2.a.isbankasi-overfit-xgboost.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
    "execution": {
     "iopub.execute_input": "2020-11-09T11:34:42.289680Z",
     "iopub.status.busy": "2020-11-09T11:34:42.288828Z",
     "iopub.status.idle": "2020-11-09T11:34:43.020075Z",
     "shell.execute_reply": "2020-11-09T11:34:43.019432Z"
    },
    "papermill": {
     "duration": 0.761189,
     "end_time": "2020-11-09T11:34:43.020191",
     "exception": false,
     "start_time": "2020-11-09T11:34:42.259002",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "#read dfs# import packages\n",
    "import csv\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import os \n",
    "from sklearn.metrics import accuracy_score, roc_curve, confusion_matrix, classification_report, roc_auc_score\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.model_selection import ShuffleSplit\n",
    "from sklearn.model_selection import RandomizedSearchCV\n",
    "import sys \n",
    "from sklearn.model_selection import learning_curve\n",
    "from sklearn.datasets import make_classification\n",
    "from sklearn.preprocessing import StandardScaler, LabelEncoder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-11-09T11:34:43.063312Z",
     "iopub.status.busy": "2020-11-09T11:34:43.062707Z",
     "iopub.status.idle": "2020-11-09T11:35:12.595864Z",
     "shell.execute_reply": "2020-11-09T11:35:12.594411Z"
    },
    "papermill": {
     "duration": 29.557625,
     "end_time": "2020-11-09T11:35:12.596000",
     "exception": false,
     "start_time": "2020-11-09T11:34:43.038375",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "train_X=pd.read_csv(\"../input/vol3-kopya-no-scaling/reduced_train_X.csv\")\n",
    "train_y=pd.read_csv(\"../input/vol3-kopya-no-scaling/train_y.csv\")\n",
    "normalized_test_df=pd.read_csv(\"../input/vol3-kopya-no-scaling/reduced_test_X.csv\")\n",
    "df_test1=pd.read_csv(\"../input/vol3-kopya-no-scaling/df_test1.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "_cell_guid": "79c7e3d0-c299-4dcb-8224-4455121ee9b0",
    "_uuid": "d629ff2d2480ee46fbb7e2d37f6b5fab8052498a",
    "execution": {
     "iopub.execute_input": "2020-11-09T11:35:12.844753Z",
     "iopub.status.busy": "2020-11-09T11:35:12.843189Z",
     "iopub.status.idle": "2020-11-09T11:35:13.051484Z",
     "shell.execute_reply": "2020-11-09T11:35:13.050903Z"
    },
    "papermill": {
     "duration": 0.437461,
     "end_time": "2020-11-09T11:35:13.051633",
     "exception": false,
     "start_time": "2020-11-09T11:35:12.614172",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "train_X=train_X.drop(columns=train_X.columns[0])\n",
    "train_y=train_y.drop(columns=train_y.columns[0])\n",
    "normalized_test_df=normalized_test_df.drop(columns=normalized_test_df.columns[0])\n",
    "df_test1=df_test1.drop(columns=df_test1.columns[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-11-09T11:35:13.094993Z",
     "iopub.status.busy": "2020-11-09T11:35:13.093975Z",
     "iopub.status.idle": "2020-11-09T11:35:13.107951Z",
     "shell.execute_reply": "2020-11-09T11:35:13.107482Z"
    },
    "papermill": {
     "duration": 0.037314,
     "end_time": "2020-11-09T11:35:13.108048",
     "exception": false,
     "start_time": "2020-11-09T11:35:13.070734",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "ISLEM_TUTARI_Y    0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_y.isna().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-11-09T11:35:13.530360Z",
     "iopub.status.busy": "2020-11-09T11:35:13.529252Z",
     "iopub.status.idle": "2020-11-09T11:35:13.654956Z",
     "shell.execute_reply": "2020-11-09T11:35:13.654351Z"
    },
    "papermill": {
     "duration": 0.528568,
     "end_time": "2020-11-09T11:35:13.655066",
     "exception": false,
     "start_time": "2020-11-09T11:35:13.126498",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "bin_dict = {'PESIN': 1, 'TAKSITLI': 0}\n",
    "train_X['ISLEM_TURU'] = train_X['ISLEM_TURU'].map(bin_dict)\n",
    "normalized_test_df['ISLEM_TURU'] = normalized_test_df['ISLEM_TURU'].map(bin_dict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-11-09T11:35:13.699393Z",
     "iopub.status.busy": "2020-11-09T11:35:13.698065Z",
     "iopub.status.idle": "2020-11-09T11:35:14.051516Z",
     "shell.execute_reply": "2020-11-09T11:35:14.050901Z"
    },
    "papermill": {
     "duration": 0.377555,
     "end_time": "2020-11-09T11:35:14.051654",
     "exception": false,
     "start_time": "2020-11-09T11:35:13.674099",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "#drop yıl_ay\n",
    "train_X=train_X.drop(columns=[\"YIL_AY\"])\n",
    "normalized_test_df=normalized_test_df.drop(columns=[\"YIL_AY\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-11-09T11:35:14.102658Z",
     "iopub.status.busy": "2020-11-09T11:35:14.101654Z",
     "iopub.status.idle": "2020-11-09T11:35:14.109541Z",
     "shell.execute_reply": "2020-11-09T11:35:14.110977Z"
    },
    "papermill": {
     "duration": 0.040509,
     "end_time": "2020-11-09T11:35:14.111153",
     "exception": false,
     "start_time": "2020-11-09T11:35:14.070644",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 3526227 entries, 0 to 3526226\n",
      "Data columns (total 28 columns):\n",
      " #   Column                             Dtype  \n",
      "---  ------                             -----  \n",
      " 0   ISLEM_ADEDI                        float64\n",
      " 1   ISLEM_TURU                         int64  \n",
      " 2   customer_id                        int64  \n",
      " 3   sector_id                          int64  \n",
      " 4   cit_bin                            float64\n",
      " 5   islem_ortalamasi_per_sek_cust_tur  float64\n",
      " 6   adet_per_cust                      float64\n",
      " 7   tutar_per_cust                     float64\n",
      " 8   islem_ort_per_cust                 float64\n",
      " 9   adet_per_sek                       float64\n",
      " 10  tutar_per_sek                      float64\n",
      " 11  islem_ort_per_sek                  float64\n",
      " 12  month                              object \n",
      " 13  season                             object \n",
      " 14  adet_per_month                     float64\n",
      " 15  tutar_per_month                    float64\n",
      " 16  islem_ort_per_month                float64\n",
      " 17  islem_ort_per_year                 float64\n",
      " 18  adet_per_ay_sek                    float64\n",
      " 19  tutar_per_ay_sek                   float64\n",
      " 20  islem_ort_per_ay_sek               float64\n",
      " 21  adet_per_mev_sek                   float64\n",
      " 22  tutar_per_mev_sek                  float64\n",
      " 23  doviz_har_ort                      float64\n",
      " 24  pca_ek1                            float64\n",
      " 25  pca_ek2                            float64\n",
      " 26  pca_ek3                            float64\n",
      " 27  pca_doviz1                         float64\n",
      "dtypes: float64(23), int64(3), object(2)\n",
      "memory usage: 753.3+ MB\n"
     ]
    }
   ],
   "source": [
    "train_X.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-11-09T11:35:14.155732Z",
     "iopub.status.busy": "2020-11-09T11:35:14.154725Z",
     "iopub.status.idle": "2020-11-09T11:35:14.641731Z",
     "shell.execute_reply": "2020-11-09T11:35:14.642224Z"
    },
    "papermill": {
     "duration": 0.511344,
     "end_time": "2020-11-09T11:35:14.642370",
     "exception": false,
     "start_time": "2020-11-09T11:35:14.131026",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "df = pd.concat([train_X, normalized_test_df], axis=0, sort=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-11-09T11:35:14.688437Z",
     "iopub.status.busy": "2020-11-09T11:35:14.687081Z",
     "iopub.status.idle": "2020-11-09T11:35:17.349904Z",
     "shell.execute_reply": "2020-11-09T11:35:17.349326Z"
    },
    "papermill": {
     "duration": 2.688133,
     "end_time": "2020-11-09T11:35:17.350012",
     "exception": false,
     "start_time": "2020-11-09T11:35:14.661879",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "df = pd.get_dummies(data=df, columns=['sector_id',\"cit_bin\",\"month\",\"season\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-11-09T11:35:17.416419Z",
     "iopub.status.busy": "2020-11-09T11:35:17.415770Z",
     "iopub.status.idle": "2020-11-09T11:35:17.419115Z",
     "shell.execute_reply": "2020-11-09T11:35:17.418647Z"
    },
    "papermill": {
     "duration": 0.049384,
     "end_time": "2020-11-09T11:35:17.419214",
     "exception": false,
     "start_time": "2020-11-09T11:35:17.369830",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "train_X = df[:len(train_X)]\n",
    "normalized_test_df = df[len(train_X):]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-11-09T11:35:17.473574Z",
     "iopub.status.busy": "2020-11-09T11:35:17.471654Z",
     "iopub.status.idle": "2020-11-09T11:35:17.476349Z",
     "shell.execute_reply": "2020-11-09T11:35:17.475547Z"
    },
    "papermill": {
     "duration": 0.037658,
     "end_time": "2020-11-09T11:35:17.476516",
     "exception": false,
     "start_time": "2020-11-09T11:35:17.438858",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 3526227 entries, 0 to 3526226\n",
      "Columns: 103 entries, ISLEM_ADEDI to season_yaz\n",
      "dtypes: float64(22), int64(2), uint8(79)\n",
      "memory usage: 938.2 MB\n"
     ]
    }
   ],
   "source": [
    "train_X.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-11-09T11:35:17.533466Z",
     "iopub.status.busy": "2020-11-09T11:35:17.532819Z",
     "iopub.status.idle": "2020-11-09T11:35:19.605327Z",
     "shell.execute_reply": "2020-11-09T11:35:19.604218Z"
    },
    "papermill": {
     "duration": 2.10681,
     "end_time": "2020-11-09T11:35:19.605480",
     "exception": false,
     "start_time": "2020-11-09T11:35:17.498670",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# cross-validation with 5 splits\n",
    "\n",
    "cv = ShuffleSplit(n_splits=5, test_size=0.2, random_state=0)\n",
    "\n",
    "#Importing XGBM Classifier \n",
    "from xgboost import XGBRegressor\n",
    "\n",
    "#creating an extreme Gradient boosting instance\n",
    "xgb = XGBRegressor(\n",
    "                   #early_stopping_rounds=100,\n",
    "                   random_state=0, \n",
    "                   tree_method='gpu_hist', \n",
    "                   verbosity=0\n",
    "                )\n",
    "\n",
    "# parameters \n",
    "parameters_overfit = {\n",
    "                \"n_estimators\": [100, 300, 500],\n",
    "                \"max_depth\": [10, 12, 16],\n",
    "                \"subsample\": [0.4, 0.5, 0.7],\n",
    "                \"reg_lambda\":[1, 5, 10],\n",
    "                \"colsample_bytree\": [0.4, 0.6, 1],\n",
    "                \"max_bin\":[32, 64,128],\n",
    "                'max_leaves': [15, 31, 63],\n",
    "                \"min_child_weight\":[1,5,10],\n",
    "                \n",
    "}\n",
    "# parameters \n",
    "parameters = {\n",
    "                \"n_estimators\": [100, 300, 500],\n",
    "                \"max_depth\": [4, 5, 7],\n",
    "                \"subsample\": [0.4, 0.5, 0.7],\n",
    "                \"reg_lambda\":[1, 5, 10],\n",
    "                \"reg_alpha\": [1, 5, 10],\n",
    "                \"colsample_bytree\": [0.4, 0.5, 0.7],\n",
    "                \"max_bin\":[32, 64,128],\n",
    "                'max_leaves': [15, 31, 63],\n",
    "                \"min_child_weight\":[1,5,10],\n",
    "                \n",
    "}\n",
    "\n",
    "xgb_overfit = XGBRegressor(\n",
    "                        random_state=0, \n",
    "                        tree_method='gpu_hist', \n",
    "                        nthread=-1,\n",
    "                        verbosity=1,\n",
    "                        n_estimators=1000,\n",
    "                        max_depth=16,\n",
    "                        subsample=1,\n",
    "                        reg_lambda=1,\n",
    "                        colsample_bytree=0.4,\n",
    "                        max_bin=128,\n",
    "                        max_leaves=255,\n",
    "                        min_child_weight=1,\n",
    "                        learning_rate=0.001,\n",
    "                       )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-11-09T11:35:19.662282Z",
     "iopub.status.busy": "2020-11-09T11:35:19.661388Z",
     "iopub.status.idle": "2020-11-09T15:58:45.741589Z",
     "shell.execute_reply": "2020-11-09T15:58:45.742136Z"
    },
    "papermill": {
     "duration": 15806.116358,
     "end_time": "2020-11-09T15:58:45.742302",
     "exception": false,
     "start_time": "2020-11-09T11:35:19.625944",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The best parameters are {'subsample': 0.7, 'reg_lambda': 5, 'reg_alpha': 5, 'n_estimators': 500, 'min_child_weight': 1, 'max_leaves': 63, 'max_depth': 7, 'max_bin': 128, 'colsample_bytree': 0.7} with a score of 0.6710\n"
     ]
    }
   ],
   "source": [
    "# grid search for parameters\n",
    "grid = RandomizedSearchCV(estimator=xgb, param_distributions=parameters, cv=cv, n_iter=100)\n",
    "grid.fit(train_X, train_y)\n",
    "\n",
    "# print best scores\n",
    "print(\"The best parameters are %s with a score of %0.4f\"\n",
    "      % (grid.best_params_, grid.best_score_))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-11-09T15:58:45.789516Z",
     "iopub.status.busy": "2020-11-09T15:58:45.788715Z",
     "iopub.status.idle": "2020-11-09T15:59:46.009623Z",
     "shell.execute_reply": "2020-11-09T15:59:46.010771Z"
    },
    "papermill": {
     "duration": 60.247924,
     "end_time": "2020-11-09T15:59:46.010945",
     "exception": false,
     "start_time": "2020-11-09T15:58:45.763021",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# extend n_estimators to 1000\n",
    "best_params_dict=grid.best_estimator_.get_params()\n",
    "best_params_dict[\"n_estimators\"]=1000\n",
    "extented_xgb = XGBRegressor().set_params(**best_params_dict)\n",
    "extented_xgb.fit(train_X, train_y)\n",
    "\n",
    "# prediction results\n",
    "y_pred = extented_xgb.predict(normalized_test_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-11-09T15:59:46.071991Z",
     "iopub.status.busy": "2020-11-09T15:59:46.071322Z",
     "iopub.status.idle": "2020-11-09T15:59:46.092317Z",
     "shell.execute_reply": "2020-11-09T15:59:46.093214Z"
    },
    "papermill": {
     "duration": 0.055807,
     "end_time": "2020-11-09T15:59:46.093348",
     "exception": false,
     "start_time": "2020-11-09T15:59:46.037541",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "213690"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(np.unique(y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-11-09T15:59:46.150494Z",
     "iopub.status.busy": "2020-11-09T15:59:46.149829Z",
     "iopub.status.idle": "2020-11-09T15:59:46.523195Z",
     "shell.execute_reply": "2020-11-09T15:59:46.522705Z"
    },
    "papermill": {
     "duration": 0.404057,
     "end_time": "2020-11-09T15:59:46.523300",
     "exception": false,
     "start_time": "2020-11-09T15:59:46.119243",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([1.00000e+00, 0.00000e+00, 0.00000e+00, 1.00000e+00, 1.00000e+00,\n",
       "        3.00000e+00, 7.10000e+01, 2.15811e+05, 2.45200e+03, 6.24000e+02,\n",
       "        2.76000e+02, 1.50000e+02, 8.70000e+01, 4.10000e+01, 3.00000e+01,\n",
       "        3.00000e+01, 1.30000e+01, 1.50000e+01, 6.00000e+00, 9.00000e+00,\n",
       "        6.00000e+00, 3.00000e+00, 2.00000e+00, 4.00000e+00, 3.00000e+00,\n",
       "        2.00000e+00, 3.00000e+00, 3.00000e+00, 2.00000e+00, 0.00000e+00,\n",
       "        1.00000e+00, 0.00000e+00, 1.00000e+00, 0.00000e+00, 1.00000e+00,\n",
       "        0.00000e+00, 1.00000e+00, 1.00000e+00, 1.00000e+00, 2.00000e+00,\n",
       "        1.00000e+00, 0.00000e+00, 0.00000e+00, 0.00000e+00, 1.00000e+00,\n",
       "        0.00000e+00, 0.00000e+00, 0.00000e+00, 0.00000e+00, 1.00000e+00,\n",
       "        1.00000e+00, 0.00000e+00, 0.00000e+00, 0.00000e+00, 0.00000e+00,\n",
       "        0.00000e+00, 1.00000e+00, 0.00000e+00, 0.00000e+00, 0.00000e+00,\n",
       "        0.00000e+00, 0.00000e+00, 0.00000e+00, 0.00000e+00, 0.00000e+00,\n",
       "        0.00000e+00, 0.00000e+00, 0.00000e+00, 0.00000e+00, 0.00000e+00,\n",
       "        0.00000e+00, 0.00000e+00, 0.00000e+00, 0.00000e+00, 0.00000e+00,\n",
       "        0.00000e+00, 0.00000e+00, 0.00000e+00, 0.00000e+00, 0.00000e+00,\n",
       "        0.00000e+00, 0.00000e+00, 0.00000e+00, 0.00000e+00, 0.00000e+00,\n",
       "        0.00000e+00, 0.00000e+00, 0.00000e+00, 0.00000e+00, 0.00000e+00,\n",
       "        0.00000e+00, 0.00000e+00, 0.00000e+00, 0.00000e+00, 0.00000e+00,\n",
       "        0.00000e+00, 0.00000e+00, 0.00000e+00, 0.00000e+00, 1.00000e+00]),\n",
       " array([-32996.45  , -28522.545 , -24048.639 , -19574.734 , -15100.829 ,\n",
       "        -10626.924 ,  -6153.0186,  -1679.1134,   2794.7917,   7268.697 ,\n",
       "         11742.602 ,  16216.507 ,  20690.412 ,  25164.316 ,  29638.223 ,\n",
       "         34112.13  ,  38586.03  ,  43059.938 ,  47533.844 ,  52007.746 ,\n",
       "         56481.652 ,  60955.56  ,  65429.465 ,  69903.37  ,  74377.27  ,\n",
       "         78851.18  ,  83325.086 ,  87798.99  ,  92272.89  ,  96746.8   ,\n",
       "        101220.7   , 105694.61  , 110168.516 , 114642.42  , 119116.33  ,\n",
       "        123590.23  , 128064.13  , 132538.05  , 137011.94  , 141485.84  ,\n",
       "        145959.75  , 150433.66  , 154907.56  , 159381.47  , 163855.38  ,\n",
       "        168329.28  , 172803.19  , 177277.1   , 181751.    , 186224.9   ,\n",
       "        190698.81  , 195172.72  , 199646.61  , 204120.52  , 208594.42  ,\n",
       "        213068.33  , 217542.23  , 222016.14  , 226490.05  , 230963.95  ,\n",
       "        235437.86  , 239911.77  , 244385.67  , 248859.58  , 253333.48  ,\n",
       "        257807.39  , 262281.28  , 266755.2   , 271229.1   , 275703.    ,\n",
       "        280176.9   , 284650.8   , 289124.72  , 293598.62  , 298072.53  ,\n",
       "        302546.44  , 307020.34  , 311494.25  , 315968.16  , 320442.06  ,\n",
       "        324915.97  , 329389.88  , 333863.78  , 338337.7   , 342811.6   ,\n",
       "        347285.5   , 351759.4   , 356233.28  , 360707.2   , 365181.1   ,\n",
       "        369655.    , 374128.9   , 378602.8   , 383076.72  , 387550.62  ,\n",
       "        392024.53  , 396498.44  , 400972.34  , 405446.25  , 409920.16  ,\n",
       "        414394.06  ], dtype=float32),\n",
       " <a list of 100 Patch objects>)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(y_pred, bins=100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-11-09T15:59:46.574720Z",
     "iopub.status.busy": "2020-11-09T15:59:46.572653Z",
     "iopub.status.idle": "2020-11-09T15:59:46.577742Z",
     "shell.execute_reply": "2020-11-09T15:59:46.578186Z"
    },
    "papermill": {
     "duration": 0.031974,
     "end_time": "2020-11-09T15:59:46.578299",
     "exception": false,
     "start_time": "2020-11-09T15:59:46.546325",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2926"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sum(y_pred<=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-11-09T15:59:46.631179Z",
     "iopub.status.busy": "2020-11-09T15:59:46.630438Z",
     "iopub.status.idle": "2020-11-09T15:59:49.369157Z",
     "shell.execute_reply": "2020-11-09T15:59:49.368373Z"
    },
    "papermill": {
     "duration": 2.767734,
     "end_time": "2020-11-09T15:59:49.369318",
     "exception": false,
     "start_time": "2020-11-09T15:59:46.601584",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f7950149990>"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x1152 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "importance_df=pd.DataFrame(zip(normalized_test_df.columns.tolist(),extented_xgb.feature_importances_),columns=[\"feature\",\"importance\"])\n",
    "\n",
    "import seaborn as sns\n",
    "plt.figure(figsize=(16, 16))\n",
    "sns.barplot(data=importance_df.sort_values(by='importance', ascending=False), x='importance', y='feature')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-11-09T15:59:49.457288Z",
     "iopub.status.busy": "2020-11-09T15:59:49.456465Z",
     "iopub.status.idle": "2020-11-09T15:59:52.401559Z",
     "shell.execute_reply": "2020-11-09T15:59:52.400996Z"
    },
    "papermill": {
     "duration": 2.995127,
     "end_time": "2020-11-09T15:59:52.401677",
     "exception": false,
     "start_time": "2020-11-09T15:59:49.406550",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f78f666c850>"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from xgboost import plot_tree\n",
    "plot_tree(extented_xgb, num_trees=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-11-09T15:59:52.468167Z",
     "iopub.status.busy": "2020-11-09T15:59:52.466147Z",
     "iopub.status.idle": "2020-11-09T15:59:55.113794Z",
     "shell.execute_reply": "2020-11-09T15:59:55.112871Z"
    },
    "papermill": {
     "duration": 2.685023,
     "end_time": "2020-11-09T15:59:55.113909",
     "exception": false,
     "start_time": "2020-11-09T15:59:52.428886",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f78f62a73d0>"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_tree(extented_xgb, num_trees=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-11-09T15:59:55.192767Z",
     "iopub.status.busy": "2020-11-09T15:59:55.191626Z",
     "iopub.status.idle": "2020-11-09T15:59:57.487707Z",
     "shell.execute_reply": "2020-11-09T15:59:57.486760Z"
    },
    "papermill": {
     "duration": 2.344863,
     "end_time": "2020-11-09T15:59:57.487820",
     "exception": false,
     "start_time": "2020-11-09T15:59:55.142957",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f78f49f4990>"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_tree(extented_xgb, num_trees=50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-11-09T15:59:57.559462Z",
     "iopub.status.busy": "2020-11-09T15:59:57.554402Z",
     "iopub.status.idle": "2020-11-09T16:00:00.044260Z",
     "shell.execute_reply": "2020-11-09T16:00:00.044961Z"
    },
    "papermill": {
     "duration": 2.526924,
     "end_time": "2020-11-09T16:00:00.045302",
     "exception": false,
     "start_time": "2020-11-09T15:59:57.518378",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f78f4a9cbd0>"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_tree(extented_xgb, num_trees=100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-11-09T16:00:00.390157Z",
     "iopub.status.busy": "2020-11-09T16:00:00.389183Z",
     "iopub.status.idle": "2020-11-09T16:00:00.556790Z",
     "shell.execute_reply": "2020-11-09T16:00:00.556179Z"
    },
    "papermill": {
     "duration": 0.480303,
     "end_time": "2020-11-09T16:00:00.556910",
     "exception": false,
     "start_time": "2020-11-09T16:00:00.076607",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# prediction results\n",
    "pred_df=pd.DataFrame(zip(df_test1['sector_id'], y_pred), columns=['sector_id', 'predictions'])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-11-09T16:00:00.627383Z",
     "iopub.status.busy": "2020-11-09T16:00:00.626504Z",
     "iopub.status.idle": "2020-11-09T16:00:00.636294Z",
     "shell.execute_reply": "2020-11-09T16:00:00.635787Z"
    },
    "papermill": {
     "duration": 0.048054,
     "end_time": "2020-11-09T16:00:00.636390",
     "exception": false,
     "start_time": "2020-11-09T16:00:00.588336",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "submission_dff=pred_df.groupby('sector_id').sum()\n",
    "ss=submission_dff.reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-11-09T16:00:00.704101Z",
     "iopub.status.busy": "2020-11-09T16:00:00.703368Z",
     "iopub.status.idle": "2020-11-09T16:00:00.707054Z",
     "shell.execute_reply": "2020-11-09T16:00:00.706593Z"
    },
    "papermill": {
     "duration": 0.039464,
     "end_time": "2020-11-09T16:00:00.707149",
     "exception": false,
     "start_time": "2020-11-09T16:00:00.667685",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "ss=ss.rename(columns={\"sector_id\": \"Sektor_Id\", \"predictions\": \"Expected\"})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-11-09T16:00:00.775155Z",
     "iopub.status.busy": "2020-11-09T16:00:00.774288Z",
     "iopub.status.idle": "2020-11-09T16:00:01.819468Z",
     "shell.execute_reply": "2020-11-09T16:00:01.818910Z"
    },
    "papermill": {
     "duration": 1.081395,
     "end_time": "2020-11-09T16:00:01.819581",
     "exception": false,
     "start_time": "2020-11-09T16:00:00.738186",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "ss.to_csv('out_random_search_xgboost.csv',index=False)\n",
    "pred_df.to_csv('pred_df_random_search_xg', index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "papermill": {
     "duration": 0.031811,
     "end_time": "2020-11-09T16:00:01.883984",
     "exception": false,
     "start_time": "2020-11-09T16:00:01.852173",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.7.6"
  },
  "papermill": {
   "duration": 15924.663295,
   "end_time": "2020-11-09T16:00:03.279099",
   "environment_variables": {},
   "exception": null,
   "input_path": "__notebook__.ipynb",
   "output_path": "__notebook__.ipynb",
   "parameters": {},
   "start_time": "2020-11-09T11:34:38.615804",
   "version": "2.1.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}