aws Baseline Notebook.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "dbdb3f00-8aca-4f36-b1cc-3cee1cf5f056",
   "metadata": {},
   "source": [
    "### NB: a five fold perfomed worse compared to 10 Kfold"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a039cd89-fc9c-4dd2-a370-5a1bc806b8af",
   "metadata": {},
   "source": [
    "## Data "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "id": "17893e4d-1b56-464c-923f-549d013b38b7",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import time\n",
    "\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "ba5dbe05-35fa-42ed-8a4d-bee5b2f9518e",
   "metadata": {},
   "outputs": [],
   "source": [
    "pd.set_option('display.max_column', None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "e21dfb46-9e1b-4908-8a09-c12c91ddc3e0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 23 ms\n",
      "Parser   : 205 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "def data_loader():\n",
    "    train = pd.read_csv('train.csv')\n",
    "    test = pd.read_csv('test.csv')\n",
    "    sub = pd.read_csv('SampleSubmission.csv')\n",
    "    policy = pd.read_csv('policies.csv')\n",
    "    data_dict = pd.read_csv('DataDictionary.csv')\n",
    "    return train,test, sub, policy, data_dict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "5f80cb6e-7409-468a-8ea6-090b59ed1f37",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 7.54 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "train,test, sub, policy, data_dict = data_loader()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "b1a08ba5-1849-40b1-b0bc-97c974fb6d22",
   "metadata": {},
   "outputs": [
    {
     "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>claim_id</th>\n",
       "      <th>policy_number</th>\n",
       "      <th>claim_number</th>\n",
       "      <th>sum_insured</th>\n",
       "      <th>product</th>\n",
       "      <th>agent</th>\n",
       "      <th>class_of_business</th>\n",
       "      <th>risk_type</th>\n",
       "      <th>loss_date</th>\n",
       "      <th>client_type</th>\n",
       "      <th>renewal_frequency</th>\n",
       "      <th>primary_cause</th>\n",
       "      <th>secondary_cause</th>\n",
       "      <th>branch</th>\n",
       "      <th>target</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ID_JZ7QWRHH57RO</td>\n",
       "      <td>pol000009680</td>\n",
       "      <td>claim0011359</td>\n",
       "      <td>300000.0</td>\n",
       "      <td>prod00027</td>\n",
       "      <td>ag00068</td>\n",
       "      <td>cob00031</td>\n",
       "      <td>rt00006</td>\n",
       "      <td>2021-05-13</td>\n",
       "      <td>ct0003</td>\n",
       "      <td>rf0001</td>\n",
       "      <td>pc0007</td>\n",
       "      <td>sc00022</td>\n",
       "      <td>br00006</td>\n",
       "      <td>82150.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ID_KQ5TJMVLJ4EP</td>\n",
       "      <td>pol000006869</td>\n",
       "      <td>claim0005272</td>\n",
       "      <td>16000.0</td>\n",
       "      <td>prod00005</td>\n",
       "      <td>ag00226</td>\n",
       "      <td>cob00010</td>\n",
       "      <td>rt00032</td>\n",
       "      <td>2018-05-13</td>\n",
       "      <td>ct0003</td>\n",
       "      <td>rf0001</td>\n",
       "      <td>pc0007</td>\n",
       "      <td>sc00023</td>\n",
       "      <td>br00002</td>\n",
       "      <td>3780.24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>ID_NYHI7WJGGIAE</td>\n",
       "      <td>pol000006135</td>\n",
       "      <td>claim0004147</td>\n",
       "      <td>3000.0</td>\n",
       "      <td>prod00029</td>\n",
       "      <td>ag00037</td>\n",
       "      <td>cob00010</td>\n",
       "      <td>rt00026</td>\n",
       "      <td>2018-08-18</td>\n",
       "      <td>ct0003</td>\n",
       "      <td>rf0001</td>\n",
       "      <td>pc0007</td>\n",
       "      <td>sc00024</td>\n",
       "      <td>br00002</td>\n",
       "      <td>2196.50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ID_ROMMDCXYUXN5</td>\n",
       "      <td>pol000008991</td>\n",
       "      <td>claim0009962</td>\n",
       "      <td>6000.0</td>\n",
       "      <td>prod00005</td>\n",
       "      <td>ag00037</td>\n",
       "      <td>cob00009</td>\n",
       "      <td>rt00032</td>\n",
       "      <td>2018-12-20</td>\n",
       "      <td>ct0003</td>\n",
       "      <td>rf0001</td>\n",
       "      <td>pc0007</td>\n",
       "      <td>sc00021</td>\n",
       "      <td>br00001</td>\n",
       "      <td>105.67</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ID_2OTD9NX8L73D</td>\n",
       "      <td>pol000007768</td>\n",
       "      <td>claim0007346</td>\n",
       "      <td>26850.0</td>\n",
       "      <td>prod00015</td>\n",
       "      <td>ag00226</td>\n",
       "      <td>cob00034</td>\n",
       "      <td>rt00023</td>\n",
       "      <td>2018-11-07</td>\n",
       "      <td>ct0001</td>\n",
       "      <td>rf0001</td>\n",
       "      <td>pc0007</td>\n",
       "      <td>sc00021</td>\n",
       "      <td>br00001</td>\n",
       "      <td>1605.64</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          claim_id policy_number  claim_number  sum_insured    product  \\\n",
       "0  ID_JZ7QWRHH57RO  pol000009680  claim0011359     300000.0  prod00027   \n",
       "1  ID_KQ5TJMVLJ4EP  pol000006869  claim0005272      16000.0  prod00005   \n",
       "2  ID_NYHI7WJGGIAE  pol000006135  claim0004147       3000.0  prod00029   \n",
       "3  ID_ROMMDCXYUXN5  pol000008991  claim0009962       6000.0  prod00005   \n",
       "4  ID_2OTD9NX8L73D  pol000007768  claim0007346      26850.0  prod00015   \n",
       "\n",
       "     agent class_of_business risk_type   loss_date client_type  \\\n",
       "0  ag00068          cob00031   rt00006  2021-05-13      ct0003   \n",
       "1  ag00226          cob00010   rt00032  2018-05-13      ct0003   \n",
       "2  ag00037          cob00010   rt00026  2018-08-18      ct0003   \n",
       "3  ag00037          cob00009   rt00032  2018-12-20      ct0003   \n",
       "4  ag00226          cob00034   rt00023  2018-11-07      ct0001   \n",
       "\n",
       "  renewal_frequency primary_cause secondary_cause   branch    target  \n",
       "0            rf0001        pc0007         sc00022  br00006  82150.00  \n",
       "1            rf0001        pc0007         sc00023  br00002   3780.24  \n",
       "2            rf0001        pc0007         sc00024  br00002   2196.50  \n",
       "3            rf0001        pc0007         sc00021  br00001    105.67  \n",
       "4            rf0001        pc0007         sc00021  br00001   1605.64  "
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "1a0b3246-76e5-453c-a27d-7f72b63cf89a",
   "metadata": {},
   "outputs": [
    {
     "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>claim_id</th>\n",
       "      <th>policy_number</th>\n",
       "      <th>claim_number</th>\n",
       "      <th>sum_insured</th>\n",
       "      <th>product</th>\n",
       "      <th>agent</th>\n",
       "      <th>class_of_business</th>\n",
       "      <th>risk_type</th>\n",
       "      <th>loss_date</th>\n",
       "      <th>client_type</th>\n",
       "      <th>renewal_frequency</th>\n",
       "      <th>primary_cause</th>\n",
       "      <th>secondary_cause</th>\n",
       "      <th>branch</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ID_4RO0KZI9BBW2</td>\n",
       "      <td>pol000007115</td>\n",
       "      <td>claim0005757</td>\n",
       "      <td>300000.0</td>\n",
       "      <td>prod00005</td>\n",
       "      <td>ag00068</td>\n",
       "      <td>cob00009</td>\n",
       "      <td>rt00032</td>\n",
       "      <td>2021-03-11</td>\n",
       "      <td>ct0003</td>\n",
       "      <td>rf0001</td>\n",
       "      <td>pc0007</td>\n",
       "      <td>sc00029</td>\n",
       "      <td>br00004</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ID_GTE9ZUJ9GZ97</td>\n",
       "      <td>pol000006008</td>\n",
       "      <td>claim0003037</td>\n",
       "      <td>2700000.0</td>\n",
       "      <td>prod00015</td>\n",
       "      <td>ag00064</td>\n",
       "      <td>cob00034</td>\n",
       "      <td>rt00023</td>\n",
       "      <td>2020-06-25</td>\n",
       "      <td>ct0001</td>\n",
       "      <td>rf0003</td>\n",
       "      <td>pc0007</td>\n",
       "      <td>sc00022</td>\n",
       "      <td>br00002</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>ID_RUVIYRKWZ4RG</td>\n",
       "      <td>pol000007651</td>\n",
       "      <td>claim0006507</td>\n",
       "      <td>4071352.0</td>\n",
       "      <td>prod00012</td>\n",
       "      <td>ag00021</td>\n",
       "      <td>cob00016</td>\n",
       "      <td>rt00015</td>\n",
       "      <td>2019-03-14</td>\n",
       "      <td>ct0001</td>\n",
       "      <td>rf0001</td>\n",
       "      <td>pc0011</td>\n",
       "      <td>sc00049</td>\n",
       "      <td>br00001</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ID_J81X8SWU6JYN</td>\n",
       "      <td>pol000010981</td>\n",
       "      <td>claim0013636</td>\n",
       "      <td>0.0</td>\n",
       "      <td>prod00002</td>\n",
       "      <td>ag00037</td>\n",
       "      <td>cob00002</td>\n",
       "      <td>rt00002</td>\n",
       "      <td>2018-12-04</td>\n",
       "      <td>ct0001</td>\n",
       "      <td>rf0002</td>\n",
       "      <td>pc0002</td>\n",
       "      <td>sc00012</td>\n",
       "      <td>br00005</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ID_94NM3JJAWWOQ</td>\n",
       "      <td>pol000005926</td>\n",
       "      <td>claim0001750</td>\n",
       "      <td>1607178.0</td>\n",
       "      <td>prod00015</td>\n",
       "      <td>ag00037</td>\n",
       "      <td>cob00034</td>\n",
       "      <td>rt00023</td>\n",
       "      <td>2019-09-18</td>\n",
       "      <td>ct0001</td>\n",
       "      <td>rf0003</td>\n",
       "      <td>pc0007</td>\n",
       "      <td>sc00022</td>\n",
       "      <td>br00002</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          claim_id policy_number  claim_number  sum_insured    product  \\\n",
       "0  ID_4RO0KZI9BBW2  pol000007115  claim0005757     300000.0  prod00005   \n",
       "1  ID_GTE9ZUJ9GZ97  pol000006008  claim0003037    2700000.0  prod00015   \n",
       "2  ID_RUVIYRKWZ4RG  pol000007651  claim0006507    4071352.0  prod00012   \n",
       "3  ID_J81X8SWU6JYN  pol000010981  claim0013636          0.0  prod00002   \n",
       "4  ID_94NM3JJAWWOQ  pol000005926  claim0001750    1607178.0  prod00015   \n",
       "\n",
       "     agent class_of_business risk_type   loss_date client_type  \\\n",
       "0  ag00068          cob00009   rt00032  2021-03-11      ct0003   \n",
       "1  ag00064          cob00034   rt00023  2020-06-25      ct0001   \n",
       "2  ag00021          cob00016   rt00015  2019-03-14      ct0001   \n",
       "3  ag00037          cob00002   rt00002  2018-12-04      ct0001   \n",
       "4  ag00037          cob00034   rt00023  2019-09-18      ct0001   \n",
       "\n",
       "  renewal_frequency primary_cause secondary_cause   branch  \n",
       "0            rf0001        pc0007         sc00029  br00004  \n",
       "1            rf0003        pc0007         sc00022  br00002  \n",
       "2            rf0001        pc0011         sc00049  br00001  \n",
       "3            rf0002        pc0002         sc00012  br00005  \n",
       "4            rf0003        pc0007         sc00022  br00002  "
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "d800b5c7-2068-4512-b483-a9943f1c6e56",
   "metadata": {},
   "outputs": [
    {
     "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>claim_id</th>\n",
       "      <th>target</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ID_4RO0KZI9BBW2</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ID_GTE9ZUJ9GZ97</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>ID_RUVIYRKWZ4RG</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ID_J81X8SWU6JYN</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ID_94NM3JJAWWOQ</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          claim_id  target\n",
       "0  ID_4RO0KZI9BBW2       0\n",
       "1  ID_GTE9ZUJ9GZ97       0\n",
       "2  ID_RUVIYRKWZ4RG       0\n",
       "3  ID_J81X8SWU6JYN       0\n",
       "4  ID_94NM3JJAWWOQ       0"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sub.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "5dfa20d7-3673-433e-b8f4-576b05027752",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(9376, 15)"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "306b8b20-b388-4b9e-9b60-d17d41064505",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(4019, 14)"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "563c83d6-1cc5-4e85-a29f-f002930ade95",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(4019, 2)"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sub.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "a761178c-669d-405d-88a0-e03173e81fb7",
   "metadata": {},
   "outputs": [
    {
     "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>policy_number</th>\n",
       "      <th>product_code</th>\n",
       "      <th>trans_date</th>\n",
       "      <th>from_date</th>\n",
       "      <th>to_date</th>\n",
       "      <th>agent</th>\n",
       "      <th>class_of_business</th>\n",
       "      <th>client_type</th>\n",
       "      <th>renewal_frequency</th>\n",
       "      <th>branch</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>pol000006995</td>\n",
       "      <td>prod00033</td>\n",
       "      <td>2017-07-01 08:24:55.800</td>\n",
       "      <td>2017-07-01 00:00:00.000</td>\n",
       "      <td>2018-07-01 00:00:00.000</td>\n",
       "      <td>ag00005</td>\n",
       "      <td>cob00035</td>\n",
       "      <td>ct0001</td>\n",
       "      <td>rf0002</td>\n",
       "      <td>br00002</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>pol000009828</td>\n",
       "      <td>prod00005</td>\n",
       "      <td>2017-07-01 09:04:02.217</td>\n",
       "      <td>2017-07-03 00:00:00.000</td>\n",
       "      <td>2017-11-02 00:00:00.000</td>\n",
       "      <td>ag00226</td>\n",
       "      <td>cob00009</td>\n",
       "      <td>ct0003</td>\n",
       "      <td>rf0001</td>\n",
       "      <td>br00006</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>pol000009828</td>\n",
       "      <td>prod00005</td>\n",
       "      <td>2017-07-01 09:04:02.217</td>\n",
       "      <td>2017-07-03 00:00:00.000</td>\n",
       "      <td>2017-11-02 00:00:00.000</td>\n",
       "      <td>ag00226</td>\n",
       "      <td>cob00031</td>\n",
       "      <td>ct0003</td>\n",
       "      <td>rf0001</td>\n",
       "      <td>br00006</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>pol000007729</td>\n",
       "      <td>prod00005</td>\n",
       "      <td>2017-07-01 09:36:53.843</td>\n",
       "      <td>2017-07-01 00:00:00.000</td>\n",
       "      <td>2017-10-31 00:00:00.000</td>\n",
       "      <td>ag00005</td>\n",
       "      <td>cob00031</td>\n",
       "      <td>ct0003</td>\n",
       "      <td>rf0001</td>\n",
       "      <td>br00002</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>pol000007729</td>\n",
       "      <td>prod00005</td>\n",
       "      <td>2017-07-01 09:36:53.843</td>\n",
       "      <td>2017-07-01 00:00:00.000</td>\n",
       "      <td>2017-10-31 00:00:00.000</td>\n",
       "      <td>ag00005</td>\n",
       "      <td>cob00009</td>\n",
       "      <td>ct0003</td>\n",
       "      <td>rf0001</td>\n",
       "      <td>br00002</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  policy_number product_code               trans_date  \\\n",
       "0  pol000006995    prod00033  2017-07-01 08:24:55.800   \n",
       "1  pol000009828    prod00005  2017-07-01 09:04:02.217   \n",
       "2  pol000009828    prod00005  2017-07-01 09:04:02.217   \n",
       "3  pol000007729    prod00005  2017-07-01 09:36:53.843   \n",
       "4  pol000007729    prod00005  2017-07-01 09:36:53.843   \n",
       "\n",
       "                 from_date                  to_date    agent  \\\n",
       "0  2017-07-01 00:00:00.000  2018-07-01 00:00:00.000  ag00005   \n",
       "1  2017-07-03 00:00:00.000  2017-11-02 00:00:00.000  ag00226   \n",
       "2  2017-07-03 00:00:00.000  2017-11-02 00:00:00.000  ag00226   \n",
       "3  2017-07-01 00:00:00.000  2017-10-31 00:00:00.000  ag00005   \n",
       "4  2017-07-01 00:00:00.000  2017-10-31 00:00:00.000  ag00005   \n",
       "\n",
       "  class_of_business client_type renewal_frequency   branch  \n",
       "0          cob00035      ct0001            rf0002  br00002  \n",
       "1          cob00009      ct0003            rf0001  br00006  \n",
       "2          cob00031      ct0003            rf0001  br00006  \n",
       "3          cob00031      ct0003            rf0001  br00002  \n",
       "4          cob00009      ct0003            rf0001  br00002  "
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "policy.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "0004aa20-4743-4471-b5f1-b4366b285d70",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(30371, 10)"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "policy.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "301d4f87-f7df-465f-a42a-5d7f99ab1ab6",
   "metadata": {},
   "outputs": [
    {
     "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>10</th>\n",
       "      <th>11</th>\n",
       "      <th>12</th>\n",
       "      <th>13</th>\n",
       "      <th>14</th>\n",
       "      <th>15</th>\n",
       "      <th>16</th>\n",
       "      <th>17</th>\n",
       "      <th>18</th>\n",
       "      <th>19</th>\n",
       "      <th>20</th>\n",
       "      <th>21</th>\n",
       "      <th>22</th>\n",
       "      <th>23</th>\n",
       "      <th>24</th>\n",
       "      <th>25</th>\n",
       "      <th>26</th>\n",
       "      <th>27</th>\n",
       "      <th>28</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Policy Data</th>\n",
       "      <td>Columns</td>\n",
       "      <td>policy number</td>\n",
       "      <td>product code</td>\n",
       "      <td>trans date</td>\n",
       "      <td>from date</td>\n",
       "      <td>to date</td>\n",
       "      <td>agency full name</td>\n",
       "      <td>class of business</td>\n",
       "      <td>client type</td>\n",
       "      <td>Renewal frequency</td>\n",
       "      <td>branch</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Claims Data</td>\n",
       "      <td>Claim Number</td>\n",
       "      <td>Policy Number</td>\n",
       "      <td>Loss Date</td>\n",
       "      <td>Sum Insured</td>\n",
       "      <td>Paid to Date</td>\n",
       "      <td>Agent</td>\n",
       "      <td>risk type</td>\n",
       "      <td>class of business</td>\n",
       "      <td>product</td>\n",
       "      <td>client type</td>\n",
       "      <td>Renewal frequency</td>\n",
       "      <td>primary cause</td>\n",
       "      <td>secondary cause</td>\n",
       "      <td>branch</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Unnamed: 1</th>\n",
       "      <td>Data type</td>\n",
       "      <td>Unique policy identifier</td>\n",
       "      <td>product code</td>\n",
       "      <td>date of policy transaction</td>\n",
       "      <td>first date of policy cover</td>\n",
       "      <td>last date of policy cover</td>\n",
       "      <td>agency</td>\n",
       "      <td>class of business</td>\n",
       "      <td>client type</td>\n",
       "      <td>policy renewal frequency</td>\n",
       "      <td>branch</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Description</td>\n",
       "      <td>Unique policy identifier</td>\n",
       "      <td>Date of loss event</td>\n",
       "      <td>Sum Insured</td>\n",
       "      <td>Total cost of claim.</td>\n",
       "      <td>Agency (categorical variable)</td>\n",
       "      <td>Risk Type</td>\n",
       "      <td>Class of business</td>\n",
       "      <td>Product</td>\n",
       "      <td>Client Type</td>\n",
       "      <td>Policy renewal frequency</td>\n",
       "      <td>Primary cause of loss event</td>\n",
       "      <td>Secondary cause of loss event</td>\n",
       "      <td>branch</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                    0                         1             2   \\\n",
       "Policy Data    Columns             policy number  product code   \n",
       "Unnamed: 1   Data type  Unique policy identifier  product code   \n",
       "\n",
       "                                     3                           4   \\\n",
       "Policy Data                  trans date                   from date   \n",
       "Unnamed: 1   date of policy transaction  first date of policy cover   \n",
       "\n",
       "                                    5                 6                  7   \\\n",
       "Policy Data                    to date  agency full name  class of business   \n",
       "Unnamed: 1   last date of policy cover            agency  class of business   \n",
       "\n",
       "                      8                         9       10   11   12   13  \\\n",
       "Policy Data  client type         Renewal frequency  branch  NaN  NaN  NaN   \n",
       "Unnamed: 1   client type  policy renewal frequency  branch  NaN  NaN  NaN   \n",
       "\n",
       "                      14            15                        16  \\\n",
       "Policy Data  Claims Data  Claim Number             Policy Number   \n",
       "Unnamed: 1           NaN   Description  Unique policy identifier   \n",
       "\n",
       "                             17           18                    19  \\\n",
       "Policy Data           Loss Date  Sum Insured          Paid to Date   \n",
       "Unnamed: 1   Date of loss event  Sum Insured  Total cost of claim.   \n",
       "\n",
       "                                        20         21                 22  \\\n",
       "Policy Data                          Agent  risk type  class of business   \n",
       "Unnamed: 1   Agency (categorical variable)  Risk Type  Class of business   \n",
       "\n",
       "                  23           24                        25  \\\n",
       "Policy Data  product  client type         Renewal frequency   \n",
       "Unnamed: 1   Product  Client Type  Policy renewal frequency   \n",
       "\n",
       "                                      26                             27  \\\n",
       "Policy Data                primary cause                secondary cause   \n",
       "Unnamed: 1   Primary cause of loss event  Secondary cause of loss event   \n",
       "\n",
       "                 28  \n",
       "Policy Data  branch  \n",
       "Unnamed: 1   branch  "
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_dict.T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "140c4ed8-d036-4546-be0a-acd88b408978",
   "metadata": {},
   "outputs": [],
   "source": [
    "same_col = [i for i in train.columns.tolist() if i in policy.columns.to_list()]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "795e5dce-c052-473e-a10c-cccf8bdeb9b1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['policy_number', 'product_code', 'trans_date', 'from_date',\n",
       "       'to_date', 'agent', 'class_of_business', 'client_type',\n",
       "       'renewal_frequency', 'branch'], dtype='<U17')"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.array(policy.columns.to_list())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 229,
   "id": "2dd16ac7-6460-4502-ab99-7be0be48cb89",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.concat([train, test]).reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 230,
   "id": "9b3a494d-da4a-473a-a7c5-09c9e39ae990",
   "metadata": {},
   "outputs": [],
   "source": [
    "ntrain = train.shape[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 231,
   "id": "aa3ecf72",
   "metadata": {},
   "outputs": [],
   "source": [
    "df['risk_business'] = df['class_of_business'] + '_' + df['risk_type']\n",
    "df['product_agent'] = df['agent'] + '_' + df['product']\n",
    "df['renwal_business'] = df['renewal_frequency'] + '_' + df['class_of_business']\n",
    "df['product_class'] = df['class_of_business'] + '_' + df['product']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 237,
   "id": "ebd5ae83",
   "metadata": {},
   "outputs": [],
   "source": [
    "df['year'] = [int(i.split('-')[0]) - 2018 for i in df.loss_date]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 238,
   "id": "3399e158",
   "metadata": {},
   "outputs": [],
   "source": [
    "df['cause'] = df['secondary_cause'] + '_' + df['primary_cause']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 239,
   "id": "56d80b70",
   "metadata": {},
   "outputs": [],
   "source": [
    "df['product_class'] = df['class_of_business'] + '_' + df['product']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 240,
   "id": "67f4b06d",
   "metadata": {},
   "outputs": [],
   "source": [
    "id_claim = [i.split('m')[-1] for i in df['claim_number']]\n",
    "df['id_claim'] = id_claim\n",
    "df['idclaim_freq'] = df['id_claim'].map(df['id_claim'].value_counts())\n",
    "df = df.drop('id_claim', axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 241,
   "id": "9d43ddbe",
   "metadata": {},
   "outputs": [],
   "source": [
    "df['loss_date'] = pd.to_datetime(df['loss_date'], format='%Y-%m-%d')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 242,
   "id": "17e0d3a3",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Hp\\AppData\\Local\\Temp/ipykernel_49288/360260820.py:2: FutureWarning: Series.dt.weekofyear and Series.dt.week have been deprecated.  Please use Series.dt.isocalendar().week instead.\n",
      "  df[attr] = getattr(df['loss_date'].dt, attr)\n"
     ]
    }
   ],
   "source": [
    "for attr in ['day', 'month', 'week', 'dayofweek', 'weekofyear', 'days_in_month', 'is_month_start', 'is_month_end', 'dayofyear']:\n",
    "    df[attr] = getattr(df['loss_date'].dt, attr)\n",
    "df['is_weekend'] = (df['dayofweek'] >= 5)*1\n",
    "df['fortnight'] = df['day']%15\n",
    "df['which_fortnight'] = df['day']//15"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 243,
   "id": "4fffac27",
   "metadata": {},
   "outputs": [],
   "source": [
    "df['branch_product'] = pd.Series(df.branch + '_' + df['product'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 244,
   "id": "077e1645",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set_theme(color_codes=True)\n",
    "sns.lmplot(x=\"target\", y=\"sum_insured\", data=train);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 281,
   "id": "5762b28a",
   "metadata": {},
   "outputs": [],
   "source": [
    "#df = df.drop(['claim_id', 'loss_date'], axis = 1)\n",
    "df = df.fillna(-999999)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 284,
   "id": "cf69ed5e-f7fb-46fa-a725-7c839b276c6b",
   "metadata": {},
   "outputs": [],
   "source": [
    "#cat_feat = [val for index, val in enumerate(features) if index in cat_feat[0]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 285,
   "id": "948989a3",
   "metadata": {},
   "outputs": [],
   "source": [
    "new_test = df[train.shape[0]:].reset_index().drop('target', axis = 1)\n",
    "new_train = df[:train.shape[0]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 286,
   "id": "af007f16",
   "metadata": {},
   "outputs": [],
   "source": [
    "features = [feat for feat in list(new_train) if feat != 'target' ]\n",
    "cat_feat = np.where(new_train[features].dtypes != float)[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 287,
   "id": "593589dc-66c7-4567-9168-b677f60036d3",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn import model_selection\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn.metrics import mean_squared_error\n",
    "from math import sqrt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 288,
   "id": "b3f8f578-eada-42a2-ba51-9c897bd24385",
   "metadata": {},
   "outputs": [],
   "source": [
    "fold = model_selection.KFold(n_splits=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 308,
   "id": "4819e841-4316-4709-b410-3d1d8ceaa9ff",
   "metadata": {},
   "outputs": [],
   "source": [
    "params = {'cat_features':cat_feat,\n",
    "         'eval_metric': 'RMSE',\n",
    "          'iterations':1000,\n",
    "         'verbose': 10,\n",
    "         'od_type':'Iter',\n",
    "         'random_state':200}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 309,
   "id": "a04e2b9a-7ff6-4746-9ee8-f813f772d412",
   "metadata": {},
   "outputs": [],
   "source": [
    "from catboost import CatBoostRegressor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 310,
   "id": "71803d57-b46e-46c1-b889-7199f3700d3e",
   "metadata": {},
   "outputs": [],
   "source": [
    "cat_model = CatBoostRegressor(**params)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 311,
   "id": "3585340e-2272-4495-b8f1-be1364722893",
   "metadata": {},
   "outputs": [],
   "source": [
    "def metric(y_val, pred):\n",
    "    return sqrt(metrics.mean_squared_error(y_val, pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 312,
   "id": "9eded473",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn import metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 313,
   "id": "596fd541-3a2e-4f84-89c7-737869da878b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 0 ns\n",
      "Parser   : 237 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "def modelling_tree(algorithm,yy,X, test_df):\n",
    "    cv_score=[]\n",
    "    test_pred=[]\n",
    "    train_score=[]\n",
    "    \n",
    "    for train_index, test_index in fold.split(X,pd.Series(yy)):\n",
    "        \n",
    "        X_trainx, X_val = X.iloc[train_index], X.iloc[test_index]\n",
    "        y_trainx, y_val = pd.Series(yy).iloc[train_index], pd.Series(yy).iloc[test_index]\n",
    "\n",
    "        algorithm.fit(X_trainx,y_trainx, eval_set=[(X_trainx,y_trainx),(X_val, y_val)], use_best_model = True)\n",
    "        \n",
    "        train_preds=algorithm.predict(X_trainx)\n",
    "        train_score.append(metric(y_trainx, train_preds))\n",
    "        \n",
    "        val_preds=algorithm.predict(X_val)\n",
    "        cv_score.append(metric(y_val, val_preds))\n",
    "\n",
    "        test_p = algorithm.predict(test_df)\n",
    "        test_pred.append(test_p)\n",
    "\n",
    "\n",
    "    print(f'Training Accuracy Score {np.mean(train_score)}')\n",
    "    print(f'Validation Accuracy Score {np.mean(cv_score)}')\n",
    "    \n",
    "    return test_pred"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 314,
   "id": "44c7db03",
   "metadata": {},
   "outputs": [
    {
     "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>policy_number</th>\n",
       "      <th>claim_number</th>\n",
       "      <th>sum_insured</th>\n",
       "      <th>product</th>\n",
       "      <th>agent</th>\n",
       "      <th>class_of_business</th>\n",
       "      <th>risk_type</th>\n",
       "      <th>client_type</th>\n",
       "      <th>renewal_frequency</th>\n",
       "      <th>primary_cause</th>\n",
       "      <th>secondary_cause</th>\n",
       "      <th>branch</th>\n",
       "      <th>target</th>\n",
       "      <th>risk_business</th>\n",
       "      <th>product_agent</th>\n",
       "      <th>renwal_business</th>\n",
       "      <th>product_class</th>\n",
       "      <th>cause</th>\n",
       "      <th>year</th>\n",
       "      <th>idclaim_freq</th>\n",
       "      <th>day</th>\n",
       "      <th>month</th>\n",
       "      <th>week</th>\n",
       "      <th>dayofweek</th>\n",
       "      <th>weekofyear</th>\n",
       "      <th>days_in_month</th>\n",
       "      <th>is_month_start</th>\n",
       "      <th>is_month_end</th>\n",
       "      <th>dayofyear</th>\n",
       "      <th>is_weekend</th>\n",
       "      <th>fortnight</th>\n",
       "      <th>which_fortnight</th>\n",
       "      <th>branch_product</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>pol000009680</td>\n",
       "      <td>claim0011359</td>\n",
       "      <td>300000.0</td>\n",
       "      <td>prod00027</td>\n",
       "      <td>ag00068</td>\n",
       "      <td>cob00031</td>\n",
       "      <td>rt00006</td>\n",
       "      <td>ct0003</td>\n",
       "      <td>rf0001</td>\n",
       "      <td>pc0007</td>\n",
       "      <td>sc00022</td>\n",
       "      <td>br00006</td>\n",
       "      <td>82150.00</td>\n",
       "      <td>cob00031_rt00006</td>\n",
       "      <td>ag00068_prod00027</td>\n",
       "      <td>rf0001_cob00031</td>\n",
       "      <td>cob00031_prod00027</td>\n",
       "      <td>sc00022_pc0007</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>13</td>\n",
       "      <td>5</td>\n",
       "      <td>19</td>\n",
       "      <td>3</td>\n",
       "      <td>19</td>\n",
       "      <td>31</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>133</td>\n",
       "      <td>0</td>\n",
       "      <td>13</td>\n",
       "      <td>0</td>\n",
       "      <td>br00006_prod00027</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>pol000006869</td>\n",
       "      <td>claim0005272</td>\n",
       "      <td>16000.0</td>\n",
       "      <td>prod00005</td>\n",
       "      <td>ag00226</td>\n",
       "      <td>cob00010</td>\n",
       "      <td>rt00032</td>\n",
       "      <td>ct0003</td>\n",
       "      <td>rf0001</td>\n",
       "      <td>pc0007</td>\n",
       "      <td>sc00023</td>\n",
       "      <td>br00002</td>\n",
       "      <td>3780.24</td>\n",
       "      <td>cob00010_rt00032</td>\n",
       "      <td>ag00226_prod00005</td>\n",
       "      <td>rf0001_cob00010</td>\n",
       "      <td>cob00010_prod00005</td>\n",
       "      <td>sc00023_pc0007</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>13</td>\n",
       "      <td>5</td>\n",
       "      <td>19</td>\n",
       "      <td>6</td>\n",
       "      <td>19</td>\n",
       "      <td>31</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>133</td>\n",
       "      <td>1</td>\n",
       "      <td>13</td>\n",
       "      <td>0</td>\n",
       "      <td>br00002_prod00005</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>pol000006135</td>\n",
       "      <td>claim0004147</td>\n",
       "      <td>3000.0</td>\n",
       "      <td>prod00029</td>\n",
       "      <td>ag00037</td>\n",
       "      <td>cob00010</td>\n",
       "      <td>rt00026</td>\n",
       "      <td>ct0003</td>\n",
       "      <td>rf0001</td>\n",
       "      <td>pc0007</td>\n",
       "      <td>sc00024</td>\n",
       "      <td>br00002</td>\n",
       "      <td>2196.50</td>\n",
       "      <td>cob00010_rt00026</td>\n",
       "      <td>ag00037_prod00029</td>\n",
       "      <td>rf0001_cob00010</td>\n",
       "      <td>cob00010_prod00029</td>\n",
       "      <td>sc00024_pc0007</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>18</td>\n",
       "      <td>8</td>\n",
       "      <td>33</td>\n",
       "      <td>5</td>\n",
       "      <td>33</td>\n",
       "      <td>31</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>230</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>br00002_prod00029</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>pol000008991</td>\n",
       "      <td>claim0009962</td>\n",
       "      <td>6000.0</td>\n",
       "      <td>prod00005</td>\n",
       "      <td>ag00037</td>\n",
       "      <td>cob00009</td>\n",
       "      <td>rt00032</td>\n",
       "      <td>ct0003</td>\n",
       "      <td>rf0001</td>\n",
       "      <td>pc0007</td>\n",
       "      <td>sc00021</td>\n",
       "      <td>br00001</td>\n",
       "      <td>105.67</td>\n",
       "      <td>cob00009_rt00032</td>\n",
       "      <td>ag00037_prod00005</td>\n",
       "      <td>rf0001_cob00009</td>\n",
       "      <td>cob00009_prod00005</td>\n",
       "      <td>sc00021_pc0007</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>20</td>\n",
       "      <td>12</td>\n",
       "      <td>51</td>\n",
       "      <td>3</td>\n",
       "      <td>51</td>\n",
       "      <td>31</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>354</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "      <td>br00001_prod00005</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>pol000007768</td>\n",
       "      <td>claim0007346</td>\n",
       "      <td>26850.0</td>\n",
       "      <td>prod00015</td>\n",
       "      <td>ag00226</td>\n",
       "      <td>cob00034</td>\n",
       "      <td>rt00023</td>\n",
       "      <td>ct0001</td>\n",
       "      <td>rf0001</td>\n",
       "      <td>pc0007</td>\n",
       "      <td>sc00021</td>\n",
       "      <td>br00001</td>\n",
       "      <td>1605.64</td>\n",
       "      <td>cob00034_rt00023</td>\n",
       "      <td>ag00226_prod00015</td>\n",
       "      <td>rf0001_cob00034</td>\n",
       "      <td>cob00034_prod00015</td>\n",
       "      <td>sc00021_pc0007</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>11</td>\n",
       "      <td>45</td>\n",
       "      <td>2</td>\n",
       "      <td>45</td>\n",
       "      <td>30</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>311</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>br00001_prod00015</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9371</th>\n",
       "      <td>pol000006054</td>\n",
       "      <td>claim0003743</td>\n",
       "      <td>42000.0</td>\n",
       "      <td>prod00015</td>\n",
       "      <td>ag00037</td>\n",
       "      <td>cob00034</td>\n",
       "      <td>rt00023</td>\n",
       "      <td>ct0001</td>\n",
       "      <td>rf0002</td>\n",
       "      <td>pc0007</td>\n",
       "      <td>sc00021</td>\n",
       "      <td>br00002</td>\n",
       "      <td>205.68</td>\n",
       "      <td>cob00034_rt00023</td>\n",
       "      <td>ag00037_prod00015</td>\n",
       "      <td>rf0002_cob00034</td>\n",
       "      <td>cob00034_prod00015</td>\n",
       "      <td>sc00021_pc0007</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>9</td>\n",
       "      <td>36</td>\n",
       "      <td>0</td>\n",
       "      <td>36</td>\n",
       "      <td>30</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>245</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>br00002_prod00015</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9372</th>\n",
       "      <td>pol000006715</td>\n",
       "      <td>claim0005088</td>\n",
       "      <td>6800.0</td>\n",
       "      <td>prod00005</td>\n",
       "      <td>ag00064</td>\n",
       "      <td>cob00009</td>\n",
       "      <td>rt00032</td>\n",
       "      <td>ct0003</td>\n",
       "      <td>rf0002</td>\n",
       "      <td>pc0007</td>\n",
       "      <td>sc00021</td>\n",
       "      <td>br00002</td>\n",
       "      <td>1133.32</td>\n",
       "      <td>cob00009_rt00032</td>\n",
       "      <td>ag00064_prod00005</td>\n",
       "      <td>rf0002_cob00009</td>\n",
       "      <td>cob00009_prod00005</td>\n",
       "      <td>sc00021_pc0007</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>12</td>\n",
       "      <td>5</td>\n",
       "      <td>19</td>\n",
       "      <td>5</td>\n",
       "      <td>19</td>\n",
       "      <td>31</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>132</td>\n",
       "      <td>1</td>\n",
       "      <td>12</td>\n",
       "      <td>0</td>\n",
       "      <td>br00002_prod00005</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9373</th>\n",
       "      <td>pol000008236</td>\n",
       "      <td>claim0008976</td>\n",
       "      <td>5000.0</td>\n",
       "      <td>prod00005</td>\n",
       "      <td>ag00064</td>\n",
       "      <td>cob00009</td>\n",
       "      <td>rt00032</td>\n",
       "      <td>ct0003</td>\n",
       "      <td>rf0001</td>\n",
       "      <td>pc0007</td>\n",
       "      <td>sc00029</td>\n",
       "      <td>br00001</td>\n",
       "      <td>2176.00</td>\n",
       "      <td>cob00009_rt00032</td>\n",
       "      <td>ag00064_prod00005</td>\n",
       "      <td>rf0001_cob00009</td>\n",
       "      <td>cob00009_prod00005</td>\n",
       "      <td>sc00029_pc0007</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>28</td>\n",
       "      <td>7</td>\n",
       "      <td>30</td>\n",
       "      <td>5</td>\n",
       "      <td>30</td>\n",
       "      <td>31</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>209</td>\n",
       "      <td>1</td>\n",
       "      <td>13</td>\n",
       "      <td>1</td>\n",
       "      <td>br00001_prod00005</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9374</th>\n",
       "      <td>pol000006847</td>\n",
       "      <td>claim0005250</td>\n",
       "      <td>6000.0</td>\n",
       "      <td>prod00005</td>\n",
       "      <td>ag00037</td>\n",
       "      <td>cob00009</td>\n",
       "      <td>rt00032</td>\n",
       "      <td>ct0003</td>\n",
       "      <td>rf0001</td>\n",
       "      <td>pc0007</td>\n",
       "      <td>sc00022</td>\n",
       "      <td>br00002</td>\n",
       "      <td>851.40</td>\n",
       "      <td>cob00009_rt00032</td>\n",
       "      <td>ag00037_prod00005</td>\n",
       "      <td>rf0001_cob00009</td>\n",
       "      <td>cob00009_prod00005</td>\n",
       "      <td>sc00022_pc0007</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>16</td>\n",
       "      <td>8</td>\n",
       "      <td>33</td>\n",
       "      <td>3</td>\n",
       "      <td>33</td>\n",
       "      <td>31</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>228</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>br00002_prod00005</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9375</th>\n",
       "      <td>pol000007450</td>\n",
       "      <td>claim0006249</td>\n",
       "      <td>10000.0</td>\n",
       "      <td>prod00027</td>\n",
       "      <td>ag00226</td>\n",
       "      <td>cob00010</td>\n",
       "      <td>rt00006</td>\n",
       "      <td>ct0003</td>\n",
       "      <td>rf0002</td>\n",
       "      <td>pc0007</td>\n",
       "      <td>sc00023</td>\n",
       "      <td>br00001</td>\n",
       "      <td>1410.00</td>\n",
       "      <td>cob00010_rt00006</td>\n",
       "      <td>ag00226_prod00027</td>\n",
       "      <td>rf0002_cob00010</td>\n",
       "      <td>cob00010_prod00027</td>\n",
       "      <td>sc00023_pc0007</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>9</td>\n",
       "      <td>36</td>\n",
       "      <td>5</td>\n",
       "      <td>36</td>\n",
       "      <td>30</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>250</td>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>br00001_prod00027</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>9376 rows × 33 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     policy_number  claim_number  sum_insured    product    agent  \\\n",
       "0     pol000009680  claim0011359     300000.0  prod00027  ag00068   \n",
       "1     pol000006869  claim0005272      16000.0  prod00005  ag00226   \n",
       "2     pol000006135  claim0004147       3000.0  prod00029  ag00037   \n",
       "3     pol000008991  claim0009962       6000.0  prod00005  ag00037   \n",
       "4     pol000007768  claim0007346      26850.0  prod00015  ag00226   \n",
       "...            ...           ...          ...        ...      ...   \n",
       "9371  pol000006054  claim0003743      42000.0  prod00015  ag00037   \n",
       "9372  pol000006715  claim0005088       6800.0  prod00005  ag00064   \n",
       "9373  pol000008236  claim0008976       5000.0  prod00005  ag00064   \n",
       "9374  pol000006847  claim0005250       6000.0  prod00005  ag00037   \n",
       "9375  pol000007450  claim0006249      10000.0  prod00027  ag00226   \n",
       "\n",
       "     class_of_business risk_type client_type renewal_frequency primary_cause  \\\n",
       "0             cob00031   rt00006      ct0003            rf0001        pc0007   \n",
       "1             cob00010   rt00032      ct0003            rf0001        pc0007   \n",
       "2             cob00010   rt00026      ct0003            rf0001        pc0007   \n",
       "3             cob00009   rt00032      ct0003            rf0001        pc0007   \n",
       "4             cob00034   rt00023      ct0001            rf0001        pc0007   \n",
       "...                ...       ...         ...               ...           ...   \n",
       "9371          cob00034   rt00023      ct0001            rf0002        pc0007   \n",
       "9372          cob00009   rt00032      ct0003            rf0002        pc0007   \n",
       "9373          cob00009   rt00032      ct0003            rf0001        pc0007   \n",
       "9374          cob00009   rt00032      ct0003            rf0001        pc0007   \n",
       "9375          cob00010   rt00006      ct0003            rf0002        pc0007   \n",
       "\n",
       "     secondary_cause   branch    target     risk_business      product_agent  \\\n",
       "0            sc00022  br00006  82150.00  cob00031_rt00006  ag00068_prod00027   \n",
       "1            sc00023  br00002   3780.24  cob00010_rt00032  ag00226_prod00005   \n",
       "2            sc00024  br00002   2196.50  cob00010_rt00026  ag00037_prod00029   \n",
       "3            sc00021  br00001    105.67  cob00009_rt00032  ag00037_prod00005   \n",
       "4            sc00021  br00001   1605.64  cob00034_rt00023  ag00226_prod00015   \n",
       "...              ...      ...       ...               ...                ...   \n",
       "9371         sc00021  br00002    205.68  cob00034_rt00023  ag00037_prod00015   \n",
       "9372         sc00021  br00002   1133.32  cob00009_rt00032  ag00064_prod00005   \n",
       "9373         sc00029  br00001   2176.00  cob00009_rt00032  ag00064_prod00005   \n",
       "9374         sc00022  br00002    851.40  cob00009_rt00032  ag00037_prod00005   \n",
       "9375         sc00023  br00001   1410.00  cob00010_rt00006  ag00226_prod00027   \n",
       "\n",
       "      renwal_business       product_class           cause  year  idclaim_freq  \\\n",
       "0     rf0001_cob00031  cob00031_prod00027  sc00022_pc0007     3             1   \n",
       "1     rf0001_cob00010  cob00010_prod00005  sc00023_pc0007     0             2   \n",
       "2     rf0001_cob00010  cob00010_prod00029  sc00024_pc0007     0             1   \n",
       "3     rf0001_cob00009  cob00009_prod00005  sc00021_pc0007     0             1   \n",
       "4     rf0001_cob00034  cob00034_prod00015  sc00021_pc0007     0             1   \n",
       "...               ...                 ...             ...   ...           ...   \n",
       "9371  rf0002_cob00034  cob00034_prod00015  sc00021_pc0007     1             1   \n",
       "9372  rf0002_cob00009  cob00009_prod00005  sc00021_pc0007     0             1   \n",
       "9373  rf0001_cob00009  cob00009_prod00005  sc00029_pc0007     0             1   \n",
       "9374  rf0001_cob00009  cob00009_prod00005  sc00022_pc0007     0             1   \n",
       "9375  rf0002_cob00010  cob00010_prod00027  sc00023_pc0007     1             1   \n",
       "\n",
       "      day  month  week  dayofweek  weekofyear  days_in_month  is_month_start  \\\n",
       "0      13      5    19          3          19             31           False   \n",
       "1      13      5    19          6          19             31           False   \n",
       "2      18      8    33          5          33             31           False   \n",
       "3      20     12    51          3          51             31           False   \n",
       "4       7     11    45          2          45             30           False   \n",
       "...   ...    ...   ...        ...         ...            ...             ...   \n",
       "9371    2      9    36          0          36             30           False   \n",
       "9372   12      5    19          5          19             31           False   \n",
       "9373   28      7    30          5          30             31           False   \n",
       "9374   16      8    33          3          33             31           False   \n",
       "9375    7      9    36          5          36             30           False   \n",
       "\n",
       "      is_month_end  dayofyear  is_weekend  fortnight  which_fortnight  \\\n",
       "0            False        133           0         13                0   \n",
       "1            False        133           1         13                0   \n",
       "2            False        230           1          3                1   \n",
       "3            False        354           0          5                1   \n",
       "4            False        311           0          7                0   \n",
       "...            ...        ...         ...        ...              ...   \n",
       "9371         False        245           0          2                0   \n",
       "9372         False        132           1         12                0   \n",
       "9373         False        209           1         13                1   \n",
       "9374         False        228           0          1                1   \n",
       "9375         False        250           1          7                0   \n",
       "\n",
       "         branch_product  \n",
       "0     br00006_prod00027  \n",
       "1     br00002_prod00005  \n",
       "2     br00002_prod00029  \n",
       "3     br00001_prod00005  \n",
       "4     br00001_prod00015  \n",
       "...                 ...  \n",
       "9371  br00002_prod00015  \n",
       "9372  br00002_prod00005  \n",
       "9373  br00001_prod00005  \n",
       "9374  br00002_prod00005  \n",
       "9375  br00001_prod00027  \n",
       "\n",
       "[9376 rows x 33 columns]"
      ]
     },
     "execution_count": 314,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 315,
   "id": "bd1af8cd",
   "metadata": {},
   "outputs": [],
   "source": [
    "XX = new_train.drop('target', axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 316,
   "id": "d947a496",
   "metadata": {},
   "outputs": [],
   "source": [
    "test_x= new_test[XX.columns.to_list()]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 317,
   "id": "8a6a325f-70a5-4baa-a16e-c1e0517f44e4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Learning rate set to 0.071173\n",
      "0:\tlearn: 208010.3431873\ttest: 208013.4274350\ttest1: 191006.0375941\tbest: 191006.0375941 (0)\ttotal: 660ms\tremaining: 10m 59s\n",
      "10:\tlearn: 196268.6849152\ttest: 195713.8941399\ttest1: 178232.0847513\tbest: 178232.0847513 (10)\ttotal: 5.45s\tremaining: 8m 9s\n",
      "20:\tlearn: 190849.0628074\ttest: 191740.2842668\ttest1: 174438.6824107\tbest: 174438.6824107 (20)\ttotal: 10.1s\tremaining: 7m 48s\n",
      "30:\tlearn: 188726.0078980\ttest: 189790.3689551\ttest1: 173148.8448212\tbest: 173148.8448212 (30)\ttotal: 14s\tremaining: 7m 17s\n",
      "40:\tlearn: 186738.2831822\ttest: 188312.7397940\ttest1: 172722.9028093\tbest: 172694.2717914 (39)\ttotal: 18.5s\tremaining: 7m 13s\n",
      "50:\tlearn: 184887.0023609\ttest: 187192.2216915\ttest1: 172054.9453130\tbest: 172026.2210606 (49)\ttotal: 22.6s\tremaining: 7m 1s\n",
      "60:\tlearn: 184156.0035789\ttest: 186496.9811932\ttest1: 171928.3189934\tbest: 171907.4036751 (58)\ttotal: 27.3s\tremaining: 7m\n",
      "70:\tlearn: 181800.1401339\ttest: 185218.4998653\ttest1: 171953.5905289\tbest: 171870.3147040 (64)\ttotal: 30.7s\tremaining: 6m 41s\n",
      "80:\tlearn: 181490.6529528\ttest: 185020.9789236\ttest1: 171907.3726273\tbest: 171870.3147040 (64)\ttotal: 32.3s\tremaining: 6m 6s\n",
      "90:\tlearn: 179296.6260349\ttest: 184471.6416420\ttest1: 171845.4520758\tbest: 171844.0812959 (89)\ttotal: 34s\tremaining: 5m 39s\n",
      "100:\tlearn: 177670.8001986\ttest: 183266.9200694\ttest1: 171289.8919017\tbest: 171282.3282070 (97)\ttotal: 35.6s\tremaining: 5m 16s\n",
      "110:\tlearn: 176976.2935840\ttest: 182865.8902313\ttest1: 171093.7899165\tbest: 171087.9630309 (109)\ttotal: 37.4s\tremaining: 4m 59s\n",
      "120:\tlearn: 175860.6228764\ttest: 182722.5364819\ttest1: 171055.3131468\tbest: 171048.6565666 (115)\ttotal: 39s\tremaining: 4m 43s\n",
      "130:\tlearn: 174645.4024074\ttest: 181602.8192722\ttest1: 171033.6367635\tbest: 171003.2995218 (125)\ttotal: 40.8s\tremaining: 4m 30s\n",
      "140:\tlearn: 173545.9943581\ttest: 180665.8831962\ttest1: 171105.8851193\tbest: 171003.2995218 (125)\ttotal: 42.5s\tremaining: 4m 18s\n",
      "Stopped by overfitting detector  (20 iterations wait)\n",
      "\n",
      "bestTest = 171003.2995\n",
      "bestIteration = 125\n",
      "\n",
      "Shrink model to first 126 iterations.\n",
      "Learning rate set to 0.071173\n",
      "0:\tlearn: 204318.7920388\ttest: 204343.3985287\ttest1: 225041.9725847\tbest: 225041.9725847 (0)\ttotal: 195ms\tremaining: 3m 14s\n",
      "10:\tlearn: 190727.5878353\ttest: 191533.4709850\ttest1: 215901.7547033\tbest: 215901.7547033 (10)\ttotal: 2.14s\tremaining: 3m 12s\n",
      "20:\tlearn: 181450.2845033\ttest: 183525.2720854\ttest1: 212680.7686319\tbest: 212680.7686319 (20)\ttotal: 3.87s\tremaining: 3m\n",
      "30:\tlearn: 174839.1538762\ttest: 178594.9997976\ttest1: 211065.0265655\tbest: 211065.0265655 (30)\ttotal: 5.59s\tremaining: 2m 54s\n",
      "40:\tlearn: 171672.1066018\ttest: 176995.4726283\ttest1: 209701.7420382\tbest: 209701.7420382 (40)\ttotal: 7.59s\tremaining: 2m 57s\n",
      "50:\tlearn: 167614.4205758\ttest: 174218.6386626\ttest1: 209876.8018503\tbest: 209599.6435363 (48)\ttotal: 9.29s\tremaining: 2m 52s\n",
      "60:\tlearn: 166061.1720763\ttest: 173389.2506216\ttest1: 209502.2844754\tbest: 209502.2844754 (60)\ttotal: 11s\tremaining: 2m 49s\n",
      "70:\tlearn: 164129.1522210\ttest: 172395.1153610\ttest1: 208167.6258552\tbest: 208165.7317316 (68)\ttotal: 12.5s\tremaining: 2m 44s\n",
      "80:\tlearn: 161877.9764843\ttest: 171233.8862158\ttest1: 208046.1730273\tbest: 207840.8524852 (74)\ttotal: 14.6s\tremaining: 2m 45s\n",
      "90:\tlearn: 158226.8778456\ttest: 168374.6298449\ttest1: 208252.7439678\tbest: 207840.8524852 (74)\ttotal: 16.5s\tremaining: 2m 45s\n",
      "Stopped by overfitting detector  (20 iterations wait)\n",
      "\n",
      "bestTest = 207840.8525\n",
      "bestIteration = 74\n",
      "\n",
      "Shrink model to first 75 iterations.\n",
      "Learning rate set to 0.071173\n",
      "0:\tlearn: 212993.2412328\ttest: 213048.6011632\ttest1: 129438.0700829\tbest: 129438.0700829 (0)\ttotal: 169ms\tremaining: 2m 48s\n",
      "10:\tlearn: 196801.7577272\ttest: 197655.7266887\ttest1: 118679.3747148\tbest: 118679.3747148 (10)\ttotal: 2.22s\tremaining: 3m 19s\n",
      "20:\tlearn: 188386.9893587\ttest: 189856.1716325\ttest1: 115925.4098615\tbest: 115838.2056451 (19)\ttotal: 3.96s\tremaining: 3m 4s\n",
      "30:\tlearn: 185499.0593857\ttest: 187285.4884163\ttest1: 116041.2257627\tbest: 115786.6099620 (25)\ttotal: 5.82s\tremaining: 3m 2s\n",
      "40:\tlearn: 181088.8703913\ttest: 183525.4269589\ttest1: 115335.1600052\tbest: 115335.1600052 (40)\ttotal: 7.59s\tremaining: 2m 57s\n",
      "50:\tlearn: 176821.0324062\ttest: 181346.8617188\ttest1: 115424.4288947\tbest: 115335.1600052 (40)\ttotal: 9.55s\tremaining: 2m 57s\n",
      "60:\tlearn: 173676.0624465\ttest: 178736.7690506\ttest1: 115591.4793979\tbest: 115335.1600052 (40)\ttotal: 11.7s\tremaining: 2m 59s\n",
      "Stopped by overfitting detector  (20 iterations wait)\n",
      "\n",
      "bestTest = 115335.16\n",
      "bestIteration = 40\n",
      "\n",
      "Shrink model to first 41 iterations.\n",
      "Learning rate set to 0.071173\n",
      "0:\tlearn: 191286.7730059\ttest: 191158.0670792\ttest1: 311116.4114010\tbest: 311116.4114010 (0)\ttotal: 194ms\tremaining: 3m 14s\n",
      "10:\tlearn: 176261.0675615\ttest: 177631.9436828\ttest1: 302542.1846868\tbest: 302542.1846868 (10)\ttotal: 2.01s\tremaining: 3m 1s\n",
      "20:\tlearn: 168163.2620801\ttest: 171193.8138869\ttest1: 299015.4724820\tbest: 299015.4724820 (20)\ttotal: 3.76s\tremaining: 2m 55s\n",
      "30:\tlearn: 165089.5426164\ttest: 168794.8468418\ttest1: 297232.4268904\tbest: 297232.4268904 (30)\ttotal: 5.83s\tremaining: 3m 2s\n",
      "40:\tlearn: 162122.0177060\ttest: 166938.1265573\ttest1: 296847.5648556\tbest: 296836.9124122 (39)\ttotal: 7.59s\tremaining: 2m 57s\n",
      "50:\tlearn: 160334.2654081\ttest: 165233.7660721\ttest1: 296775.2290239\tbest: 296735.7111243 (48)\ttotal: 9.37s\tremaining: 2m 54s\n",
      "60:\tlearn: 157493.8937972\ttest: 163092.0987274\ttest1: 295922.2908770\tbest: 295922.2908770 (60)\ttotal: 11.2s\tremaining: 2m 52s\n",
      "70:\tlearn: 155837.7520581\ttest: 162427.7746583\ttest1: 295766.6394200\tbest: 295766.6394200 (70)\ttotal: 12.9s\tremaining: 2m 49s\n",
      "80:\tlearn: 154223.4484759\ttest: 161459.8401152\ttest1: 295812.2112993\tbest: 295766.6394200 (70)\ttotal: 14.7s\tremaining: 2m 46s\n",
      "90:\tlearn: 153455.0254969\ttest: 160876.8010789\ttest1: 295759.3267654\tbest: 295733.4824631 (89)\ttotal: 16.3s\tremaining: 2m 42s\n",
      "100:\tlearn: 151972.2246915\ttest: 160168.3562074\ttest1: 295856.2041259\tbest: 295733.4824631 (89)\ttotal: 18s\tremaining: 2m 40s\n",
      "110:\tlearn: 150530.2985176\ttest: 158679.8406682\ttest1: 295670.0887162\tbest: 295654.0699142 (109)\ttotal: 19.8s\tremaining: 2m 38s\n",
      "120:\tlearn: 148742.3875279\ttest: 157421.4877093\ttest1: 295419.9815015\tbest: 295321.2906899 (116)\ttotal: 21.6s\tremaining: 2m 37s\n",
      "130:\tlearn: 148385.3181118\ttest: 157120.6321386\ttest1: 295272.3186760\tbest: 295272.3186760 (130)\ttotal: 23.5s\tremaining: 2m 36s\n",
      "140:\tlearn: 147405.2785630\ttest: 156545.0607922\ttest1: 295236.5881339\tbest: 295097.7468746 (136)\ttotal: 25.6s\tremaining: 2m 36s\n",
      "150:\tlearn: 146542.0746624\ttest: 156075.7772931\ttest1: 295377.4279109\tbest: 295097.7468746 (136)\ttotal: 27.4s\tremaining: 2m 34s\n",
      "Stopped by overfitting detector  (20 iterations wait)\n",
      "\n",
      "bestTest = 295097.7469\n",
      "bestIteration = 136\n",
      "\n",
      "Shrink model to first 137 iterations.\n",
      "Learning rate set to 0.071173\n",
      "0:\tlearn: 205715.0922052\ttest: 206511.0018061\ttest1: 206641.0450341\tbest: 206641.0450341 (0)\ttotal: 179ms\tremaining: 2m 58s\n",
      "10:\tlearn: 190876.0260750\ttest: 193671.0671648\ttest1: 195729.2007847\tbest: 195729.2007847 (10)\ttotal: 2.21s\tremaining: 3m 19s\n",
      "20:\tlearn: 182482.8999604\ttest: 186491.7706212\ttest1: 190396.5925523\tbest: 190396.5925523 (20)\ttotal: 4.02s\tremaining: 3m 7s\n",
      "30:\tlearn: 178382.9102759\ttest: 183954.4531988\ttest1: 184887.9408728\tbest: 184853.3692945 (29)\ttotal: 5.93s\tremaining: 3m 5s\n",
      "40:\tlearn: 174563.3091353\ttest: 181198.8071391\ttest1: 183499.1695249\tbest: 183402.9965207 (38)\ttotal: 7.59s\tremaining: 2m 57s\n",
      "50:\tlearn: 171830.7630112\ttest: 180641.8483820\ttest1: 182957.3899162\tbest: 182708.0681871 (42)\ttotal: 9.24s\tremaining: 2m 52s\n",
      "60:\tlearn: 169591.3501888\ttest: 179224.1624042\ttest1: 179724.4162555\tbest: 179724.4162555 (60)\ttotal: 11.2s\tremaining: 2m 51s\n",
      "70:\tlearn: 167948.5521671\ttest: 178271.7891064\ttest1: 179709.6018824\tbest: 179406.2047583 (65)\ttotal: 13.5s\tremaining: 2m 56s\n",
      "80:\tlearn: 165502.5940199\ttest: 176527.6537769\ttest1: 180162.0641869\tbest: 179406.2047583 (65)\ttotal: 15.4s\tremaining: 2m 54s\n",
      "Stopped by overfitting detector  (20 iterations wait)\n",
      "\n",
      "bestTest = 179406.2048\n",
      "bestIteration = 65\n",
      "\n",
      "Shrink model to first 66 iterations.\n",
      "Learning rate set to 0.071173\n",
      "0:\tlearn: 213341.7662533\ttest: 213997.4854090\ttest1: 123037.7225798\tbest: 123037.7225798 (0)\ttotal: 168ms\tremaining: 2m 48s\n",
      "10:\tlearn: 199290.2224152\ttest: 203315.2509360\ttest1: 111706.7266145\tbest: 111706.7266145 (10)\ttotal: 1.95s\tremaining: 2m 54s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "20:\tlearn: 191374.9091840\ttest: 196879.8492860\ttest1: 107385.0476600\tbest: 107385.0476600 (20)\ttotal: 3.77s\tremaining: 2m 56s\n",
      "30:\tlearn: 186561.3363040\ttest: 194023.2373397\ttest1: 105582.5475821\tbest: 105394.3144030 (29)\ttotal: 5.65s\tremaining: 2m 56s\n",
      "40:\tlearn: 182918.7947918\ttest: 191445.9370822\ttest1: 104731.8209318\tbest: 104728.3263930 (39)\ttotal: 7.46s\tremaining: 2m 54s\n",
      "50:\tlearn: 178055.2777564\ttest: 188808.6937286\ttest1: 103464.4853158\tbest: 103464.4853158 (50)\ttotal: 9.16s\tremaining: 2m 50s\n",
      "60:\tlearn: 174214.5604521\ttest: 186704.6445673\ttest1: 102554.2540492\tbest: 102471.9136559 (59)\ttotal: 11s\tremaining: 2m 49s\n",
      "70:\tlearn: 171335.9014353\ttest: 185756.6035439\ttest1: 101444.4086379\tbest: 101444.4086379 (70)\ttotal: 12.7s\tremaining: 2m 46s\n",
      "80:\tlearn: 168789.8526125\ttest: 183746.4604411\ttest1: 100807.9109380\tbest: 100807.9109380 (80)\ttotal: 14.6s\tremaining: 2m 45s\n",
      "90:\tlearn: 166398.9964185\ttest: 183210.8607101\ttest1: 100542.0898691\tbest: 100506.9459183 (88)\ttotal: 16.2s\tremaining: 2m 41s\n",
      "100:\tlearn: 164144.1948227\ttest: 181362.7513132\ttest1: 100424.1147693\tbest: 100302.7134947 (98)\ttotal: 17.8s\tremaining: 2m 38s\n",
      "110:\tlearn: 162310.6882176\ttest: 180333.3936726\ttest1: 100658.6788875\tbest: 100302.7134947 (98)\ttotal: 19.5s\tremaining: 2m 36s\n",
      "Stopped by overfitting detector  (20 iterations wait)\n",
      "\n",
      "bestTest = 100302.7135\n",
      "bestIteration = 98\n",
      "\n",
      "Shrink model to first 99 iterations.\n",
      "Learning rate set to 0.071174\n",
      "0:\tlearn: 206513.4207707\ttest: 206574.2502485\ttest1: 202798.8646202\tbest: 202798.8646202 (0)\ttotal: 171ms\tremaining: 2m 50s\n",
      "10:\tlearn: 195424.9535244\ttest: 195195.9753468\ttest1: 191722.1246547\tbest: 191722.1246547 (10)\ttotal: 2.03s\tremaining: 3m 2s\n",
      "20:\tlearn: 190968.0840816\ttest: 191457.4897443\ttest1: 187833.0259309\tbest: 187833.0259309 (20)\ttotal: 3.73s\tremaining: 2m 53s\n",
      "30:\tlearn: 187997.8049973\ttest: 189523.3333358\ttest1: 186815.2929040\tbest: 186803.0254250 (29)\ttotal: 5.46s\tremaining: 2m 50s\n",
      "40:\tlearn: 186479.4724259\ttest: 188712.7952920\ttest1: 185845.9303227\tbest: 185845.9303227 (40)\ttotal: 7.35s\tremaining: 2m 51s\n",
      "50:\tlearn: 185459.8467693\ttest: 188177.9231400\ttest1: 184252.3618600\tbest: 184252.3618600 (50)\ttotal: 8.99s\tremaining: 2m 47s\n",
      "60:\tlearn: 184633.0746741\ttest: 187545.2843218\ttest1: 184070.5666002\tbest: 184070.5666002 (60)\ttotal: 10.9s\tremaining: 2m 47s\n",
      "70:\tlearn: 182133.4211558\ttest: 186601.1162823\ttest1: 182817.1128241\tbest: 182817.1128241 (70)\ttotal: 12.5s\tremaining: 2m 43s\n",
      "80:\tlearn: 180815.0361263\ttest: 185645.9627170\ttest1: 181641.7213789\tbest: 181602.3615003 (77)\ttotal: 14.4s\tremaining: 2m 43s\n",
      "90:\tlearn: 180186.7210486\ttest: 185063.0250668\ttest1: 181620.4966733\tbest: 181602.3615003 (77)\ttotal: 16.4s\tremaining: 2m 43s\n",
      "100:\tlearn: 178912.4251982\ttest: 184119.7733795\ttest1: 180502.8059786\tbest: 180419.3376682 (94)\ttotal: 18.4s\tremaining: 2m 43s\n",
      "110:\tlearn: 178128.7885617\ttest: 183411.0024420\ttest1: 180400.1938556\tbest: 180400.1938556 (110)\ttotal: 19.8s\tremaining: 2m 38s\n",
      "120:\tlearn: 177131.8351052\ttest: 183229.6645416\ttest1: 180232.3814399\tbest: 180232.3814399 (120)\ttotal: 21.6s\tremaining: 2m 36s\n",
      "130:\tlearn: 175734.2323725\ttest: 182810.1165215\ttest1: 180123.8425094\tbest: 180095.2508351 (129)\ttotal: 23.5s\tremaining: 2m 36s\n",
      "140:\tlearn: 174673.9075803\ttest: 181778.5855631\ttest1: 178089.0695595\tbest: 178077.5301524 (139)\ttotal: 25.5s\tremaining: 2m 35s\n",
      "150:\tlearn: 173888.3489563\ttest: 181665.0095742\ttest1: 177924.5273149\tbest: 177924.5273149 (150)\ttotal: 27.5s\tremaining: 2m 34s\n",
      "160:\tlearn: 172354.6217183\ttest: 180754.8074153\ttest1: 177719.3807479\tbest: 177719.3807479 (160)\ttotal: 29.2s\tremaining: 2m 32s\n",
      "170:\tlearn: 170973.2685014\ttest: 180509.7488589\ttest1: 177446.0444423\tbest: 177399.2929182 (168)\ttotal: 31.2s\tremaining: 2m 31s\n",
      "180:\tlearn: 169951.6580704\ttest: 180016.7009489\ttest1: 177153.8878312\tbest: 177153.8878312 (180)\ttotal: 33s\tremaining: 2m 29s\n",
      "190:\tlearn: 167592.3892884\ttest: 178049.0475407\ttest1: 177245.3343825\tbest: 177147.5497875 (181)\ttotal: 35s\tremaining: 2m 28s\n",
      "200:\tlearn: 166424.3340712\ttest: 177207.4625219\ttest1: 176323.4893024\tbest: 176323.4893024 (200)\ttotal: 36.9s\tremaining: 2m 26s\n",
      "210:\tlearn: 165032.7323987\ttest: 176473.5501292\ttest1: 176224.0369007\tbest: 176224.0369007 (210)\ttotal: 38.7s\tremaining: 2m 24s\n",
      "220:\tlearn: 163479.4413287\ttest: 175056.7715335\ttest1: 176122.4896880\tbest: 176120.1225213 (218)\ttotal: 40.5s\tremaining: 2m 22s\n",
      "230:\tlearn: 162919.1899458\ttest: 174542.4454947\ttest1: 176054.5005112\tbest: 176054.5005112 (230)\ttotal: 42.5s\tremaining: 2m 21s\n",
      "240:\tlearn: 161649.9915602\ttest: 173127.0167673\ttest1: 176099.1271046\tbest: 176054.5005112 (230)\ttotal: 44.3s\tremaining: 2m 19s\n",
      "250:\tlearn: 160167.8643008\ttest: 171789.9416501\ttest1: 175575.9963547\tbest: 175491.3857984 (249)\ttotal: 46s\tremaining: 2m 17s\n",
      "260:\tlearn: 159066.2565904\ttest: 171088.7491242\ttest1: 175305.2678194\tbest: 175173.3388835 (252)\ttotal: 47.7s\tremaining: 2m 15s\n",
      "270:\tlearn: 158615.5227844\ttest: 170655.6660676\ttest1: 175187.2880407\tbest: 175173.3388835 (252)\ttotal: 49.5s\tremaining: 2m 13s\n",
      "280:\tlearn: 157959.4498114\ttest: 170105.7982047\ttest1: 175121.4936662\tbest: 175105.0373141 (272)\ttotal: 51.3s\tremaining: 2m 11s\n",
      "290:\tlearn: 156148.0715095\ttest: 168225.9029123\ttest1: 174849.9650990\tbest: 174790.4194468 (287)\ttotal: 53.2s\tremaining: 2m 9s\n",
      "300:\tlearn: 155606.3577903\ttest: 167998.8271662\ttest1: 174413.6438359\tbest: 174413.6438359 (300)\ttotal: 54.9s\tremaining: 2m 7s\n",
      "310:\tlearn: 154839.7926360\ttest: 167483.0531584\ttest1: 174237.7679969\tbest: 174201.2653916 (308)\ttotal: 56.5s\tremaining: 2m 5s\n",
      "320:\tlearn: 154507.3286644\ttest: 167193.8708419\ttest1: 174122.5043425\tbest: 174113.4866370 (317)\ttotal: 58.2s\tremaining: 2m 3s\n",
      "330:\tlearn: 153916.8441687\ttest: 167100.1778462\ttest1: 174122.0937633\tbest: 174100.0740251 (327)\ttotal: 59.9s\tremaining: 2m 1s\n",
      "340:\tlearn: 153103.9993884\ttest: 166760.2303878\ttest1: 174235.5378131\tbest: 174042.1374718 (331)\ttotal: 1m 1s\tremaining: 1m 59s\n",
      "350:\tlearn: 152532.2766376\ttest: 166436.6129855\ttest1: 174449.2651853\tbest: 174042.1374718 (331)\ttotal: 1m 3s\tremaining: 1m 57s\n",
      "Stopped by overfitting detector  (20 iterations wait)\n",
      "\n",
      "bestTest = 174042.1375\n",
      "bestIteration = 331\n",
      "\n",
      "Shrink model to first 332 iterations.\n",
      "Learning rate set to 0.071174\n",
      "0:\tlearn: 214187.6027314\ttest: 214259.8020444\ttest1: 108141.0531825\tbest: 108141.0531825 (0)\ttotal: 171ms\tremaining: 2m 50s\n",
      "10:\tlearn: 200595.1608190\ttest: 201332.0354088\ttest1: 98597.8922832\tbest: 98597.8922832 (10)\ttotal: 2.02s\tremaining: 3m 1s\n",
      "20:\tlearn: 195415.4712356\ttest: 197502.1907287\ttest1: 96472.5221860\tbest: 96472.5221860 (20)\ttotal: 3.48s\tremaining: 2m 42s\n",
      "30:\tlearn: 190732.2315912\ttest: 193298.7962980\ttest1: 96327.4455785\tbest: 96038.4047928 (25)\ttotal: 4.86s\tremaining: 2m 32s\n",
      "40:\tlearn: 186512.4693739\ttest: 189972.9036628\ttest1: 95718.1040881\tbest: 95666.3846783 (37)\ttotal: 6.61s\tremaining: 2m 34s\n",
      "50:\tlearn: 183808.1585280\ttest: 188161.8368352\ttest1: 95868.6176876\tbest: 95594.5799508 (42)\ttotal: 8.31s\tremaining: 2m 34s\n",
      "60:\tlearn: 180850.6102417\ttest: 185504.1754807\ttest1: 95993.5099553\tbest: 95594.5799508 (42)\ttotal: 10.1s\tremaining: 2m 35s\n",
      "Stopped by overfitting detector  (20 iterations wait)\n",
      "\n",
      "bestTest = 95594.57995\n",
      "bestIteration = 42\n",
      "\n",
      "Shrink model to first 43 iterations.\n",
      "Learning rate set to 0.071174\n",
      "0:\tlearn: 192504.2458797\ttest: 192444.7305947\ttest1: 306228.0641826\tbest: 306228.0641826 (0)\ttotal: 173ms\tremaining: 2m 52s\n",
      "10:\tlearn: 176526.9411354\ttest: 178619.7125927\ttest1: 298950.2816011\tbest: 298950.2816011 (10)\ttotal: 1.93s\tremaining: 2m 53s\n",
      "20:\tlearn: 171097.2212105\ttest: 174620.6840217\ttest1: 296327.9445275\tbest: 296327.9445275 (20)\ttotal: 3.62s\tremaining: 2m 48s\n",
      "30:\tlearn: 166096.3940885\ttest: 171119.1309869\ttest1: 294652.6695107\tbest: 294652.6695107 (30)\ttotal: 5.15s\tremaining: 2m 41s\n",
      "40:\tlearn: 163535.2630539\ttest: 169034.8718989\ttest1: 293797.4138624\tbest: 293797.4138624 (40)\ttotal: 7.07s\tremaining: 2m 45s\n",
      "50:\tlearn: 160621.8359545\ttest: 167870.7829597\ttest1: 292889.8160667\tbest: 292889.8160667 (50)\ttotal: 9.03s\tremaining: 2m 48s\n",
      "60:\tlearn: 158643.5796988\ttest: 167196.3191259\ttest1: 292622.2495050\tbest: 292600.9967696 (55)\ttotal: 10.8s\tremaining: 2m 45s\n",
      "70:\tlearn: 157314.7141501\ttest: 165729.5775774\ttest1: 292111.5007150\tbest: 292084.8386395 (68)\ttotal: 12.3s\tremaining: 2m 41s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "80:\tlearn: 155369.2958075\ttest: 164425.5150416\ttest1: 292098.7612614\tbest: 292014.2116535 (74)\ttotal: 14.3s\tremaining: 2m 42s\n",
      "90:\tlearn: 153647.7694290\ttest: 163173.7444899\ttest1: 291897.0368034\tbest: 291897.0368034 (90)\ttotal: 16.4s\tremaining: 2m 44s\n",
      "100:\tlearn: 151582.3091861\ttest: 160960.8637133\ttest1: 291824.5344254\tbest: 291824.5344254 (100)\ttotal: 18s\tremaining: 2m 40s\n",
      "110:\tlearn: 150227.6349926\ttest: 160704.1969604\ttest1: 291622.4712915\tbest: 291622.4712915 (110)\ttotal: 19.6s\tremaining: 2m 37s\n",
      "120:\tlearn: 148868.2668378\ttest: 159790.5806897\ttest1: 291169.3951407\tbest: 291121.5461529 (115)\ttotal: 21.5s\tremaining: 2m 35s\n",
      "130:\tlearn: 148158.3058911\ttest: 159555.0795605\ttest1: 291110.1397050\tbest: 291110.1397050 (130)\ttotal: 23.2s\tremaining: 2m 33s\n",
      "140:\tlearn: 147397.7671323\ttest: 159460.7141411\ttest1: 291114.5081620\tbest: 291082.7199384 (137)\ttotal: 25.2s\tremaining: 2m 33s\n",
      "150:\tlearn: 146534.8590335\ttest: 158777.8428797\ttest1: 291042.6262342\tbest: 291042.6262342 (150)\ttotal: 27s\tremaining: 2m 31s\n",
      "160:\tlearn: 145892.0349911\ttest: 158472.8690831\ttest1: 290883.1273336\tbest: 290883.1273336 (160)\ttotal: 28.9s\tremaining: 2m 30s\n",
      "170:\tlearn: 143863.8339224\ttest: 157594.2996954\ttest1: 290229.8853876\tbest: 290227.5008686 (169)\ttotal: 30.9s\tremaining: 2m 29s\n",
      "180:\tlearn: 143094.2345927\ttest: 157090.2087079\ttest1: 290247.5190754\tbest: 290157.6972083 (176)\ttotal: 33.1s\tremaining: 2m 29s\n",
      "190:\tlearn: 140761.1495392\ttest: 156033.6734334\ttest1: 289969.8528822\tbest: 289659.7994404 (183)\ttotal: 35.2s\tremaining: 2m 28s\n",
      "200:\tlearn: 139185.0408320\ttest: 155342.0551483\ttest1: 290525.3220436\tbest: 289659.7994404 (183)\ttotal: 37s\tremaining: 2m 27s\n",
      "Stopped by overfitting detector  (20 iterations wait)\n",
      "\n",
      "bestTest = 289659.7994\n",
      "bestIteration = 183\n",
      "\n",
      "Shrink model to first 184 iterations.\n",
      "Learning rate set to 0.071174\n",
      "0:\tlearn: 211946.6673278\ttest: 211968.8016774\ttest1: 153893.3678946\tbest: 153893.3678946 (0)\ttotal: 172ms\tremaining: 2m 52s\n",
      "10:\tlearn: 196940.2243871\ttest: 197374.8310102\ttest1: 147226.5747861\tbest: 147226.5747861 (10)\ttotal: 1.93s\tremaining: 2m 53s\n",
      "20:\tlearn: 189199.1038465\ttest: 190569.3635213\ttest1: 144872.0211059\tbest: 144872.0211059 (20)\ttotal: 3.57s\tremaining: 2m 46s\n",
      "30:\tlearn: 185152.4102044\ttest: 187422.3070833\ttest1: 143059.3445074\tbest: 143054.5959013 (29)\ttotal: 5.26s\tremaining: 2m 44s\n",
      "40:\tlearn: 182241.5339098\ttest: 185787.6419142\ttest1: 142994.1884502\tbest: 142914.0625619 (39)\ttotal: 6.89s\tremaining: 2m 41s\n",
      "50:\tlearn: 179790.8136756\ttest: 184239.4434393\ttest1: 143376.0259588\tbest: 142821.1261901 (42)\ttotal: 8.44s\tremaining: 2m 37s\n",
      "60:\tlearn: 177652.7291946\ttest: 182300.3833495\ttest1: 143491.3890911\tbest: 142821.1261901 (42)\ttotal: 9.99s\tremaining: 2m 33s\n",
      "Stopped by overfitting detector  (20 iterations wait)\n",
      "\n",
      "bestTest = 142821.1262\n",
      "bestIteration = 42\n",
      "\n",
      "Shrink model to first 43 iterations.\n",
      "Training Accuracy Score 175331.0932095681\n",
      "Validation Accuracy Score 177110.36256680655\n",
      "Wall time: 4min 46s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "test_pred = modelling_tree(algorithm = cat_model, yy = train.target, X = XX, test_df = test_x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 319,
   "id": "269e2206",
   "metadata": {},
   "outputs": [],
   "source": [
    "cat_sub = np.mean(test_pred, axis = 0)\n",
    "sub['target'] = cat_sub\n",
    "sub['target'].describe()\n",
    "\n",
    "sub.to_csv('1000est_base_cat.csv', index = False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "687bf627",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 334,
   "id": "43c04ccf",
   "metadata": {},
   "outputs": [],
   "source": [
    "feat_imp = pd.DataFrame({'Feature': cat_model.feature_names_, 'Score': cat_model.feature_importances_})\n",
    "feat_imp = feat_imp.sort_values(by = 'Score', ascending = False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 335,
   "id": "cc5ffd3e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize = (10,10))\n",
    "sns.barplot(y = 'Feature', x = 'Score', data = feat_imp);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ad664a5b",
   "metadata": {},
   "outputs": [],
   "source": [
    "[feat_imp['Score'][i] for i in range(feat_imp.shape[0])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 347,
   "id": "83ff47d4",
   "metadata": {},
   "outputs": [],
   "source": [
    "feat_to_use = [feat_imp['Feature'][i] for i in range(feat_imp.shape[0]) if feat_imp['Score'][i] > 0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 349,
   "id": "bdedca61",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['sum_insured', 'product', 'class_of_business', 'risk_type', 'client_type', 'renewal_frequency', 'primary_cause', 'secondary_cause', 'branch', 'risk_business', 'product_agent', 'renwal_business', 'product_class', 'cause', 'year', 'idclaim_freq', 'day', 'month', 'week', 'dayofweek', 'weekofyear', 'days_in_month', 'dayofyear', 'fortnight', 'which_fortnight', 'branch_product']\n"
     ]
    }
   ],
   "source": [
    "print(feat_to_use)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 307,
   "id": "9f4c63da-5a4c-428b-b218-422a0446016d",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "\"\"\"\n",
    "This Submission was \n",
    "\n",
    "params = {'cat_features':cat_feat,\n",
    "         'eval_metric': 'RMSE',\n",
    "          'iterations':3000,\n",
    "         'verbose': 100,\n",
    "         'od_type':'Iter',\n",
    "         'random_state':200}\n",
    "\"\"\"\n",
    "\n",
    "\n",
    "cat_sub = np.mean(test_pred, axis = 0)\n",
    "sub['target'] = cat_sub\n",
    "sub['target'].describe()\n",
    "\n",
    "sub.to_csv('aws_base_cat.csv', index = False)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}