ajaytiwari-isb/tweedie-regression.ipynb

## tweedie-regression.ipynb
{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Tweedie Regression",
      "provenance": [],
      "collapsed_sections": [],
      "machine_shape": "hm",
      "authorship_tag": "ABX9TyMo5+IsdTjliNNj8lgIaDHH",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/gist/ajaytiwari-isb/dd78b382c44d69332574c381e016af2a/tweedie-regression.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "31kjk-Bf25C4",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 72
        },
        "outputId": "2ea8beda-7ca3-48da-e6aa-281fc74fbd38"
      },
      "source": [
        "%matplotlib inline\n",
        "import pandas as pd\n",
        "from patsy import dmatrices\n",
        "import numpy as np\n",
        "import statsmodels.api as sm\n",
        "import matplotlib as mpl\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "from statsmodels.compat import lzip\n",
        "from statsmodels.graphics.api import abline_plot\n",
        "from statsmodels.genmod.generalized_linear_model import GLM"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/statsmodels/tools/_testing.py:19: FutureWarning: pandas.util.testing is deprecated. Use the functions in the public API at pandas.testing instead.\n",
            "  import pandas.util.testing as tm\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "T9oFXygO3V4A",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Reading csv data\n",
        "df = pd.read_csv('data1.csv')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qWZoRewdaaNl",
        "colab_type": "text"
      },
      "source": [
        "Data Check"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NUEN3NORanOZ",
        "colab_type": "code",
        "outputId": "68e669a5-7a9d-40ec-81ae-95ca865c70a1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "# Rows containing duplicate data\n",
        "duplicate_rows_df = df[df.duplicated()]\n",
        "print(\"number of duplicate rows:\", duplicate_rows_df.shape)"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "number of duplicate rows: (378, 11)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UtHc0LXka3vs",
        "colab_type": "code",
        "outputId": "d8b29bd9-1869-4244-b316-e050e57d358f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 197
        }
      },
      "source": [
        "# Dropping duplicate rows\n",
        "df1 = df.drop_duplicates()\n",
        "df1.head(5)"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<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>veh_value</th>\n",
              "      <th>exposure</th>\n",
              "      <th>clm</th>\n",
              "      <th>numclaims</th>\n",
              "      <th>claimcst0</th>\n",
              "      <th>veh_body</th>\n",
              "      <th>veh_age</th>\n",
              "      <th>gender</th>\n",
              "      <th>area</th>\n",
              "      <th>agecat</th>\n",
              "      <th>X_OBSTAT_</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1.06</td>\n",
              "      <td>0.303901</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>HBACK</td>\n",
              "      <td>3</td>\n",
              "      <td>F</td>\n",
              "      <td>C</td>\n",
              "      <td>2</td>\n",
              "      <td>01101    0    0    0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1.03</td>\n",
              "      <td>0.648871</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>HBACK</td>\n",
              "      <td>2</td>\n",
              "      <td>F</td>\n",
              "      <td>A</td>\n",
              "      <td>4</td>\n",
              "      <td>01101    0    0    0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>3.26</td>\n",
              "      <td>0.569473</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>UTE</td>\n",
              "      <td>2</td>\n",
              "      <td>F</td>\n",
              "      <td>E</td>\n",
              "      <td>2</td>\n",
              "      <td>01101    0    0    0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>4.14</td>\n",
              "      <td>0.317591</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>STNWG</td>\n",
              "      <td>2</td>\n",
              "      <td>F</td>\n",
              "      <td>D</td>\n",
              "      <td>2</td>\n",
              "      <td>01101    0    0    0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0.72</td>\n",
              "      <td>0.648871</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>HBACK</td>\n",
              "      <td>4</td>\n",
              "      <td>F</td>\n",
              "      <td>C</td>\n",
              "      <td>2</td>\n",
              "      <td>01101    0    0    0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "   veh_value  exposure  clm  ...  area  agecat             X_OBSTAT_\n",
              "0       1.06  0.303901    0  ...     C       2  01101    0    0    0\n",
              "1       1.03  0.648871    0  ...     A       4  01101    0    0    0\n",
              "2       3.26  0.569473    0  ...     E       2  01101    0    0    0\n",
              "3       4.14  0.317591    0  ...     D       2  01101    0    0    0\n",
              "4       0.72  0.648871    0  ...     C       2  01101    0    0    0\n",
              "\n",
              "[5 rows x 11 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZnWa7nghA-d9",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Dropping observations with zero vehicle value\n",
        "df2 = df1[df1.veh_value !=0]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Gm4VQ5mlaeBw",
        "colab_type": "text"
      },
      "source": [
        "Exploring Data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Vf8L3XCAeFzY",
        "colab_type": "code",
        "outputId": "c5ad9037-87c6-4238-dd6b-ca6f6e327568",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 241
        }
      },
      "source": [
        "#Gender\n",
        "fig, (ax1, ax2) = plt.subplots(ncols=2)\n",
        "fig.set_size_inches(10,3)\n",
        "g = sns.countplot(x='gender',data=df2.sort_values('gender'), ax=ax1)\n",
        "title = g.set_title('Overall Data')\n",
        "g = sns.countplot(x='gender',data=df2[df2['numclaims']>0].sort_values('gender'), ax=ax2)\n",
        "title = g.set_title('Excluded Zero Values')"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x216 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vS1tJP16hh-U",
        "colab_type": "code",
        "outputId": "9aed3f07-8017-4f40-9e52-037b28a3d3c3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 348
        }
      },
      "source": [
        "# Converting vehcile value as categorical\n",
        "df2['veh_value_cat'] = df2['veh_value'].apply(lambda x:1 if x<=1 else 2 if x<=2 else 3 if x<=3 else 4 if x<=4 else 5)\n",
        "#Graphs\n",
        "fig, (ax1, ax2) = plt.subplots(ncols=2)\n",
        "fig.set_size_inches(10,3)\n",
        "g = sns.countplot(x='veh_value_cat',data=df2, ax=ax1)\n",
        "title = g.set_title('Overall Data')\n",
        "g = sns.countplot(x='veh_value_cat',data=df2[df2['numclaims']>0], ax=ax2)\n",
        "title = g.set_title('Excluded Zero Values')"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:1: SettingWithCopyWarning: \n",
            "A value is trying to be set on a copy of a slice from a DataFrame.\n",
            "Try using .loc[row_indexer,col_indexer] = value instead\n",
            "\n",
            "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
            "  \"\"\"Entry point for launching an IPython kernel.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x216 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zuQrIA0GxJo9",
        "colab_type": "code",
        "outputId": "7418685c-12aa-4dc8-a70c-57baddcdac4e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 242
        }
      },
      "source": [
        "# Vehicle Age\n",
        "fig, (ax1, ax2) = plt.subplots(ncols=2)\n",
        "fig.set_size_inches(10,3)\n",
        "g = sns.countplot(x='veh_age',data=df2,ax=ax1)\n",
        "title = g.set_title('Overall Data')\n",
        "g = sns.countplot(x='veh_age',data=df2[df2['numclaims']>0], ax=ax2)\n",
        "title = g.set_title('Excluded Zero Values')"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x216 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mxr3tJf8yaqD",
        "colab_type": "code",
        "outputId": "63472b6c-21ba-4e6c-9dd5-e817afc477e1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 241
        }
      },
      "source": [
        "# Driver Age\n",
        "fig, (ax1, ax2) = plt.subplots(ncols=2)\n",
        "fig.set_size_inches(10,3)\n",
        "g = sns.countplot(x='agecat',data=df2, ax=ax1)\n",
        "title = g.set_title('Overall Data')\n",
        "g = sns.countplot(x='agecat',data=df2[df2['numclaims']>0],ax=ax2)\n",
        "title = g.set_title('Excluded Zero Values')"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x216 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hN67mKCI7YZi",
        "colab_type": "code",
        "outputId": "d935a654-42b3-41a1-c4d8-7c12dbd09de7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 264
        }
      },
      "source": [
        "# Vehicle Body Type\n",
        "fig, (ax1, ax2) = plt.subplots(ncols=2)\n",
        "fig.set_size_inches(10,3)\n",
        "g = sns.countplot(x='veh_body',data=df2,ax=ax1)\n",
        "label = g.set_xticklabels(g.get_xticklabels(), rotation=45)\n",
        "title = g.set_title('Overall Data')\n",
        "g = sns.countplot(x='veh_body',data=df2[df2['numclaims']>0], ax=ax2)\n",
        "label = g.set_xticklabels(g.get_xticklabels(), rotation=45)\n",
        "title = g.set_title('Excluded Zero Values')"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x216 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_DZcVOIJ8U2t",
        "colab_type": "code",
        "outputId": "6bfdbd18-ab0f-4004-8d1a-2e34f8bbf7ef",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 123
        }
      },
      "source": [
        "#Grouping bottom categories as \"others\"\n",
        "series = pd.value_counts(df2.veh_body)\n",
        "mask = (series/series.sum() * 100).lt(1)\n",
        "# To replace df['column'] use np.where I.e \n",
        "df2['veh_body'] = np.where(df2['veh_body'].isin(series[mask].index),'Other',df2['veh_body'])"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:4: SettingWithCopyWarning: \n",
            "A value is trying to be set on a copy of a slice from a DataFrame.\n",
            "Try using .loc[row_indexer,col_indexer] = value instead\n",
            "\n",
            "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
            "  after removing the cwd from sys.path.\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YxJw6CQLTHfh",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Converting into string\n",
        "df3 = df2.astype({\"veh_body\":str,\"veh_age\":str,\"gender\":str,\"area\":str,\"agecat\":str,\"veh_value_cat\":str})"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "U3C-SO2BtNsH",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#Extracting subset of data having a claim\n",
        "#df4 = df3[df3.clm == 1]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5S9rsYgg3a8K",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "##from sklearn.preprocessing import MinMaxScaler\n",
        "#scaler = MinMaxScaler() \n",
        "#df_excl_0_99[['veh_value']]= scaler.fit_transform(df_excl_0_99[['veh_value']])\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0VWa8WeK3gJ6",
        "colab_type": "code",
        "outputId": "807f4302-3beb-4d9b-8398-5f0a847a7fdd",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        }
      },
      "source": [
        "# Training and test split\n",
        "mask = np.random.rand(len(df3)) < 0.8\n",
        "df_train = df3[mask]\n",
        "df_test = df3[~mask]\n",
        "print('Training data set length='+str(len(df_train)))\n",
        "print('Testing data set length='+str(len(df_test)))"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Training data set length=54175\n",
            "Testing data set length=13251\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SemoNia-9is8",
        "colab_type": "text"
      },
      "source": [
        "Statsmodels - Poisson"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5bflV40u-tRp",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Model expression\n",
        "expr = \"\"\"claimcst0 ~ veh_value_cat+veh_body+veh_age+gender+area+agecat\"\"\""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "n6AmUtGZ-wbS",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Converting data into dmatrices\n",
        "y_train, X_train = dmatrices(expr, df_train, return_type='dataframe')\n",
        "y_test, X_test = dmatrices(expr, df_test, return_type='dataframe')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "p4twDOVF8LAa",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Training model\n",
        "tweedie_model = sm.GLM(y_train, X_train, exposure = df_train.exposure, family=sm.families.Tweedie(link=None,var_power=1.6,eql=True))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fnNeC41srasR",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "tweedie_result = tweedie_model.fit()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "S01T1FM4uejO",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#tweedie_result = tweedie_model.fit_regularized(method='elastic_net', alpha=0.0, L1_wt=1, start_params=None, refit=True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "n4bqVVSW3pce",
        "colab_type": "code",
        "outputId": "c08a2395-41e4-4fb2-8adf-b8d3175e9285",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 781
        }
      },
      "source": [
        "print(tweedie_result.summary2())"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "                Results: Generalized linear model\n",
            "==================================================================\n",
            "Model:               GLM              AIC:            994762.4441 \n",
            "Link Function:       log              BIC:            2323319.5037\n",
            "Dependent Variable:  claimcst0        Log-Likelihood: -4.9735e+05 \n",
            "Date:                2020-04-20 05:27 LL-Null:        -5.0013e+05 \n",
            "No. Observations:    54175            Deviance:       2.9135e+06  \n",
            "Df Model:            27               Pearson chi2:   3.27e+08    \n",
            "Df Residuals:        54147            Scale:          6041.8      \n",
            "Method:              IRLS                                         \n",
            "------------------------------------------------------------------\n",
            "                    Coef.  Std.Err.    z    P>|z|   [0.025  0.975]\n",
            "------------------------------------------------------------------\n",
            "Intercept           7.0313   1.2736  5.5210 0.0000  4.5352  9.5274\n",
            "veh_value_cat[T.2] -0.1309   0.3892 -0.3363 0.7366 -0.8938  0.6320\n",
            "veh_value_cat[T.3] -0.2585   0.5690 -0.4543 0.6496 -1.3738  0.8568\n",
            "veh_value_cat[T.4] -0.2753   0.7496 -0.3672 0.7134 -1.7446  1.1940\n",
            "veh_value_cat[T.5]  0.1263   0.8169  0.1547 0.8771 -1.4748  1.7275\n",
            "veh_body[T.HBACK]  -0.7178   1.0858 -0.6611 0.5086 -2.8460  1.4104\n",
            "veh_body[T.HDTOP]  -0.7897   1.3125 -0.6017 0.5474 -3.3621  1.7827\n",
            "veh_body[T.MIBUS]  -0.4441   1.5986 -0.2778 0.7812 -3.5774  2.6892\n",
            "veh_body[T.Other]  -1.0597   2.3605 -0.4489 0.6535 -5.6863  3.5668\n",
            "veh_body[T.PANVN]  -0.4403   1.4956 -0.2944 0.7685 -3.3716  2.4910\n",
            "veh_body[T.SEDAN]  -0.4414   1.0684 -0.4132 0.6795 -2.5354  1.6525\n",
            "veh_body[T.STNWG]  -0.8074   1.0800 -0.7476 0.4547 -2.9241  1.3093\n",
            "veh_body[T.TRUCK]  -0.1784   1.2455 -0.1433 0.8861 -2.6195  2.2627\n",
            "veh_body[T.UTE]    -0.8898   1.1532 -0.7716 0.4403 -3.1500  1.3704\n",
            "veh_age[T.2]        0.2659   0.3939  0.6750 0.4997 -0.5061  1.0378\n",
            "veh_age[T.3]        0.1295   0.4277  0.3028 0.7621 -0.7088  0.9678\n",
            "veh_age[T.4]        0.0273   0.5436  0.0503 0.9599 -1.0381  1.0928\n",
            "gender[T.M]         0.1894   0.2547  0.7434 0.4573 -0.3099  0.6887\n",
            "area[T.B]           0.0068   0.3847  0.0178 0.9858 -0.7471  0.7608\n",
            "area[T.C]           0.3161   0.3358  0.9413 0.3465 -0.3421  0.9742\n",
            "area[T.D]           0.4001   0.4295  0.9316 0.3516 -0.4417  1.2418\n",
            "area[T.E]           0.3107   0.4852  0.6403 0.5219 -0.6403  1.2617\n",
            "area[T.F]           0.8972   0.5443  1.6484 0.0993 -0.1696  1.9641\n",
            "agecat[T.2]        -0.6394   0.4596 -1.3911 0.1642 -1.5403  0.2615\n",
            "agecat[T.3]        -0.7456   0.4481 -1.6637 0.0962 -1.6239  0.1327\n",
            "agecat[T.4]        -0.7478   0.4446 -1.6820 0.0926 -1.6192  0.1236\n",
            "agecat[T.5]        -0.9775   0.4874 -2.0057 0.0449 -1.9328 -0.0223\n",
            "agecat[T.6]        -0.9214   0.5497 -1.6761 0.0937 -1.9988  0.1561\n",
            "==================================================================\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nhBiRfphOA20",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#GLM.estimate_tweedie_power(tweedie_result.mu, method='brentq', low=1.01, high=5.0)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Tbo9mxmcDimc",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "ypred = tweedie_result.predict(X_test)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "W8jzqm_RMu7R",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "f200b962-de9f-4457-d7db-ab4764afb74f"
      },
      "source": [
        "# Calculating RMSE\n",
        "from sklearn.metrics import mean_squared_error \n",
        "rmse =np.sqrt(mean_squared_error(y_test,ypred))\n",
        "rmse"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "967.1472146811549"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 23
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PAghfX5EdjZ2",
        "colab_type": "code",
        "outputId": "d060b9d6-7a30-4ebf-e42c-4fe7af96d387",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        }
      },
      "source": [
        "model_predictions = tweedie_result.get_prediction()\n",
        "#.summary_frame() returns a pandas DataFrame\n",
        "predictions_summary_frame = model_predictions.summary_frame()\n",
        "print(predictions_summary_frame)"
      ],
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "             mean     mean_se  mean_ci_lower  mean_ci_upper\n",
            "0      398.921262  186.213028     159.791471     995.911559\n",
            "1      299.063945  134.273880     124.048399     721.002806\n",
            "2      331.403026  270.151158      67.062390    1637.698339\n",
            "3      588.030453  413.391912     148.250347    2332.404748\n",
            "4      410.555074  189.634232     166.036599    1015.170570\n",
            "...           ...         ...            ...            ...\n",
            "67850  427.675136  282.330652     117.271232    1559.683634\n",
            "67852  516.057732  267.434995     186.888577    1424.996574\n",
            "67853  357.898487  176.197092     136.365535     939.323318\n",
            "67854  454.713000  220.772978     175.574179    1177.644196\n",
            "67855  666.115473  370.900849     223.661171    1983.848255\n",
            "\n",
            "[54175 rows x 4 columns]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ThHGha15dwVq",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 294
        },
        "outputId": "925a842c-c435-4b70-c092-b9229dd8dcea"
      },
      "source": [
        "predicted_counts=predictions_summary_frame['mean']\n",
        "actual_counts = y_train['claimcst0']\n",
        "fig = plt.figure()\n",
        "fig.suptitle('Predicted versus actual')\n",
        "predicted, = plt.plot(X_train.index, predicted_counts, 'go-', label='Predicted amount')\n",
        "actual, = plt.plot(X_train.index, actual_counts, 'ro-', label='Actual amount')\n",
        "plt.legend(handles=[predicted, actual])\n",
        "plt.show()"
      ],
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "K1g_idmlvzMN",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 325
        },
        "outputId": "d60ecaef-1e63-4140-ef7f-d1f896096927"
      },
      "source": [
        "plt.clf()\n",
        "fig = plt.figure()\n",
        "fig.suptitle('Scatter plot of Actual versus Predicted amount')\n",
        "plt.scatter(x=predicted_counts, y=actual_counts, marker='.')\n",
        "plt.xlabel('Predicted Amount')\n",
        "plt.ylabel('Actual Amount')\n",
        "plt.show()"
      ],
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 0 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nRzSy-VLHZNM",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 281
        },
        "outputId": "7925722b-0428-4d54-854c-bbc468ab829f"
      },
      "source": [
        "from scipy import stats\n",
        "\n",
        "fig, ax = plt.subplots()\n",
        "\n",
        "resid = tweedie_result.resid_deviance.copy()\n",
        "resid_std = stats.zscore(resid)\n",
        "ax.hist(resid_std, bins=25)\n",
        "ax.set_title('Histogram of standardized deviance residuals');\n"
      ],
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYMAAAEICAYAAAC9E5gJAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAasUlEQVR4nO3de7hddX3n8ffHBBS5GJCUIqEGNdZBZkSNgCN1rBcIqA121IJtCZaKDtiqdarotAUvtDi1YpkqDh0YQEVEvJBaFDMg46WiBEW5aYlcJBFIJASM1Avy7R/rd2RxPJedkHP2Rt6v59nPWeu3fmut71pn7/3Z63L2SVUhSXpoe9iwC5AkDZ9hIEkyDCRJhoEkCcNAkoRhIEnCMJgRSa5O8pxh1zFMSV6S5OYkG5M8ddj1jElyfJIPbcHlnZHknW34t5J8Z0stu7eOSvKEAftuke1L8oEkf/lAlzOKpnp9JnlOktVbaD2XJPnjLbGs2WAYbKIkNyZ5/ri2I5J8aWy8qp5cVZdMs5yF7UU+d4ZKHbZ3A6+tqu2q6huDztR/c32wqaovVtVvDruOLaGqXlNV7xh2HTNhkNfnQ5Fh8CtqBELmscDVQ65hi0oyZ9g1qDMCz+9fOYbBDOgfPSTZJ8nKJHcluS3Je1q3L7SfG9qplGcmeViSv0hyU5K1Sc5K8qjecg9v025P8pfj1nN8kvOSfCjJXcARbd1fSbIhyS1J/iHJ1r3lVZKjk1yX5IdJ3pHk8Un+pdV7br//uG2csNYkD0+yEZgDfDPJdyeYN0lOavPdleTKJHslOQr4feBNbZ/8U+t/bJLvthqvSfKS3rKOSPKlJO9OckeSG5Ic1Ju+R5L/3+ZdAew8rpaPJbk1yZ1JvpDkyb1pZyQ5JckFSX4E/HaSpyb5elveR4FH9Pr/4hRDkt9r2zD2+EmSS9q0h7d6v9eeEx9Isk1vOX/efl/fT/JHE+3/Tdi+/drvc0OSb6adHmn1rRzX9w1Jlve2fez0145JPp1kXdvHn06yoDffJe258+VWx+eS7Nybvn+vhpuTHDHIfhhX2xFt+ScluR04fqr5k+zc6tyQZH2SLyZ5WJvWf91s07b1jiTXAM8Yt977naLblP0ybjlPaL+nO5P8oD13RktV+diEB3Aj8PxxbUcAX5qoD/AV4A/b8HbAfm14IVDA3N58fwSsAh7X+n4C+GCbtiewEdgf2JruNMzPeus5vo0fQhfy2wBPB/YD5rb1XQu8vre+As4HdgCeDPwEuKit/1HANcCySfbDpLX2lv2ESeY9ELgcmAcE+A/Arm3aGcA7x/V/GfCYtl2/B/yo1/+Itt2vogug/wZ8H0hv/78HeDjwbOCHwIfGbcf2bfp7gSt6084A7gSe1da9A3AT8AZgK+Clbd3vbP2fA6yeYHt3aPv+1W38JGA5sFNb9z8Bf9OmLQFuA/YCtgXOnmZfTrp9wG7A7cDBrf4XtPH5wCNb30W9ZV0GHDr+9wA8GvivbZ7tgY8Bn+rNdwnwXeCJdM+7S4AT27THtvUc1vbZo4G9p9sPE2znEcA9wJ/QPZ+3mWY//g3wgbbOrYDf6j0nbuS+182JwBfbMnYHrur/Dsfv+83YL3/chj8C/I/2e3gEsP+w38t+aR8Pu4AH26M9kTYCG3qPu5k8DL4AvA3YedxyFvLLYXARcHRv/Dfp3mzmAn8FfKQ37ZHAT7l/GHxhmtpfD3yyN17As3rjlwNv7o3/HfDeSZY1aa29ZU/2BvZc4F/pguph46b94sU2xXZcASxtw0cAq8btlwJ+HfgNujeQbXvTz6YXBuOWO6/N+6heLWf1pj+bXtC0tn9hijBoL/5PA6e08dCF2eN7fZ4J3NCGT6e9kbbxJ062L6fbPuDN9AK6tV1IC3jgQ8BfteFFdG/aj5zu9wDsDdzRG78E+Ive+NHAZ9vwW/rPuV6fKffDBP2PAL436PzA2+k+6Ey0327kvtfN9cCS3rSjGDAMBtwvY2FwFnAqsGCq5/YwH54m2jyHVNW8sQfdk38yR9K9oL+d5LIkL5qi72PoPnmOuYkuCHZp024em1BVd9N9yuu7uT+S5Int0PXWdKeO/ppxpxHoPoWO+bcJxrfbjFqnVFUXA/8AvA9Ym+TUJDtM1j/d6bEr2iH/BrpPzf3tuLW37Lvb4Hatxjuq6kfj6hxb7pwkJ6Y7BXUX3ZsE45bd36ePAdZUe3WPX94kTqD71PinbXzsU/nlve35bGsfW0d/nVMtf8rto/tU/rKx9bR17Q/s2qafTfeJHeAVdJ9q72acJI9M8r/TnRK8i+4Dzrzc/xrKrb3hu7nvebM73VHDeNPth4n098t08/8t3ZHr55Jcn+TYSZa5Kfv7fgbcL2PeRBdgX0t3N9OUp/+GwTCYYVV1XVUdBvwa8C7gvCTb0n3iGO/7dC/gMWOf/G4DbgH652m3oTtMvd/qxo2fAnyb7lTADsBb6Z6QW8JUtU6rqk6uqqfTnf56IvDnY5P6/ZI8FvhH4LXAo1v4XsVg23ELsGPb3/06x7wCWAo8n+602MKx1fZLHbe83ZL0p/eXdz9JDqV7s31pVf2sNf+ALmSf3PtA8aiqGnvzvIXuDXTa5Q+wfTfTHRnM6z22raoT2/QVwPwke7c6z55kPW+kO/Lbtz2Pnj22iVPU1q/h8RO0T7cfJtL/XUw5f1X9sKreWFWPA34H+LMkz5tgmdPt77vpQmfMr/eGB94vVXVrVb2qqh4DvBp4fwa8XXi2GAYzLMkfJJlfVffSnVICuBdY134+rtf9I8Ab0l0U3I7uk/xHq+oe4DzgxUn+c7qLuscz/Ytxe+AuYGOSJ9GdT99Spqp1SkmekWTfJFvRHer/mG5fQBcm/X0yFpzr2ryvpDsymFZV3QSsBN6WZOsk+wMv7nXZnu46ye10L/i/nmaRX6ELvD9NslWS3wX2mWQbnwr8L7qjyHW9mu6lC7eTkvxa67tbkgNbl3PpLv7vmeSRwHEPYPs+RPecObAdBT0i3UXuBW3+n9Gd5/5bunPmKyZZ1fZ0b7wbkuw0VU0T+DDw/CQvTzI3yaOT7D3AfpjSdPMneVG7aBu66z4/577nWN+5wFvaxeAFdNck+q4AXtH23xLgv/SmDbxfkrysd3H5Drrn9ET1DI1hMPOWAFenu8Pm7+ku0P1bOxw/AfhyO8zdj+588QfpDjdvoHuT/BOAqrq6DZ9D92lmI7CW7s1sMv+d7tPvD+leOFvyDoZJax3ADq2eO+gOy2+ne0MCOA3Ys+2TT1XVNXTXLr5CFxT/EfjyJtT5CmBfYD3di/Ws3rSz2vrX0F0sv3SqBVXVT4HfpTt/vZ7uYvYnJum+FNgR+FLuu6PoM23am+lOYVzaTi/8P7pPmFTVZ+guZF/c+ly8udtXVTe3Ot5KF6Y30x2B9V/3Z9MdGX1siiB/L90F2x/Q7aPPTlPTL1TV9+guYL+x1XgF8JQ2edL9MKCp5l/UxjfSPXfeX1Wfn2AZb6N7DtwAfI7uOd33OrqA3UB3p9unetM2Zb88A/hqex9YDryuqq4fbDNnx9jVdT3ItE/jG+hOAd0w7HokPbh5ZPAgkuTF7aLVtnS3ll7JfRc9JWmzGQYPLkvpLtx+n+4w+NDy0E7SFuBpIkmSRwaSpO6PhB6Udt5551q4cOGwy5CkB43LL7/8B1U14R/2PWjDYOHChaxcuXL6jpIkAJJM+hfWniaSJBkGkiTDQJKEYSBJwjCQJGEYSJIwDCRJDBgG6f6B9JXtv02tbG07JVmR7p+pr0iyY2tPkpOTrEryrSRP6y1nWet/XZJlvfant+WvavNuqX/AIkkawKYcGfx2Ve1dVYvb+LHARVW1iO7/4Y79W7mD6L5EbRHd/xM9BbrwoPu+9X3p/iHIcWMB0vq8qjffks3eIknSJnsgf4G8lO4fgAOcSffPn9/c2s9q36Z5aZJ5SXZtfVdU1XqAJCuAJUkuAXaoqktb+1nAIcBnmCELj/3nTep/44kvnKFKJGk0DHpkUHT/WPryJEe1tl2q6pY2fCv3/SP03bj/P5he3dqmal89QfsvSXJUkpVJVq5bt26iLpKkzTDokcH+VbWm/a/RFUm+3Z9YVZVkxr8Lu6pOBU4FWLx4sd+9LUlbyEBHBlW1pv1cC3yS7pz/be30D+3n2tZ9DbB7b/YFrW2q9gUTtEuSZsm0YZBk2yTbjw0DBwBX0f1T57E7gpYB57fh5cDh7a6i/YA72+mkC4EDkuzYLhwfAFzYpt2VZL92F9HhvWVJkmbBIKeJdgE+2e72nAucXVWfTXIZcG6SI4GbgJe3/hcABwOrgLuBVwJU1fok7wAua/3ePnYxGTgaOAPYhu7C8YxdPJYk/bJpw6CqrgeeMkH77cDzJmgv4JhJlnU6cPoE7SuBvQaoV5I0A/wLZEmSYSBJMgwkSRgGkiQMA0kShoEkCcNAkoRhIEnCMJAkYRhIkjAMJEkYBpIkDANJEoaBJAnDQJKEYSBJwjCQJGEYSJIwDCRJGAaSJAwDSRKGgSQJw0CShGEgScIwkCRhGEiSMAwkSRgGkiQMA0kShoEkCcNAkoRhIEliE8IgyZwk30jy6Ta+R5KvJlmV5KNJtm7tD2/jq9r0hb1lvKW1fyfJgb32Ja1tVZJjt9zmSZIGsSlHBq8Dru2Nvws4qaqeANwBHNnajwTuaO0ntX4k2RM4FHgysAR4fwuYOcD7gIOAPYHDWl9J0iwZKAySLABeCPyfNh7gucB5rcuZwCFteGkbp01/Xuu/FDinqn5SVTcAq4B92mNVVV1fVT8Fzml9JUmzZNAjg/cCbwLubeOPBjZU1T1tfDWwWxveDbgZoE2/s/X/Rfu4eSZr/yVJjkqyMsnKdevWDVi6JGk604ZBkhcBa6vq8lmoZ0pVdWpVLa6qxfPnzx92OZL0K2PuAH2eBfxOkoOBRwA7AH8PzEsyt336XwCsaf3XALsDq5PMBR4F3N5rH9OfZ7J2SdIsmPbIoKreUlULqmoh3QXgi6vq94HPAy9t3ZYB57fh5W2cNv3iqqrWfmi722gPYBHwNeAyYFG7O2nrto7lW2TrJEkDGeTIYDJvBs5J8k7gG8Bprf004INJVgHr6d7cqaqrk5wLXAPcAxxTVT8HSPJa4EJgDnB6VV39AOqSJG2iTQqDqroEuKQNX093J9D4Pj8GXjbJ/CcAJ0zQfgFwwabUIknacvwLZEmSYSBJMgwkSRgGkiQMA0kShoEkCcNAkoRhIEnCMJAkYRhIkjAMJEkYBpIkDANJEoaBJAnDQJKEYSBJwjCQJGEYSJIwDCRJGAaSJAwDSRKGgSQJw0CShGEgScIwkCRhGEiSMAwkSRgGkiQMA0kShoEkCcNAkoRhIEligDBI8ogkX0vyzSRXJ3lba98jyVeTrEry0SRbt/aHt/FVbfrC3rLe0tq/k+TAXvuS1rYqybFbfjMlSVMZ5MjgJ8Bzq+opwN7AkiT7Ae8CTqqqJwB3AEe2/kcCd7T2k1o/kuwJHAo8GVgCvD/JnCRzgPcBBwF7Aoe1vpKkWTJtGFRnYxvdqj0KeC5wXms/EzikDS9t47Tpz0uS1n5OVf2kqm4AVgH7tMeqqrq+qn4KnNP6SpJmyUDXDNon+CuAtcAK4LvAhqq6p3VZDezWhncDbgZo0+8EHt1vHzfPZO0T1XFUkpVJVq5bt26Q0iVJAxgoDKrq51W1N7CA7pP8k2a0qsnrOLWqFlfV4vnz5w+jBEn6lbRJdxNV1Qbg88AzgXlJ5rZJC4A1bXgNsDtAm/4o4PZ++7h5JmuXJM2SQe4mmp9kXhveBngBcC1dKLy0dVsGnN+Gl7dx2vSLq6pa+6HtbqM9gEXA14DLgEXt7qSt6S4yL98SGydJGszc6buwK3Bmu+vnYcC5VfXpJNcA5yR5J/AN4LTW/zTgg0lWAevp3typqquTnAtcA9wDHFNVPwdI8lrgQmAOcHpVXb3FtlCSNK1pw6CqvgU8dYL26+muH4xv/zHwskmWdQJwwgTtFwAXDFCvJGkG+BfIkiTDQJJkGEiSMAwkSRgGkiQMA0kShoEkCcNAkoRhIEnCMJAkYRhIkjAMJEkYBpIkDANJEoaBJAnDQJKEYSBJwjCQJGEYSJIwDCRJGAaSJAwDSRKGgSQJw0CShGEgScIwkCRhGEiSMAwkSRgGkiQMA0kShoEkCcNAksQAYZBk9ySfT3JNkquTvK6175RkRZLr2s8dW3uSnJxkVZJvJXlab1nLWv/rkizrtT89yZVtnpOTZCY2VpI0sUGODO4B3lhVewL7Acck2RM4FrioqhYBF7VxgIOARe1xFHAKdOEBHAfsC+wDHDcWIK3Pq3rzLXngmyZJGtS0YVBVt1TV19vwD4Frgd2ApcCZrduZwCFteClwVnUuBeYl2RU4EFhRVeur6g5gBbCkTduhqi6tqgLO6i1LkjQLNumaQZKFwFOBrwK7VNUtbdKtwC5teDfg5t5sq1vbVO2rJ2ifaP1HJVmZZOW6des2pXRJ0hQGDoMk2wEfB15fVXf1p7VP9LWFa/slVXVqVS2uqsXz58+f6dVJ0kPGQGGQZCu6IPhwVX2iNd/WTvHQfq5t7WuA3XuzL2htU7UvmKBdkjRLBrmbKMBpwLVV9Z7epOXA2B1By4Dze+2Ht7uK9gPubKeTLgQOSLJju3B8AHBhm3ZXkv3aug7vLUuSNAvmDtDnWcAfAlcmuaK1vRU4ETg3yZHATcDL27QLgIOBVcDdwCsBqmp9kncAl7V+b6+q9W34aOAMYBvgM+0hSZol04ZBVX0JmOy+/+dN0L+AYyZZ1unA6RO0rwT2mq4WSdLM8C+QJUmGgSTJMJAkYRhIkjAMJEkYBpIkDANJEoaBJAnDQJKEYSBJwjCQJGEYSJIwDCRJGAaSJAwDSRKGgSQJw0CShGEgScIwkCRhGEiSMAwkSRgGkiQMA0kShoEkCcNAkoRhIEnCMJAkYRhIkjAMJEkYBpIkDANJEoaBJIkBwiDJ6UnWJrmq17ZTkhVJrms/d2ztSXJyklVJvpXkab15lrX+1yVZ1mt/epIr2zwnJ8mW3khJ0tQGOTI4A1gyru1Y4KKqWgRc1MYBDgIWtcdRwCnQhQdwHLAvsA9w3FiAtD6v6s03fl2SpBk2bRhU1ReA9eOalwJntuEzgUN67WdV51JgXpJdgQOBFVW1vqruAFYAS9q0Harq0qoq4KzesiRJs2RzrxnsUlW3tOFbgV3a8G7Azb1+q1vbVO2rJ2iXJM2iB3wBuX2iry1Qy7SSHJVkZZKV69atm41VStJDwuaGwW3tFA/t59rWvgbYvddvQWubqn3BBO0TqqpTq2pxVS2eP3/+ZpYuSRpvc8NgOTB2R9Ay4Pxe++HtrqL9gDvb6aQLgQOS7NguHB8AXNim3ZVkv3YX0eG9ZUmSZsnc6Tok+QjwHGDnJKvp7go6ETg3yZHATcDLW/cLgIOBVcDdwCsBqmp9kncAl7V+b6+qsYvSR9PdsbQN8Jn2kCTNomnDoKoOm2TS8yboW8AxkyzndOD0CdpXAntNV4ckaeb4F8iSJMNAkmQYSJIwDCRJGAaSJAwDSRKGgSQJw0CShGEgScIwkCRhGEiSMAwkSRgGkiQMA0kShoEkCcNAkoRhIEnCMJAkYRhIkjAMJEkYBpIkDANJEoaBJAnDQJKEYSBJwjCQJGEYSJKAucMu4MFg4bH/vEn9bzzxhTNUiSTNDI8MJEmGgSTJMJAkYRhIkvAC8ozwgrOkB5uROTJIsiTJd5KsSnLssOuRpIeSkTgySDIHeB/wAmA1cFmS5VV1zXArmx0eSUgatpEIA2AfYFVVXQ+Q5BxgKfCQCINNtanhsTkMHOmhZVTCYDfg5t74amDf8Z2SHAUc1UY3JvnODNWzM/CDGVr2ljDj9eVdmz3rQ37fPQCjXBuMdn2jXBuMTn2PnWzCqITBQKrqVODUmV5PkpVVtXim17O5Rrm+Ua4NRru+Ua4NRru+Ua4NRr8+GJ0LyGuA3XvjC1qbJGkWjEoYXAYsSrJHkq2BQ4HlQ65Jkh4yRuI0UVXdk+S1wIXAHOD0qrp6iCXN+KmoB2iU6xvl2mC06xvl2mC06xvl2mD06yNVNewaJElDNiqniSRJQ2QYSJIMg/FG+WsxktyY5MokVyRZOQL1nJ5kbZKrem07JVmR5Lr2c8cRqu34JGva/rsiycHDqK3VsnuSzye5JsnVSV7X2oe+/6aobST2X5JHJPlakm+2+t7W2vdI8tX22v1ouxllVGo7I8kNvX2392zXNh2vGfS0r8X4V3pfiwEcNipfi5HkRmBxVY3CH6+Q5NnARuCsqtqrtf1PYH1VndjCdMeqevOI1HY8sLGq3j3b9YyXZFdg16r6epLtgcuBQ4AjGPL+m6K2lzMC+y9JgG2ramOSrYAvAa8D/gz4RFWdk+QDwDer6pQRqe01wKer6rzZrGdTeGRwf7/4Woyq+ikw9rUYmkBVfQFYP655KXBmGz6T7k1k1k1S28ioqluq6utt+IfAtXR/iT/0/TdFbSOhOhvb6FbtUcBzgbE322Htu8lqG3mGwf1N9LUYI/MioHtSfS7J5e2rOUbRLlV1Sxu+FdhlmMVM4LVJvtVOIw3lFNZ4SRYCTwW+yojtv3G1wYjsvyRzklwBrAVWAN8FNlTVPa3L0F6742urqrF9d0LbdyclefgwapuKYfDgsn9VPQ04CDimnQoZWdWdgxylT0WnAI8H9gZuAf5uuOVAku2AjwOvr6q7+tOGvf8mqG1k9l9V/byq9qb7toJ9gCcNq5bxxteWZC/gLXQ1PgPYCZj1U6fTMQzub6S/FqOq1rSfa4FP0r0IRs1t7Zzz2LnntUOu5xeq6rb2Qr0X+EeGvP/aOeWPAx+uqk+05pHYfxPVNmr7r9W0Afg88ExgXpKxP6Qd+mu3V9uSduqtquonwP9lBPbdeIbB/Y3s12Ik2bZdzCPJtsABwFVTzzUUy4FlbXgZcP4Qa7mfsTfZ5iUMcf+1C42nAddW1Xt6k4a+/yarbVT2X5L5Sea14W3obvi4lu6N96Wt27D23US1fbsX8KG7ljFyr13vJhqn3S73Xu77WowThlwSAEkeR3c0AN3XiJw97NqSfAR4Dt3X894GHAd8CjgX+A3gJuDlVTXrF3Inqe05dKc4CrgReHXv/Pxs17c/8EXgSuDe1vxWunPzQ91/U9R2GCOw/5L8J7oLxHPoPtCeW1Vvb6+Rc+hOw3wD+IP2SXwUarsYmA8EuAJ4Te9C80gwDCRJniaSJBkGkiQMA0kShoEkCcNAkoRhIEnCMJAkAf8O6BPbGyuVutMAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XsTXvXKwHeRx",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 541
        },
        "outputId": "2362ccd4-6a5e-45dc-968f-0596825c0f12"
      },
      "source": [
        "from statsmodels import graphics\n",
        "graphics.gofplots.qqplot(resid, line='r')"
      ],
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 28
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AiUPZRxZnBQ_",
        "colab_type": "text"
      },
      "source": [
        "XGBOOST"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "s5lMVOXCnmXi",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "df4=df3.drop([\"veh_value\",\"clm\",\"numclaims\",\"X_OBSTAT_\"], axis = 1) "
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "d5H_2Fa-nCoK",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 217
        },
        "outputId": "71b3a802-45fb-4e24-ca43-e21500e48809"
      },
      "source": [
        "data = pd.get_dummies(df4, prefix_sep=\"_\",columns =[\"veh_value_cat\",\"veh_body\",\"veh_age\",\"gender\",\"area\",\"agecat\"],\n",
        "drop_first=True)\n",
        "data.head()"
      ],
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<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>exposure</th>\n",
              "      <th>claimcst0</th>\n",
              "      <th>veh_value_cat_2</th>\n",
              "      <th>veh_value_cat_3</th>\n",
              "      <th>veh_value_cat_4</th>\n",
              "      <th>veh_value_cat_5</th>\n",
              "      <th>veh_body_HBACK</th>\n",
              "      <th>veh_body_HDTOP</th>\n",
              "      <th>veh_body_MIBUS</th>\n",
              "      <th>veh_body_Other</th>\n",
              "      <th>veh_body_PANVN</th>\n",
              "      <th>veh_body_SEDAN</th>\n",
              "      <th>veh_body_STNWG</th>\n",
              "      <th>veh_body_TRUCK</th>\n",
              "      <th>veh_body_UTE</th>\n",
              "      <th>veh_age_2</th>\n",
              "      <th>veh_age_3</th>\n",
              "      <th>veh_age_4</th>\n",
              "      <th>gender_M</th>\n",
              "      <th>area_B</th>\n",
              "      <th>area_C</th>\n",
              "      <th>area_D</th>\n",
              "      <th>area_E</th>\n",
              "      <th>area_F</th>\n",
              "      <th>agecat_2</th>\n",
              "      <th>agecat_3</th>\n",
              "      <th>agecat_4</th>\n",
              "      <th>agecat_5</th>\n",
              "      <th>agecat_6</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0.303901</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0.648871</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0.569473</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0.317591</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0.648871</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "   exposure  claimcst0  veh_value_cat_2  ...  agecat_4  agecat_5  agecat_6\n",
              "0  0.303901        0.0                1  ...         0         0         0\n",
              "1  0.648871        0.0                1  ...         1         0         0\n",
              "2  0.569473        0.0                0  ...         0         0         0\n",
              "3  0.317591        0.0                0  ...         0         0         0\n",
              "4  0.648871        0.0                0  ...         0         0         0\n",
              "\n",
              "[5 rows x 29 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 30
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6ODy9kExoPA1",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "test_df = data.pop('claimcst0')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fVeKzChHofKD",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 217
        },
        "outputId": "e3d1f524-151f-4b8f-8364-50a83187c36b"
      },
      "source": [
        "data['claimcst0']=test_df\n",
        "data.head()"
      ],
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<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>exposure</th>\n",
              "      <th>veh_value_cat_2</th>\n",
              "      <th>veh_value_cat_3</th>\n",
              "      <th>veh_value_cat_4</th>\n",
              "      <th>veh_value_cat_5</th>\n",
              "      <th>veh_body_HBACK</th>\n",
              "      <th>veh_body_HDTOP</th>\n",
              "      <th>veh_body_MIBUS</th>\n",
              "      <th>veh_body_Other</th>\n",
              "      <th>veh_body_PANVN</th>\n",
              "      <th>veh_body_SEDAN</th>\n",
              "      <th>veh_body_STNWG</th>\n",
              "      <th>veh_body_TRUCK</th>\n",
              "      <th>veh_body_UTE</th>\n",
              "      <th>veh_age_2</th>\n",
              "      <th>veh_age_3</th>\n",
              "      <th>veh_age_4</th>\n",
              "      <th>gender_M</th>\n",
              "      <th>area_B</th>\n",
              "      <th>area_C</th>\n",
              "      <th>area_D</th>\n",
              "      <th>area_E</th>\n",
              "      <th>area_F</th>\n",
              "      <th>agecat_2</th>\n",
              "      <th>agecat_3</th>\n",
              "      <th>agecat_4</th>\n",
              "      <th>agecat_5</th>\n",
              "      <th>agecat_6</th>\n",
              "      <th>claimcst0</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0.303901</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0.648871</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0.569473</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0.317591</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0.648871</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "   exposure  veh_value_cat_2  veh_value_cat_3  ...  agecat_5  agecat_6  claimcst0\n",
              "0  0.303901                1                0  ...         0         0        0.0\n",
              "1  0.648871                1                0  ...         0         0        0.0\n",
              "2  0.569473                0                0  ...         0         0        0.0\n",
              "3  0.317591                0                0  ...         0         0        0.0\n",
              "4  0.648871                0                0  ...         0         0        0.0\n",
              "\n",
              "[5 rows x 29 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 32
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9j2lx5RzoxEI",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import xgboost as xgb\n",
        "from sklearn.metrics import mean_squared_error\n",
        "from sklearn.model_selection import train_test_split"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9RFwjDEeosb3",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X, y = data.iloc[:,:-1],data.iloc[:,-1]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AL8NjXYEDPwg",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X_train, X_test, y_train, y_test = train_test_split(X,y,test_size=.2, random_state=123)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5rbYiNs6C5WO",
        "colab_type": "text"
      },
      "source": [
        ""
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vBTMh_Aiom-k",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "dtrain = xgb.DMatrix(data=X_train.iloc[:,1:28],label=y_train)\n",
        "dtest = xgb.DMatrix(data=X_test.iloc[:,1:28],label=y_test)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pafHwOSnN3NY",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "dtrain.set_base_margin(np.log(X_train['exposure']))\n",
        "dtest.set_base_margin(np.log(X_test['exposure']))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HOnwm0rug7iK",
        "colab_type": "text"
      },
      "source": [
        "Modeling"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Enq5Vn8nZRSl",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "params = {\"objective\":\"reg:tweedie\",'colsample_bytree': 1.0,'learning_rate': 0.01,\n",
        "                'gamma':1.5,'max_depth': 2, 'subsample':0.6,'reg_alpha': 0, 'reg_lambda':1,'min_child_weight':5,'n_estimators':2000,\n",
        "          \"tweedie_variance_power\":1.6}\n",
        "\n",
        "cv_results = xgb.cv(dtrain=dtrain, params=params, nfold=3,\n",
        "                    num_boost_round=2000,early_stopping_rounds=10,metrics=\"rmse\", as_pandas=True, seed=123)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RzihpdYaZYLJ",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 197
        },
        "outputId": "bdbd80e5-0ca9-461b-f7b6-7447f0256b45"
      },
      "source": [
        "cv_results.head()"
      ],
      "execution_count": 39,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<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>train-rmse-mean</th>\n",
              "      <th>train-rmse-std</th>\n",
              "      <th>test-rmse-mean</th>\n",
              "      <th>test-rmse-std</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1041.407247</td>\n",
              "      <td>44.692083</td>\n",
              "      <td>1038.612793</td>\n",
              "      <td>88.373932</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1041.406006</td>\n",
              "      <td>44.692110</td>\n",
              "      <td>1038.611511</td>\n",
              "      <td>88.373961</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1041.404745</td>\n",
              "      <td>44.692163</td>\n",
              "      <td>1038.610291</td>\n",
              "      <td>88.374060</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1041.403402</td>\n",
              "      <td>44.692163</td>\n",
              "      <td>1038.608928</td>\n",
              "      <td>88.374141</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>1041.402120</td>\n",
              "      <td>44.692242</td>\n",
              "      <td>1038.607625</td>\n",
              "      <td>88.374193</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "   train-rmse-mean  train-rmse-std  test-rmse-mean  test-rmse-std\n",
              "0      1041.407247       44.692083     1038.612793      88.373932\n",
              "1      1041.406006       44.692110     1038.611511      88.373961\n",
              "2      1041.404745       44.692163     1038.610291      88.374060\n",
              "3      1041.403402       44.692163     1038.608928      88.374141\n",
              "4      1041.402120       44.692242     1038.607625      88.374193"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 39
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OtARy3JPZcbe",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        },
        "outputId": "04c640f3-1ca9-4b2a-f3d7-a24ae96ee204"
      },
      "source": [
        "print((cv_results[\"test-rmse-mean\"]).tail(1))"
      ],
      "execution_count": 40,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "437    1029.123637\n",
            "Name: test-rmse-mean, dtype: float64\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pdD27M55ZhGV",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "xg_reg = xgb.train(params=params, dtrain=dtrain, num_boost_round=10000)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QN37HT5PqKUn",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "preds = xg_reg.predict(dtrain)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Nh0ZmDHBZt1O",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import xgboost as xgb\n",
        "from sklearn.metrics import mean_squared_error"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8a7PyrQkZy9c",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "155f5bcd-1007-4b15-86e8-59e0b3b64b35"
      },
      "source": [
        "rmse = np.sqrt(mean_squared_error(y_train, preds))\n",
        "print(\"RMSE: %f\" % (rmse))"
      ],
      "execution_count": 44,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "RMSE: 1039.792499\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}