0m |\n",
            "| \u001b[0m 5       \u001b[0m | \u001b[0m 0.9966  \u001b[0m | \u001b[0m 0.1303  \u001b[0m | \u001b[0m 6.76    \u001b[0m | \u001b[0m 1.0     \u001b[0m | \u001b[0m 16.69   \u001b[0m | \u001b[0m 15.6    \u001b[0m | \u001b[0m 0.5     \u001b[0m |\n",
            "| \u001b[0m 6       \u001b[0m | \u001b[0m 0.9968  \u001b[0m | \u001b[0m 1.0     \u001b[0m | \u001b[0m 0.0     \u001b[0m | \u001b[0m 1.0     \u001b[0m | \u001b[0m 19.51   \u001b[0m | \u001b[0m 10.84   \u001b[0m | \u001b[0m 0.5     \u001b[0m |\n",
            "| \u001b[0m 7       \u001b[0m | \u001b[0m 0.9968  \u001b[0m | \u001b[0m 1.0     \u001b[0m | \u001b[0m 0.0     \u001b[0m | \u001b[0m 1.0     \u001b[0m | \u001b[0m 20.0    \u001b[0m | \u001b[0m 20.0    \u001b[0m | \u001b[0m 0.5052  \u001b[0m |\n",
            "=================================================================================================\n"
          ]
        }
      ],
      "source": [
        "# BorderlineSMOTE\n",
        "dtrain = xgb.DMatrix(X_train_res_bsm, y_train_res_bsm, feature_names=X_train_res_bsm.columns.values)\n",
        "optimizer = BayesianOptimization(hyper_xgboost, hps, random_state=1)\n",
        "optimizer.maximize(init_points=2, n_iter=5)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 45,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LllDAiYH4xIK",
        "outputId": "c95ebe11-7f7e-4eae-ab9f-4345013a8678",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "{'colsample_bytree': 0.2676341902399038,\n",
              " 'gamma': 3.4556072704304777,\n",
              " 'learning_rate': 0.39676747423066994,\n",
              " 'max_depth': 11.698701212060424,\n",
              " 'min_child_weight': 10.126306744856013,\n",
              " 'subsample': 0.8426097501983798}"
            ]
          },
          "execution_count": 45,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "optimizer.max['params']"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 71,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7o5bbTfG4xIL",
        "outputId": "445fa15e-af9f-469f-b087-f2c9394a6f74",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "XGBClassifier(colsample_bytree=0.2676341902399038, eval_metric='auc',\n",
              "              gamma=3.4556072704304777, learning_rate=0.39676747423066994,\n",
              "              max_depth=12, min_child_weight=10, n_jobs=-1, nthread=-1,\n",
              "              random_state=100, subsample=0.8426097501983798)"
            ]
          },
          "execution_count": 71,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# Based on the best parameter tuning above\n",
        "params = {\n",
        "    'colsample_bytree': 0.2676341902399038,\n",
        "    'gamma': 3.4556072704304777,\n",
        "    'learning_rate': 0.39676747423066994,\n",
        "    'max_depth': 12,\n",
        "    'min_child_weight': 10,\n",
        "    'subsample': 0.8426097501983798,\n",
        "    'objective': 'binary:logistic',\n",
        "    'eval_metric':'auc',\n",
        "    'n_jobs':-1\n",
        "}\n",
        "xgbr_ov =  xgb.XGBClassifier(**params, random_state=100, nthread=-1)\n",
        "xgbr_ov.fit(X_train_res_bsm, y_train_res_bsm)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 72,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "km8WYJ3-4xIL",
        "outputId": "2ef7f324-26ae-43cd-d263-a1a29f1367c5",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Test Accuracy score 0.980050531634909\n",
            "ROC AUC Score 0.5202395988905484\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "         0.0       0.99      0.99      0.99     18748\n",
            "         1.0       0.08      0.05      0.06       250\n",
            "\n",
            "    accuracy                           0.98     18998\n",
            "   macro avg       0.53      0.52      0.52     18998\n",
            "weighted avg       0.98      0.98      0.98     18998\n",
            "\n"
          ]
        }
      ],
      "source": [
        "y_pred = xgbr_ov.predict(X_test)\n",
        "testing_accuracy = accuracy_score(y_test, y_pred)\n",
        "roc_auc = roc_auc_score(y_test,y_pred)\n",
        "print(\"Test Accuracy score {0}\".format(testing_accuracy))\n",
        "print(\"ROC AUC Score {0}\".format(roc_auc))\n",
        "print(classification_report(y_test, y_pred))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "RandomUnderSampling"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 108,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cdcVPO6cQaZl",
        "outputId": "5b7f648f-36a4-42ec-abd6-13660638deed",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "|   iter    |  target   | colsam... |   gamma   | learni... | max_depth | min_ch... | subsample |\n",
            "-------------------------------------------------------------------------------------------------\n",
            "| \u001b[0m 1       \u001b[0m | \u001b[0m 0.8199  \u001b[0m | \u001b[0m 0.4753  \u001b[0m | \u001b[0m 7.203   \u001b[0m | \u001b[0m 0.000114\u001b[0m | \u001b[0m 7.442   \u001b[0m | \u001b[0m 5.495   \u001b[0m | \u001b[0m 0.5462  \u001b[0m |\n",
            "| \u001b[0m 2       \u001b[0m | \u001b[0m 0.819   \u001b[0m | \u001b[0m 0.2676  \u001b[0m | \u001b[0m 3.456   \u001b[0m | \u001b[0m 0.3968  \u001b[0m | \u001b[0m 11.7    \u001b[0m | \u001b[0m 10.13   \u001b[0m | \u001b[0m 0.8426  \u001b[0m |\n",
            "| \u001b[95m 3       \u001b[0m | \u001b[95m 0.8211  \u001b[0m | \u001b[95m 0.4054  \u001b[0m | \u001b[95m 7.817   \u001b[0m | \u001b[95m 0.1203  \u001b[0m | \u001b[95m 7.538   \u001b[0m | \u001b[95m 5.589   \u001b[0m | \u001b[95m 0.9949  \u001b[0m |\n",
            "| \u001b[0m 4       \u001b[0m | \u001b[0m 0.806   \u001b[0m | \u001b[0m 0.6379  \u001b[0m | \u001b[0m 3.184   \u001b[0m | \u001b[0m 0.6631  \u001b[0m | \u001b[0m 10.07   \u001b[0m | \u001b[0m 5.927   \u001b[0m | \u001b[0m 0.6515  \u001b[0m |\n",
            "| \u001b[0m 5       \u001b[0m | \u001b[0m 0.8068  \u001b[0m | \u001b[0m 0.1075  \u001b[0m | \u001b[0m 9.696   \u001b[0m | \u001b[0m 0.02058 \u001b[0m | \u001b[0m 4.872   \u001b[0m | \u001b[0m 6.028   \u001b[0m | \u001b[0m 0.5944  \u001b[0m |\n",
            "| \u001b[0m 6       \u001b[0m | \u001b[0m 0.8209  \u001b[0m | \u001b[0m 0.8096  \u001b[0m | \u001b[0m 8.579   \u001b[0m | \u001b[0m 0.2866  \u001b[0m | \u001b[0m 9.447   \u001b[0m | \u001b[0m 5.336   \u001b[0m | \u001b[0m 0.5578  \u001b[0m |\n",
            "| \u001b[95m 7       \u001b[0m | \u001b[95m 0.8228  \u001b[0m | \u001b[95m 0.6531  \u001b[0m | \u001b[95m 8.305   \u001b[0m | \u001b[95m 0.1256  \u001b[0m | \u001b[95m 8.591   \u001b[0m | \u001b[95m 3.011   \u001b[0m | \u001b[95m 0.9088  \u001b[0m |\n",
            "=================================================================================================\n"
          ]
        }
      ],
      "source": [
        "# RandomUnderSampling\n",
        "dtrain = xgb.DMatrix(X_train_res_rus, y_train_res_rus, feature_names=X_train_res_rus.columns.values)\n",
        "optimizer = BayesianOptimization(hyper_xgboost, hps, random_state=1)\n",
        "optimizer.maximize(init_points=2, n_iter=5)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 109,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "RrHUyg5tUUuG",
        "outputId": "11273dd3-4741-417d-d60f-e884a9d15016",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "{'colsample_bytree': 0.6530757505740343,\n",
              " 'gamma': 8.304814324499533,\n",
              " 'learning_rate': 0.12556992929644617,\n",
              " 'max_depth': 8.590533394055313,\n",
              " 'min_child_weight': 3.0106534893767374,\n",
              " 'subsample': 0.9088292575533716}"
            ]
          },
          "execution_count": 109,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "optimizer.max['params']"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 110,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "QBlkJKQMUXgh",
        "outputId": "dc1eed91-30da-4efd-c6c3-bdf377fa54da",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "XGBClassifier(colsample_bytree=0.6530757505740343, eval_metric='auc',\n",
              "              gamma=8.304814324499533, learning_rate=0.12556992929644617,\n",
              "              max_depth=9, min_child_weight=3, n_jobs=-1, nthread=-1,\n",
              "              random_state=100, subsample=0.9088292575533716)"
            ]
          },
          "execution_count": 110,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# Based on the best parameter tuning above\n",
        "params = {\n",
        "    'colsample_bytree': 0.6530757505740343,\n",
        "    'gamma': 8.304814324499533,\n",
        "    'learning_rate': 0.12556992929644617,\n",
        "    'max_depth': 9,\n",
        "    'min_child_weight': 3,\n",
        "    'subsample': 0.9088292575533716,\n",
        "    'objective': 'binary:logistic',\n",
        "    'eval_metric':'auc',\n",
        "    'n_jobs':-1\n",
        "}\n",
        "xgbr_un =  xgb.XGBClassifier(**params, random_state=100, nthread=-1)\n",
        "xgbr_un.fit(X_train_res_rus, y_train_res_rus)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 111,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "kbSxNArGVGqm",
        "outputId": "21348b41-9865-440d-c0cd-97af330614df",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Test Accuracy score 0.7587241433759672\n",
            "ROC AUC Score 0.7463949927864344\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "         0.0       1.00      0.76      0.86     12497\n",
            "         1.0       0.04      0.73      0.08       169\n",
            "\n",
            "    accuracy                           0.76     12666\n",
            "   macro avg       0.52      0.75      0.47     12666\n",
            "weighted avg       0.98      0.76      0.85     12666\n",
            "\n"
          ]
        }
      ],
      "source": [
        "y_pred = xgbr_un.predict(X_test)\n",
        "testing_accuracy = accuracy_score(y_test, y_pred)\n",
        "roc_auc = roc_auc_score(y_test,y_pred)\n",
        "print(\"Test Accuracy score {0}\".format(testing_accuracy))\n",
        "print(\"ROC AUC Score {0}\".format(roc_auc))\n",
        "print(classification_report(y_test, y_pred))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "SMOTEENN"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 83,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "24XspE2SDtmq",
        "outputId": "9090bc09-d028-41bf-d932-13e7515c9990",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "|   iter    |  target   | colsam... |   gamma   | learni... | max_depth | min_ch... | subsample |\n",
            "-------------------------------------------------------------------------------------------------\n",
            "| \u001b[0m 1       \u001b[0m | \u001b[0m 0.9772  \u001b[0m | \u001b[0m 0.4753  \u001b[0m | \u001b[0m 7.203   \u001b[0m | \u001b[0m 0.000114\u001b[0m | \u001b[0m 7.442   \u001b[0m | \u001b[0m 5.495   \u001b[0m | \u001b[0m 0.5462  \u001b[0m |\n",
            "| \u001b[95m 2       \u001b[0m | \u001b[95m 0.9983  \u001b[0m | \u001b[95m 0.2676  \u001b[0m | \u001b[95m 3.456   \u001b[0m | \u001b[95m 0.3968  \u001b[0m | \u001b[95m 11.7    \u001b[0m | \u001b[95m 10.13   \u001b[0m | \u001b[95m 0.8426  \u001b[0m |\n",
            "| \u001b[0m 3       \u001b[0m | \u001b[0m 0.998   \u001b[0m | \u001b[0m 0.2908  \u001b[0m | \u001b[0m 3.604   \u001b[0m | \u001b[0m 0.8477  \u001b[0m | \u001b[0m 11.16   \u001b[0m | \u001b[0m 10.07   \u001b[0m | \u001b[0m 0.6585  \u001b[0m |\n",
            "| \u001b[0m 4       \u001b[0m | \u001b[0m 0.9979  \u001b[0m | \u001b[0m 0.3302  \u001b[0m | \u001b[0m 0.2223  \u001b[0m | \u001b[0m 1.0     \u001b[0m | \u001b[0m 11.33   \u001b[0m | \u001b[0m 15.57   \u001b[0m | \u001b[0m 0.5     \u001b[0m |\n",
            "| \u001b[0m 5       \u001b[0m | \u001b[0m 0.9975  \u001b[0m | \u001b[0m 0.2118  \u001b[0m | \u001b[0m 6.783   \u001b[0m | \u001b[0m 1.0     \u001b[0m | \u001b[0m 16.23   \u001b[0m | \u001b[0m 15.93   \u001b[0m | \u001b[0m 0.5     \u001b[0m |\n",
            "| \u001b[0m 6       \u001b[0m | \u001b[0m 0.9982  \u001b[0m | \u001b[0m 1.0     \u001b[0m | \u001b[0m 0.0     \u001b[0m | \u001b[0m 1.0     \u001b[0m | \u001b[0m 19.64   \u001b[0m | \u001b[0m 11.36   \u001b[0m | \u001b[0m 0.5     \u001b[0m |\n",
            "| \u001b[0m 7       \u001b[0m | \u001b[0m 0.9981  \u001b[0m | \u001b[0m 1.0     \u001b[0m | \u001b[0m 0.0     \u001b[0m | \u001b[0m 1.0     \u001b[0m | \u001b[0m 20.0    \u001b[0m | \u001b[0m 20.0    \u001b[0m | \u001b[0m 0.5025  \u001b[0m |\n",
            "=================================================================================================\n"
          ]
        }
      ],
      "source": [
        "# SMOTEENN\n",
        "dtrain = xgb.DMatrix(X_train_res_com, y_train_res_com, feature_names=X_train_res_com.columns.values)\n",
        "optimizer = BayesianOptimization(hyper_xgboost, hps, random_state=1)\n",
        "optimizer.maximize(init_points=2, n_iter=5)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 84,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "WO8V0_PND7jn",
        "outputId": "28961cbf-3aa5-4676-c737-8e886adf287d",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "{'colsample_bytree': 0.2676341902399038,\n",
              " 'gamma': 3.4556072704304777,\n",
              " 'learning_rate': 0.39676747423066994,\n",
              " 'max_depth': 11.698701212060424,\n",
              " 'min_child_weight': 10.126306744856013,\n",
              " 'subsample': 0.8426097501983798}"
            ]
          },
          "execution_count": 84,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "optimizer.max['params']"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 87,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xvqIX6edD_oL",
        "outputId": "dec36333-06ca-438d-8d18-35319b5187a8",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "XGBClassifier(colsample_bytree=0.2676341902399038, eval_metric='auc',\n",
              "              gamma=3.4556072704304777, learning_rate=0.39676747423066994,\n",
              "              max_depth=12, min_child_weight=10, n_jobs=-1, nthread=-1,\n",
              "              random_state=100, subsample=0.8426097501983798)"
            ]
          },
          "execution_count": 87,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# Based on the best parameter tuning above\n",
        "params = {\n",
        "    'colsample_bytree': 0.2676341902399038,\n",
        "    'gamma': 3.4556072704304777,\n",
        "    'learning_rate': 0.39676747423066994,\n",
        "    'max_depth': 12,\n",
        "    'min_child_weight': 10,\n",
        "    'subsample': 0.8426097501983798,\n",
        "    'objective': 'binary:logistic',\n",
        "    'eval_metric':'auc',\n",
        "    'n_jobs':-1\n",
        "}\n",
        "xgbr_com =  xgb.XGBClassifier(**params, random_state=100, nthread=-1)\n",
        "xgbr_com.fit(X_train_res_com, y_train_res_com)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 88,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "IS8J9hVjEDG4",
        "outputId": "0c52ae24-dfac-45df-da75-b1ab4f8c49df",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Test Accuracy score 0.9659964206758607\n",
            "ROC AUC Score 0.5802120759547684\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "         0.0       0.99      0.98      0.98     18748\n",
            "         1.0       0.09      0.18      0.12       250\n",
            "\n",
            "    accuracy                           0.97     18998\n",
            "   macro avg       0.54      0.58      0.55     18998\n",
            "weighted avg       0.98      0.97      0.97     18998\n",
            "\n"
          ]
        }
      ],
      "source": [
        "y_pred = xgbr_com.predict(X_test)\n",
        "testing_accuracy = accuracy_score(y_test, y_pred)\n",
        "roc_auc = roc_auc_score(y_test,y_pred)\n",
        "print(\"Test Accuracy score {0}\".format(testing_accuracy))\n",
        "print(\"ROC AUC Score {0}\".format(roc_auc))\n",
        "print(classification_report(y_test, y_pred))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aXSA8bkuFZv0"
      },
      "source": [
        "From the three results above, we can conclude that RandomUnderSampling will generate the highest ROC-AUC score among the other methods. Even though the accuracy score is not the highest, the main thing that is focused in this study is the ability of the model to predict the minority class (claims)."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 112,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "u37mKsxM4xIM",
        "outputId": "0c96031d-f49e-449d-d448-6f1ebfd7cbda",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[[9486 3011]\n",
            " [  45  124]]\n"
          ]
        }
      ],
      "source": [
        "# After re-running xgb modelling for RandomUnderSampling\n",
        "print(confusion_matrix(y_test,y_pred))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "From the confusion matrix above, we can see that the model has a 76% acurracy and 73% recall, which is a good thing since in cost-sensitive learning, a conservative model is needed. On other words, since the cost of false negative (wrongly predict someone is not claiming) is **higher** than the cost of false positive (wrongly predict someone is claiming), a high recall is a good news."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "### **Model Interpretation**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "Below is a feature importance plot of the XGBoost model. In this plot, we can see which features are the most significant predictors of whether a policyholder files a claim. It is shown that Net Sales, Duration, Age, Agency_C2B, and Product Name_Cancellation Plan are among the top 5 features."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 113,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "fuq9cW9o4xIM",
        "outputId": "ce0c7662-2f14-429d-c459-4359bb419cb9",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "# XGBoost Feature Importance\n",
        "plot_importance(xgbr_un)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "As an additional analysis to see the dynamic significance of each individual data to the prediction, a SHAP summary plot is also generated below. Some of the results below are also confirming the feature importance above."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 114,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 693
        },
        "id": "MQNe2iaoi_Pa",
        "outputId": "aba7ac01-85d5-4832-8fe9-4d8fc3065e68",
        "vscode": {
          "languageId": "python"
        }
      },
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 576x684 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "shap_values = shap.TreeExplainer(xgbr_un).shap_values(X)\n",
        "shap.summary_plot(shap_values, X)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "That's it folks! A detailed explanation of our analysis will be written in the business proposal. Thank you :smile:"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "name": "travel_insurance_claims_prediction.ipynb",
      "provenance": []
    },
    "kernelspec": {
      "display_name": "masa-kernel",
      "language": "python",
      "name": "masa-kernel"
    },
    "orig_nbformat": 4
  },
  "nbformat": 4,
  "nbformat_minor": 0
}