model-06-build-lgb-search-lr.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Import packages"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/lightgbm/__init__.py:45: FutureWarning: Starting from version 2.2.1, the library file in distribution wheels for macOS will be built by the Apple Clang compiler.\n",
      "This means that in case of installing LightGBM from PyPI via the ``pip install lightgbm`` command, you won't need to install the gcc compiler anymore.\n",
      "Instead of that, you'll need to install the OpenMP library, which is required for running LightGBM on the system with the Apple Clang compiler.\n",
      "You can install the OpenMP library by the following command: ``brew install libomp``.\n",
      "  \"You can install the OpenMP library by the following command: ``brew install libomp``.\", FutureWarning)\n"
     ]
    }
   ],
   "source": [
    "import time\n",
    "import pickle\n",
    "import itertools\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import lightgbm as lgb\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.model_selection import KFold\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import mean_squared_error\n",
    "from sklearn.metrics import mean_absolute_error\n",
    "from utilities import cal_mape\n",
    "\n",
    "from utilities import cal_score, convert_types, create_dt_feats"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Params"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "is_log_target = True\n",
    "is_per_area_target = True\n",
    "\n",
    "map_target = {\n",
    "    (True, True):   'log_per_area_total_price',\n",
    "    (True, False):  'log_total_price',\n",
    "    (False, True):  'per_area_total_price',\n",
    "    (False, False): 'total_price'\n",
    "}\n",
    "col_target = map_target[(is_log_target, is_per_area_target)]\n",
    "\n",
    "#cols_feats_add = ['encoded_latlon', 'building_per_land_area'] # add feats into model\n",
    "cols_num_add = ['building_per_land_area']\n",
    "cols_cat_add = ['encoded_latlon']\n",
    "\n",
    "#cols_feats_rm = [] # TODO: implement before using it\n",
    "#['day_in_year_txn_dt', 'day_in_year_building_complete_dt', 'dt_diff_building_complete_dt_txn_dt'] # dt num feats\n",
    "#['dow_txn_dt', 'dow_building_complete_dt', 'month_txn_dt', 'month_building_complete_dt',\n",
    "# 'quarter_txn_dt', 'quarter_building_complete_dt', 'year_txn_dt', 'year_building_complete_dt'] # dt cat feats"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Processing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def pre_processing(df, cols_num, cols_cat): #, cols_feats\n",
    "    # Convert types\n",
    "    df[cols_num] = df[cols_num].astype('float32')\n",
    "    \n",
    "    # generate target\n",
    "    if 'total_price' in df:\n",
    "        df['total_price'] = df['total_price'].astype('float32')\n",
    "        df['log_total_price'] = np.log1p(df['total_price'])\n",
    "        df['per_area_total_price'] = df['total_price'] / df['building_area']\n",
    "        df['log_per_area_total_price'] = np.log1p(df['per_area_total_price'])\n",
    "        df['target'] = df[col_target]\n",
    "\n",
    "    # generate lat/lon encoded\n",
    "    kmeans = pickle.load(open('./object_kmeans_k6_combined.pkl', 'rb'))\n",
    "    df['encoded_latlon'] = kmeans.predict(df[['lat','lon']])\n",
    "    \n",
    "    # generate dt feats\n",
    "    df = create_dt_feats(df, 'txn_dt')\n",
    "    df = create_dt_feats(df, 'building_complete_dt')\n",
    "    df['dt_diff_building_complete_dt_txn_dt'] = df['txn_dt'] - df['building_complete_dt']\n",
    "    \n",
    "    # generate building to land ratio\n",
    "    df['building_per_land_area'] = df['building_area'] / df['land_area']\n",
    " \n",
    "    #cols_feats = cols_num + cols_cat + cols_feats_add\n",
    "    \n",
    "    return df  #, cols_feats"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "from feature_engineering import CategoricalColumnsEncoder\n",
    "from feature_engineering import TargetMeanEncoding\n",
    "from feature_engineering import MultiLabelEncoding\n",
    "\n",
    "class FeatureEngineering():\n",
    "    def __init__(self, cols_num, cols_cat, cols_feats, col_target, n_ran_encode=4):\n",
    "        # cols\n",
    "        self.cols_num = cols_num\n",
    "        self.cols_cat = cols_cat\n",
    "        self.cols_feats = cols_feats\n",
    "        self.col_target = col_target\n",
    "        \n",
    "        # Numbers\n",
    "        self.n_ran_encode = n_ran_encode\n",
    "        \n",
    "        # encoders\n",
    "        #self.tar_encoder = TargetMeanEncoding(col_target = self.col_target)\n",
    "        #self.ran_encoder = MultiLabelEncoding(self.n_ran_encode)\n",
    "        self.cat_astyper = CategoricalColumnsEncoder(mode='pandas')\n",
    "        \n",
    "    def fit_transform(self, df):\n",
    "        df = df.copy()\n",
    "    \n",
    "#        encoded1 = self.tar_encoder.fit_transform(df[self.cols_cat], df[self.col_target])\n",
    "#        encoded2 = self.ran_encoder.fit_transform(df[self.cols_cat])\n",
    "        self.cat_astyper.fit_transform(df, self.cols_cat)\n",
    "    \n",
    "        return df\n",
    "#        return pd.concat([df, encoded1, encoded2], axis=1)\n",
    "    \n",
    "    def transform(self, df):\n",
    "        df = df.copy()\n",
    "#        encoded1 = self.tar_encoder.transform(df[self.cols_cat])\n",
    "#        encoded2 = self.ran_encoder.transform(df[self.cols_cat])     # TODO: add generated feats into cols_feats\n",
    "        self.cat_astyper.transform(df)\n",
    "\n",
    "        return df\n",
    "#        return pd.concat([df, encoded1, encoded2], axis=1)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def post_processing(y_pred, df):\n",
    "    \"\"\"Args: \n",
    "          y_pred (1d np.array): raw prediction y\n",
    "          df (pd.Dataframe): input feature df\n",
    "       Return: np.array\n",
    "    \"\"\"\n",
    "    y_pred_final = y_pred.copy()\n",
    "    \n",
    "    if is_log_target:\n",
    "        y_pred_final = np.expm1(y_pred_final)\n",
    "        y_pred_final = np.clip(y_pred_final, 0, None)\n",
    "\n",
    "    if is_per_area_target:\n",
    "        y_pred_final = y_pred_final * df['building_area'].values\n",
    "        \n",
    "    return y_pred_final"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "cols_num = ['txn_dt', 'total_floor', 'building_complete_dt', 'parking_area', 'parking_price', 'txn_floor',\n",
    "            'land_area', 'building_area', 'lat', 'lon', 'village_income_median', 'town_population', 'town_area',\n",
    "            'town_population_density', 'doc_rate', 'master_rate', 'bachelor_rate', 'jobschool_rate',\n",
    "            'highschool_rate', 'junior_rate', 'elementary_rate', 'born_rate', 'death_rate', 'marriage_rate',\n",
    "            'divorce_rate', 'N_50', 'N_500', 'N_1000', 'N_5000', 'N_10000', 'I_10', 'I_50', 'I_100', 'I_250',\n",
    "            'I_500', 'I_1000', 'I_5000', 'I_10000', 'I_MIN', 'II_10', 'II_50', 'II_100', 'II_250', 'II_500',\n",
    "            'II_1000', 'II_5000', 'II_10000', 'II_MIN', 'III_10', 'III_50', 'III_100', 'III_250', 'III_500',\n",
    "            'III_1000', 'III_5000', 'III_10000', 'III_MIN', 'IV_10', 'IV_50', 'IV_100', 'IV_250', 'IV_500',\n",
    "            'IV_1000', 'IV_5000', 'IV_10000', 'IV_MIN', 'V_10', 'V_50', 'V_100', 'V_250', 'V_500', 'V_1000',\n",
    "            'V_5000', 'V_10000', 'V_MIN', 'VI_50', 'VI_100', 'VI_250', 'VI_500', 'VI_1000', 'VI_5000', 'VI_10000',\n",
    "            'VI_MIN', 'VII_10', 'VII_50', 'VII_100', 'VII_250', 'VII_500', 'VII_1000', 'VII_5000', 'VII_10000',\n",
    "            'VII_MIN', 'VIII_10', 'VIII_50', 'VIII_100', 'VIII_250', 'VIII_500', 'VIII_1000', 'VIII_5000',\n",
    "            'VIII_10000', 'VIII_MIN', 'IX_10', 'IX_50', 'IX_100', 'IX_250', 'IX_500', 'IX_1000', 'IX_5000',\n",
    "            'IX_10000', 'IX_MIN', 'X_10', 'X_50', 'X_100', 'X_250', 'X_500', 'X_1000', 'X_5000', 'X_10000', 'X_MIN',\n",
    "            'XI_10', 'XI_50', 'XI_100', 'XI_250', 'XI_500', 'XI_1000', 'XI_5000', 'XI_10000', 'XI_MIN', 'XII_10',\n",
    "            'XII_50', 'XII_100', 'XII_250', 'XII_500', 'XII_1000', 'XII_5000', 'XII_10000', 'XII_MIN', 'XIII_10',\n",
    "            'XIII_50', 'XIII_100', 'XIII_250', 'XIII_500', 'XIII_1000', 'XIII_5000', 'XIII_10000', 'XIII_MIN',\n",
    "            'XIV_10', 'XIV_50', 'XIV_100', 'XIV_250', 'XIV_500', 'XIV_1000', 'XIV_5000', 'XIV_10000', 'XIV_MIN']\n",
    "cols_cat = ['building_material', 'city', 'building_type', 'building_use', 'parking_way', 'town', 'village',\n",
    "            'I_index_50', 'I_index_500', 'I_index_1000', 'II_index_50', 'II_index_500', 'II_index_1000',\n",
    "            'III_index_50', 'III_index_500', 'III_index_1000', 'IV_index_50', 'IV_index_500', 'IV_index_1000',\n",
    "            'IV_index_5000', 'V_index_50', 'V_index_500', 'V_index_1000', 'VI_10', 'VI_index_50', 'VI_index_500',\n",
    "            'VI_index_1000', 'VII_index_50', 'VII_index_500', 'VII_index_1000', 'VIII_index_50', 'VIII_index_500',\n",
    "            'VIII_index_1000', 'IX_index_50', 'IX_index_500', 'IX_index_1000', 'IX_index_5000', 'X_index_50',\n",
    "            'X_index_500', 'X_index_1000', 'XI_index_50', 'XI_index_500', 'XI_index_1000', 'XI_index_5000',\n",
    "            'XI_index_10000', 'XII_index_50', 'XII_index_500', 'XII_index_1000', 'XIII_index_50', 'XIII_index_500',\n",
    "            'XIII_index_1000', 'XIII_index_5000', 'XIII_index_10000', 'XIV_index_50', 'XIV_index_500',\n",
    "            'XIV_index_1000']\n",
    "cols_feats = cols_num + cols_cat\n",
    "\n",
    "# informational cols\n",
    "cols_dt = ['txn_dt', 'building_complete_dt']\n",
    "cols_bin = ['I_index_50', 'I_index_500', 'I_index_1000', 'II_index_50', 'II_index_500', 'II_index_1000',\n",
    "            'III_index_50', 'III_index_500', 'III_index_1000', 'IV_index_50', 'IV_index_500', 'IV_index_1000',\n",
    "            'IV_index_5000', 'V_index_50', 'V_index_500', 'V_index_1000', 'VI_10', 'VI_index_50', 'VI_index_500',\n",
    "            'VI_index_1000', 'VII_index_50', 'VII_index_500', 'VII_index_1000', 'VIII_index_50', 'VIII_index_500',\n",
    "            'VIII_index_1000', 'IX_index_50', 'IX_index_500', 'IX_index_1000', 'IX_index_5000', 'X_index_50',\n",
    "            'X_index_500', 'X_index_1000', 'XI_index_50', 'XI_index_500', 'XI_index_1000', 'XI_index_5000',\n",
    "            'XI_index_10000', 'XII_index_50', 'XII_index_500', 'XII_index_1000', 'XIII_index_50', 'XIII_index_500',\n",
    "            'XIII_index_1000', 'XIII_index_5000', 'XIII_index_10000', 'XIV_index_50', 'XIV_index_500',\n",
    "            'XIV_index_1000']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Read data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>building_id</th>\n",
       "      <th>building_material</th>\n",
       "      <th>city</th>\n",
       "      <th>txn_dt</th>\n",
       "      <th>total_floor</th>\n",
       "      <th>building_type</th>\n",
       "      <th>building_use</th>\n",
       "      <th>building_complete_dt</th>\n",
       "      <th>parking_way</th>\n",
       "      <th>parking_area</th>\n",
       "      <th>...</th>\n",
       "      <th>XIV_500</th>\n",
       "      <th>XIV_index_500</th>\n",
       "      <th>XIV_1000</th>\n",
       "      <th>XIV_index_1000</th>\n",
       "      <th>XIV_5000</th>\n",
       "      <th>XIV_index_5000</th>\n",
       "      <th>XIV_10000</th>\n",
       "      <th>XIV_index_10000</th>\n",
       "      <th>XIV_MIN</th>\n",
       "      <th>total_price</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>e3mMIMR3JJqCaXz1</td>\n",
       "      <td>8</td>\n",
       "      <td>21</td>\n",
       "      <td>18674</td>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>6271</td>\n",
       "      <td>2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>58</td>\n",
       "      <td>1</td>\n",
       "      <td>157</td>\n",
       "      <td>1</td>\n",
       "      <td>2483</td>\n",
       "      <td>1</td>\n",
       "      <td>6011</td>\n",
       "      <td>1</td>\n",
       "      <td>34.469802848528126</td>\n",
       "      <td>647603.7704719775</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>LgwzgklNvy4QCtq5</td>\n",
       "      <td>8</td>\n",
       "      <td>7</td>\n",
       "      <td>18800</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>7885</td>\n",
       "      <td>2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>28</td>\n",
       "      <td>1</td>\n",
       "      <td>115</td>\n",
       "      <td>1</td>\n",
       "      <td>15872</td>\n",
       "      <td>1</td>\n",
       "      <td>32221</td>\n",
       "      <td>1</td>\n",
       "      <td>40.07357348973039</td>\n",
       "      <td>3321452.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>ucIR2NLLsC3T650L</td>\n",
       "      <td>8</td>\n",
       "      <td>7</td>\n",
       "      <td>19289</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>6028</td>\n",
       "      <td>2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>78</td>\n",
       "      <td>1</td>\n",
       "      <td>212</td>\n",
       "      <td>1</td>\n",
       "      <td>15760</td>\n",
       "      <td>1</td>\n",
       "      <td>32228</td>\n",
       "      <td>1</td>\n",
       "      <td>54.46208108983661</td>\n",
       "      <td>9570885.316206373</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>jre1pJhcQj91Kdky</td>\n",
       "      <td>8</td>\n",
       "      <td>21</td>\n",
       "      <td>20385</td>\n",
       "      <td>24</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>18325</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>20</td>\n",
       "      <td>1</td>\n",
       "      <td>125</td>\n",
       "      <td>1</td>\n",
       "      <td>2568</td>\n",
       "      <td>1</td>\n",
       "      <td>7271</td>\n",
       "      <td>1</td>\n",
       "      <td>99.62896580763416</td>\n",
       "      <td>14215011.127100045</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>rQpYpY9nRG7X5mmr</td>\n",
       "      <td>1</td>\n",
       "      <td>21</td>\n",
       "      <td>20657</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "      <td>6880</td>\n",
       "      <td>2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>18</td>\n",
       "      <td>1</td>\n",
       "      <td>47</td>\n",
       "      <td>1</td>\n",
       "      <td>2587</td>\n",
       "      <td>1</td>\n",
       "      <td>7442</td>\n",
       "      <td>1</td>\n",
       "      <td>124.1312362968124</td>\n",
       "      <td>762711.9757135696</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 235 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        building_id building_material city txn_dt total_floor building_type  \\\n",
       "0  e3mMIMR3JJqCaXz1                 8   21  18674           4             3   \n",
       "1  LgwzgklNvy4QCtq5                 8    7  18800           5             1   \n",
       "2  ucIR2NLLsC3T650L                 8    7  19289           4             1   \n",
       "3  jre1pJhcQj91Kdky                 8   21  20385          24             0   \n",
       "4  rQpYpY9nRG7X5mmr                 1   21  20657           2             4   \n",
       "\n",
       "  building_use building_complete_dt parking_way parking_area  \\\n",
       "0            2                 6271           2          NaN   \n",
       "1            2                 7885           2          NaN   \n",
       "2            2                 6028           2          NaN   \n",
       "3            2                18325           0          NaN   \n",
       "4            2                 6880           2          NaN   \n",
       "\n",
       "          ...         XIV_500 XIV_index_500 XIV_1000 XIV_index_1000 XIV_5000  \\\n",
       "0         ...              58             1      157              1     2483   \n",
       "1         ...              28             1      115              1    15872   \n",
       "2         ...              78             1      212              1    15760   \n",
       "3         ...              20             1      125              1     2568   \n",
       "4         ...              18             1       47              1     2587   \n",
       "\n",
       "  XIV_index_5000 XIV_10000 XIV_index_10000             XIV_MIN  \\\n",
       "0              1      6011               1  34.469802848528126   \n",
       "1              1     32221               1   40.07357348973039   \n",
       "2              1     32228               1   54.46208108983661   \n",
       "3              1      7271               1   99.62896580763416   \n",
       "4              1      7442               1   124.1312362968124   \n",
       "\n",
       "          total_price  \n",
       "0   647603.7704719775  \n",
       "1           3321452.0  \n",
       "2   9570885.316206373  \n",
       "3  14215011.127100045  \n",
       "4   762711.9757135696  \n",
       "\n",
       "[5 rows x 235 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv('dataset/train.csv', dtype=object)\n",
    "df.head(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/sklearn/base.py:251: UserWarning: Trying to unpickle estimator KMeans from version 0.20.3 when using version 0.20.1. This might lead to breaking code or invalid results. Use at your own risk.\n",
      "  UserWarning)\n"
     ]
    }
   ],
   "source": [
    "df = pre_processing(df, cols_num, cols_cat)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "cols_num = cols_num + [c for c in cols_num_add if c not in cols_num]\n",
    "cols_cat = cols_cat + [c for c in cols_cat_add if c not in cols_cat]\n",
    "cols_feats = cols_num + cols_cat"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Grid search"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# grid search\n",
    "params_fix = {'task': 'train',\n",
    "              'boosting_type': 'gbdt',\n",
    "              'objective': 'mse',\n",
    "              'metric': 'mape',}\n",
    "#              'learning_rate': 0.01 }\n",
    "\n",
    "lgb_other_params = {'num_boost_round': 500000,\n",
    "                    'verbose_eval': 2000,\n",
    "                    'early_stopping_rounds': 1000 }\n",
    "\n",
    "# prvs baseline best estimator\n",
    "# ('objective', 'mse'), ('metric', 'mape'),\n",
    "# ('learning_rate', 0.015), ('num_leaves', 255),\n",
    "# ('feature_fraction', 0.5), ('min_data_in_leaf', 5),\n",
    "# ('lambda_l1', 0.1), ('lambda_l2', 0.1)),\n",
    "# 5524.863045737181,\n",
    "# [5528.862755759428, 5530.862894274216, 5514.863487177898]]\n",
    "\n",
    "#params_gsearch = {'max_depth':[-1, 8],\n",
    "#                  'num_leaves': [200, 255]}\n",
    "\n",
    "params_gsearch = {'learning_rate': [0.1, 0.07, 0.05, 0.03, 0.01, 0.007, 0.005, 0.003],\n",
    "                  'max_depth': [16],\n",
    "                  'num_leaves': [300],\n",
    "                  'feature_fraction': [0.5],\n",
    "                  'min_data_in_leaf': [10],\n",
    "                  'lambda_l1': [0.01],\n",
    "                  'lambda_l2': [0.1]\n",
    "                 }\n",
    "\n",
    "#params_gsearch = {'max_depth': [-1, 8, 16, 24],\n",
    "#                  'num_leaves': [200, 255, 300],           # may reduce in dim-reduction exp\n",
    "#                  'feature_fraction': [0.5, 0.75, 1.0],\n",
    "#                  'min_data_in_leaf': [5, 10, 20],\n",
    "#                  'lambda_l1': [0, 0.01, 0.1],\n",
    "#                  'lambda_l2': [0, 0.01, 0.1],\n",
    "#                 }\n",
    "N_fold = 3\n",
    "N_iter = 8\n",
    "score_low_bound = 5600 # if score smaller than this, skip the param set"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Round 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Round 1\n",
    "rand_params_set = []\n",
    "for i in range(N_iter):\n",
    "    params_rand = {key: np.random.choice(params_gsearch[key]) for key in params_gsearch}\n",
    "    while params_rand in rand_params_set:\n",
    "        params_rand = {key: np.random.choice(params_gsearch[key]) for key in params_gsearch}\n",
    "    rand_params_set.append(params_rand)\n",
    "gsearch = {}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "#rand_params_set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "==== Fold 1 ====\n",
      "Training until validation scores don't improve for 1000 rounds.\n",
      "[2000]\tvalid_0's mape: 0.00971187\n",
      "[4000]\tvalid_0's mape: 0.00952097\n",
      "[6000]\tvalid_0's mape: 0.00947285\n",
      "[8000]\tvalid_0's mape: 0.00945633\n",
      "[10000]\tvalid_0's mape: 0.00944843\n",
      "[12000]\tvalid_0's mape: 0.00944395\n",
      "[14000]\tvalid_0's mape: 0.00944167\n",
      "[16000]\tvalid_0's mape: 0.0094401\n",
      "[18000]\tvalid_0's mape: 0.0094391\n",
      "[20000]\tvalid_0's mape: 0.00943832\n",
      "[22000]\tvalid_0's mape: 0.00943779\n",
      "[24000]\tvalid_0's mape: 0.00943744\n",
      "[26000]\tvalid_0's mape: 0.00943733\n",
      "Early stopping, best iteration is:\n",
      "[25049]\tvalid_0's mape: 0.00943733\n",
      "params: {'task': 'train', 'boosting_type': 'gbdt', 'objective': 'mse', 'metric': 'mape', 'learning_rate': 0.01, 'max_depth': 16, 'num_leaves': 300, 'feature_fraction': 0.5, 'min_data_in_leaf': 10, 'lambda_l1': 0.01, 'lambda_l2': 0.1}\n",
      "score, mse, mae, mape: 5767.86923, 0.03853, 0.12549, 0.00943.\n",
      "Training until validation scores don't improve for 1000 rounds.\n",
      "[2000]\tvalid_0's mape: 0.0106538\n",
      "[4000]\tvalid_0's mape: 0.00995857\n",
      "[6000]\tvalid_0's mape: 0.00972948\n",
      "[8000]\tvalid_0's mape: 0.00961535\n",
      "[10000]\tvalid_0's mape: 0.00954961\n",
      "[12000]\tvalid_0's mape: 0.00950767\n",
      "[14000]\tvalid_0's mape: 0.00947989\n",
      "[16000]\tvalid_0's mape: 0.00946103\n",
      "[18000]\tvalid_0's mape: 0.00944874\n",
      "[20000]\tvalid_0's mape: 0.00943993\n",
      "[22000]\tvalid_0's mape: 0.00943344\n",
      "[24000]\tvalid_0's mape: 0.00942863\n",
      "[26000]\tvalid_0's mape: 0.00942461\n",
      "[28000]\tvalid_0's mape: 0.00942137\n",
      "[30000]\tvalid_0's mape: 0.00941896\n",
      "[32000]\tvalid_0's mape: 0.00941703\n",
      "[34000]\tvalid_0's mape: 0.0094153\n",
      "[36000]\tvalid_0's mape: 0.0094138\n",
      "[38000]\tvalid_0's mape: 0.0094125\n",
      "[40000]\tvalid_0's mape: 0.00941137\n",
      "[42000]\tvalid_0's mape: 0.00941039\n",
      "[44000]\tvalid_0's mape: 0.00940962\n",
      "[46000]\tvalid_0's mape: 0.009409\n",
      "[48000]\tvalid_0's mape: 0.00940832\n",
      "[50000]\tvalid_0's mape: 0.00940781\n",
      "[52000]\tvalid_0's mape: 0.0094074\n",
      "[54000]\tvalid_0's mape: 0.00940698\n",
      "[56000]\tvalid_0's mape: 0.0094066\n",
      "[58000]\tvalid_0's mape: 0.00940625\n",
      "[60000]\tvalid_0's mape: 0.009406\n",
      "[62000]\tvalid_0's mape: 0.00940578\n",
      "[64000]\tvalid_0's mape: 0.00940554\n",
      "[66000]\tvalid_0's mape: 0.00940536\n",
      "[68000]\tvalid_0's mape: 0.0094052\n",
      "[70000]\tvalid_0's mape: 0.00940503\n",
      "[72000]\tvalid_0's mape: 0.00940488\n",
      "[74000]\tvalid_0's mape: 0.00940475\n",
      "[76000]\tvalid_0's mape: 0.00940463\n",
      "[78000]\tvalid_0's mape: 0.00940453\n",
      "[80000]\tvalid_0's mape: 0.00940448\n",
      "[82000]\tvalid_0's mape: 0.00940444\n",
      "Early stopping, best iteration is:\n",
      "[81323]\tvalid_0's mape: 0.00940444\n",
      "params: {'task': 'train', 'boosting_type': 'gbdt', 'objective': 'mse', 'metric': 'mape', 'learning_rate': 0.003, 'max_depth': 16, 'num_leaves': 300, 'feature_fraction': 0.5, 'min_data_in_leaf': 10, 'lambda_l1': 0.01, 'lambda_l2': 0.1}\n",
      "score, mse, mae, mape: 5760.86973, 0.03824, 0.12506, 0.00940.\n",
      "Training until validation scores don't improve for 1000 rounds.\n",
      "[2000]\tvalid_0's mape: 0.00955048\n",
      "[4000]\tvalid_0's mape: 0.00952073\n",
      "[6000]\tvalid_0's mape: 0.00951583\n",
      "[8000]\tvalid_0's mape: 0.00951423\n",
      "Early stopping, best iteration is:\n",
      "[8742]\tvalid_0's mape: 0.00951397\n",
      "params: {'task': 'train', 'boosting_type': 'gbdt', 'objective': 'mse', 'metric': 'mape', 'learning_rate': 0.03, 'max_depth': 16, 'num_leaves': 300, 'feature_fraction': 0.5, 'min_data_in_leaf': 10, 'lambda_l1': 0.01, 'lambda_l2': 0.1}\n",
      "score, mse, mae, mape: 5729.86817, 0.03907, 0.12650, 0.00951.\n",
      "Training until validation scores don't improve for 1000 rounds.\n",
      "[2000]\tvalid_0's mape: 0.0098718\n",
      "[4000]\tvalid_0's mape: 0.00958185\n",
      "[6000]\tvalid_0's mape: 0.0094941\n",
      "[8000]\tvalid_0's mape: 0.00945903\n",
      "[10000]\tvalid_0's mape: 0.00944223\n",
      "[12000]\tvalid_0's mape: 0.00943385\n",
      "[14000]\tvalid_0's mape: 0.00942833\n",
      "[16000]\tvalid_0's mape: 0.00942447\n",
      "[18000]\tvalid_0's mape: 0.00942192\n",
      "[20000]\tvalid_0's mape: 0.00942007\n",
      "[22000]\tvalid_0's mape: 0.0094187\n",
      "[24000]\tvalid_0's mape: 0.00941768\n",
      "[26000]\tvalid_0's mape: 0.00941697\n",
      "[28000]\tvalid_0's mape: 0.00941639\n",
      "[30000]\tvalid_0's mape: 0.00941598\n",
      "[32000]\tvalid_0's mape: 0.00941566\n",
      "[34000]\tvalid_0's mape: 0.00941545\n",
      "Early stopping, best iteration is:\n",
      "[34816]\tvalid_0's mape: 0.00941538\n",
      "params: {'task': 'train', 'boosting_type': 'gbdt', 'objective': 'mse', 'metric': 'mape', 'learning_rate': 0.007, 'max_depth': 16, 'num_leaves': 300, 'feature_fraction': 0.5, 'min_data_in_leaf': 10, 'lambda_l1': 0.01, 'lambda_l2': 0.1}\n",
      "score, mse, mae, mape: 5756.86956, 0.03841, 0.12518, 0.00941.\n",
      "Training until validation scores don't improve for 1000 rounds.\n",
      "[2000]\tvalid_0's mape: 0.0101011\n",
      "[4000]\tvalid_0's mape: 0.00968896\n",
      "[6000]\tvalid_0's mape: 0.009555\n",
      "[8000]\tvalid_0's mape: 0.00949212\n",
      "[10000]\tvalid_0's mape: 0.00946018\n",
      "[12000]\tvalid_0's mape: 0.00944391\n",
      "[14000]\tvalid_0's mape: 0.00943264\n",
      "[16000]\tvalid_0's mape: 0.00942558\n",
      "[18000]\tvalid_0's mape: 0.00942074\n",
      "[20000]\tvalid_0's mape: 0.00941716\n",
      "[22000]\tvalid_0's mape: 0.00941476\n",
      "[24000]\tvalid_0's mape: 0.00941319\n",
      "[26000]\tvalid_0's mape: 0.00941174\n",
      "[28000]\tvalid_0's mape: 0.0094106\n",
      "[30000]\tvalid_0's mape: 0.00940964\n",
      "[32000]\tvalid_0's mape: 0.00940893\n",
      "[34000]\tvalid_0's mape: 0.00940829\n",
      "[36000]\tvalid_0's mape: 0.00940776\n",
      "[38000]\tvalid_0's mape: 0.00940744\n",
      "[40000]\tvalid_0's mape: 0.00940707\n",
      "[42000]\tvalid_0's mape: 0.0094068\n",
      "[44000]\tvalid_0's mape: 0.00940661\n",
      "[46000]\tvalid_0's mape: 0.00940643\n",
      "[48000]\tvalid_0's mape: 0.0094063\n",
      "[50000]\tvalid_0's mape: 0.00940618\n",
      "Early stopping, best iteration is:\n",
      "[50229]\tvalid_0's mape: 0.00940617\n",
      "params: {'task': 'train', 'boosting_type': 'gbdt', 'objective': 'mse', 'metric': 'mape', 'learning_rate': 0.005, 'max_depth': 16, 'num_leaves': 300, 'feature_fraction': 0.5, 'min_data_in_leaf': 10, 'lambda_l1': 0.01, 'lambda_l2': 0.1}\n",
      "score, mse, mae, mape: 5764.86972, 0.03833, 0.12508, 0.00940.\n",
      "Training until validation scores don't improve for 1000 rounds.\n",
      "[2000]\tvalid_0's mape: 0.00957238\n",
      "[4000]\tvalid_0's mape: 0.00956263\n",
      "[6000]\tvalid_0's mape: 0.00956142\n",
      "Early stopping, best iteration is:\n",
      "[5556]\tvalid_0's mape: 0.00956142\n",
      "params: {'task': 'train', 'boosting_type': 'gbdt', 'objective': 'mse', 'metric': 'mape', 'learning_rate': 0.05, 'max_depth': 16, 'num_leaves': 300, 'feature_fraction': 0.5, 'min_data_in_leaf': 10, 'lambda_l1': 0.01, 'lambda_l2': 0.1}\n",
      "score, mse, mae, mape: 5704.86741, 0.03940, 0.12712, 0.00955.\n",
      "Training until validation scores don't improve for 1000 rounds.\n",
      "[2000]\tvalid_0's mape: 0.00980767\n",
      "[4000]\tvalid_0's mape: 0.00980586\n",
      "Early stopping, best iteration is:\n",
      "[3110]\tvalid_0's mape: 0.00980586\n",
      "params: {'task': 'train', 'boosting_type': 'gbdt', 'objective': 'mse', 'metric': 'mape', 'learning_rate': 0.1, 'max_depth': 16, 'num_leaves': 300, 'feature_fraction': 0.5, 'min_data_in_leaf': 10, 'lambda_l1': 0.01, 'lambda_l2': 0.1}\n",
      "score, mse, mae, mape: 5554.86431, 0.04076, 0.13040, 0.00980.\n",
      "Training until validation scores don't improve for 1000 rounds.\n",
      "[2000]\tvalid_0's mape: 0.00966466\n",
      "[4000]\tvalid_0's mape: 0.00965937\n",
      "Early stopping, best iteration is:\n",
      "[4088]\tvalid_0's mape: 0.00965933\n",
      "params: {'task': 'train', 'boosting_type': 'gbdt', 'objective': 'mse', 'metric': 'mape', 'learning_rate': 0.07, 'max_depth': 16, 'num_leaves': 300, 'feature_fraction': 0.5, 'min_data_in_leaf': 10, 'lambda_l1': 0.01, 'lambda_l2': 0.1}\n",
      "score, mse, mae, mape: 5625.86616, 0.03979, 0.12843, 0.00965.\n",
      "==== Fold 2 ====\n",
      "Training until validation scores don't improve for 1000 rounds.\n",
      "[2000]\tvalid_0's mape: 0.00973492\n",
      "[4000]\tvalid_0's mape: 0.00955267\n",
      "[6000]\tvalid_0's mape: 0.00950321\n",
      "[8000]\tvalid_0's mape: 0.00948393\n",
      "[10000]\tvalid_0's mape: 0.00947474\n",
      "[12000]\tvalid_0's mape: 0.00946941\n",
      "[14000]\tvalid_0's mape: 0.00946636\n",
      "[16000]\tvalid_0's mape: 0.0094648\n",
      "[18000]\tvalid_0's mape: 0.00946362\n",
      "[20000]\tvalid_0's mape: 0.00946286\n",
      "[22000]\tvalid_0's mape: 0.00946238\n",
      "[24000]\tvalid_0's mape: 0.00946206\n",
      "[26000]\tvalid_0's mape: 0.00946197\n",
      "Early stopping, best iteration is:\n",
      "[25031]\tvalid_0's mape: 0.00946197\n",
      "params: {'task': 'train', 'boosting_type': 'gbdt', 'objective': 'mse', 'metric': 'mape', 'learning_rate': 0.01, 'max_depth': 16, 'num_leaves': 300, 'feature_fraction': 0.5, 'min_data_in_leaf': 10, 'lambda_l1': 0.01, 'lambda_l2': 0.1}\n",
      "score, mse, mae, mape: 5754.87062, 0.04007, 0.12605, 0.00945.\n",
      "Training until validation scores don't improve for 1000 rounds.\n",
      "[2000]\tvalid_0's mape: 0.010594\n",
      "[4000]\tvalid_0's mape: 0.00996317\n",
      "[6000]\tvalid_0's mape: 0.00974526\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[8000]\tvalid_0's mape: 0.00963913\n",
      "[10000]\tvalid_0's mape: 0.00958073\n",
      "[12000]\tvalid_0's mape: 0.00954211\n",
      "[14000]\tvalid_0's mape: 0.00951639\n",
      "[16000]\tvalid_0's mape: 0.00949814\n",
      "[18000]\tvalid_0's mape: 0.00948465\n",
      "[20000]\tvalid_0's mape: 0.00947565\n",
      "[22000]\tvalid_0's mape: 0.0094686\n",
      "[24000]\tvalid_0's mape: 0.00946291\n",
      "[26000]\tvalid_0's mape: 0.00945861\n",
      "[28000]\tvalid_0's mape: 0.00945518\n",
      "[30000]\tvalid_0's mape: 0.00945254\n",
      "[32000]\tvalid_0's mape: 0.00945072\n",
      "[34000]\tvalid_0's mape: 0.0094487\n",
      "[36000]\tvalid_0's mape: 0.00944656\n",
      "[38000]\tvalid_0's mape: 0.00944473\n",
      "[40000]\tvalid_0's mape: 0.00944355\n",
      "[42000]\tvalid_0's mape: 0.00944254\n",
      "[44000]\tvalid_0's mape: 0.00944159\n",
      "[46000]\tvalid_0's mape: 0.00944084\n",
      "[48000]\tvalid_0's mape: 0.00944007\n",
      "[50000]\tvalid_0's mape: 0.00943945\n",
      "[52000]\tvalid_0's mape: 0.00943904\n",
      "[54000]\tvalid_0's mape: 0.00943867\n",
      "[56000]\tvalid_0's mape: 0.00943833\n",
      "[58000]\tvalid_0's mape: 0.00943802\n",
      "[60000]\tvalid_0's mape: 0.0094377\n",
      "[62000]\tvalid_0's mape: 0.00943744\n",
      "[64000]\tvalid_0's mape: 0.00943723\n",
      "[66000]\tvalid_0's mape: 0.00943703\n",
      "[68000]\tvalid_0's mape: 0.00943685\n",
      "[70000]\tvalid_0's mape: 0.0094367\n",
      "[72000]\tvalid_0's mape: 0.00943659\n",
      "[74000]\tvalid_0's mape: 0.00943648\n",
      "[76000]\tvalid_0's mape: 0.00943638\n",
      "[78000]\tvalid_0's mape: 0.00943631\n",
      "[80000]\tvalid_0's mape: 0.00943623\n",
      "[82000]\tvalid_0's mape: 0.00943619\n",
      "Early stopping, best iteration is:\n",
      "[81906]\tvalid_0's mape: 0.00943619\n",
      "params: {'task': 'train', 'boosting_type': 'gbdt', 'objective': 'mse', 'metric': 'mape', 'learning_rate': 0.003, 'max_depth': 16, 'num_leaves': 300, 'feature_fraction': 0.5, 'min_data_in_leaf': 10, 'lambda_l1': 0.01, 'lambda_l2': 0.1}\n",
      "score, mse, mae, mape: 5772.87093, 0.03996, 0.12569, 0.00943.\n",
      "Training until validation scores don't improve for 1000 rounds.\n",
      "[2000]\tvalid_0's mape: 0.00959889\n",
      "[4000]\tvalid_0's mape: 0.00956758\n",
      "[6000]\tvalid_0's mape: 0.0095615\n",
      "[8000]\tvalid_0's mape: 0.0095598\n",
      "Early stopping, best iteration is:\n",
      "[8887]\tvalid_0's mape: 0.00955955\n",
      "params: {'task': 'train', 'boosting_type': 'gbdt', 'objective': 'mse', 'metric': 'mape', 'learning_rate': 0.03, 'max_depth': 16, 'num_leaves': 300, 'feature_fraction': 0.5, 'min_data_in_leaf': 10, 'lambda_l1': 0.01, 'lambda_l2': 0.1}\n",
      "score, mse, mae, mape: 5694.86929, 0.04079, 0.12737, 0.00955.\n",
      "Training until validation scores don't improve for 1000 rounds.\n",
      "[2000]\tvalid_0's mape: 0.0098919\n",
      "[4000]\tvalid_0's mape: 0.0096155\n",
      "[6000]\tvalid_0's mape: 0.00953518\n",
      "[8000]\tvalid_0's mape: 0.00950033\n",
      "[10000]\tvalid_0's mape: 0.00948264\n",
      "[12000]\tvalid_0's mape: 0.00947268\n",
      "[14000]\tvalid_0's mape: 0.0094671\n",
      "[16000]\tvalid_0's mape: 0.00946355\n",
      "[18000]\tvalid_0's mape: 0.00946104\n",
      "[20000]\tvalid_0's mape: 0.00945913\n",
      "[22000]\tvalid_0's mape: 0.00945803\n",
      "[24000]\tvalid_0's mape: 0.00945708\n",
      "[26000]\tvalid_0's mape: 0.00945636\n",
      "[28000]\tvalid_0's mape: 0.00945581\n",
      "[30000]\tvalid_0's mape: 0.00945544\n",
      "[32000]\tvalid_0's mape: 0.00945513\n",
      "[34000]\tvalid_0's mape: 0.00945495\n",
      "[36000]\tvalid_0's mape: 0.00945481\n",
      "Early stopping, best iteration is:\n",
      "[35768]\tvalid_0's mape: 0.00945481\n",
      "params: {'task': 'train', 'boosting_type': 'gbdt', 'objective': 'mse', 'metric': 'mape', 'learning_rate': 0.007, 'max_depth': 16, 'num_leaves': 300, 'feature_fraction': 0.5, 'min_data_in_leaf': 10, 'lambda_l1': 0.01, 'lambda_l2': 0.1}\n",
      "score, mse, mae, mape: 5738.87069, 0.04004, 0.12595, 0.00945.\n",
      "Training until validation scores don't improve for 1000 rounds.\n",
      "[2000]\tvalid_0's mape: 0.0101032\n",
      "[4000]\tvalid_0's mape: 0.00971369\n",
      "[6000]\tvalid_0's mape: 0.00959125\n",
      "[8000]\tvalid_0's mape: 0.00953426\n",
      "[10000]\tvalid_0's mape: 0.00950368\n",
      "[12000]\tvalid_0's mape: 0.0094849\n",
      "[14000]\tvalid_0's mape: 0.00947324\n",
      "[16000]\tvalid_0's mape: 0.00946563\n",
      "[18000]\tvalid_0's mape: 0.0094609\n",
      "[20000]\tvalid_0's mape: 0.00945738\n",
      "[22000]\tvalid_0's mape: 0.00945472\n",
      "[24000]\tvalid_0's mape: 0.00945255\n",
      "[26000]\tvalid_0's mape: 0.00945083\n",
      "[28000]\tvalid_0's mape: 0.00944956\n",
      "[30000]\tvalid_0's mape: 0.00944862\n",
      "[32000]\tvalid_0's mape: 0.0094479\n",
      "[34000]\tvalid_0's mape: 0.00944717\n",
      "[36000]\tvalid_0's mape: 0.00944665\n",
      "[38000]\tvalid_0's mape: 0.00944621\n",
      "[40000]\tvalid_0's mape: 0.00944588\n",
      "[42000]\tvalid_0's mape: 0.00944557\n",
      "[44000]\tvalid_0's mape: 0.00944536\n",
      "[46000]\tvalid_0's mape: 0.00944517\n",
      "[48000]\tvalid_0's mape: 0.00944504\n",
      "[50000]\tvalid_0's mape: 0.00944496\n",
      "Early stopping, best iteration is:\n",
      "[49632]\tvalid_0's mape: 0.00944496\n",
      "params: {'task': 'train', 'boosting_type': 'gbdt', 'objective': 'mse', 'metric': 'mape', 'learning_rate': 0.005, 'max_depth': 16, 'num_leaves': 300, 'feature_fraction': 0.5, 'min_data_in_leaf': 10, 'lambda_l1': 0.01, 'lambda_l2': 0.1}\n",
      "score, mse, mae, mape: 5763.87084, 0.04001, 0.12580, 0.00944.\n",
      "Training until validation scores don't improve for 1000 rounds.\n",
      "[2000]\tvalid_0's mape: 0.00962348\n",
      "[4000]\tvalid_0's mape: 0.00961353\n",
      "[6000]\tvalid_0's mape: 0.00961233\n",
      "Early stopping, best iteration is:\n",
      "[5504]\tvalid_0's mape: 0.00961233\n",
      "params: {'task': 'train', 'boosting_type': 'gbdt', 'objective': 'mse', 'metric': 'mape', 'learning_rate': 0.05, 'max_depth': 16, 'num_leaves': 300, 'feature_fraction': 0.5, 'min_data_in_leaf': 10, 'lambda_l1': 0.01, 'lambda_l2': 0.1}\n",
      "score, mse, mae, mape: 5692.86858, 0.04097, 0.12809, 0.00961.\n",
      "Training until validation scores don't improve for 1000 rounds.\n",
      "[2000]\tvalid_0's mape: 0.00971804\n",
      "[4000]\tvalid_0's mape: 0.00971237\n",
      "Early stopping, best iteration is:\n",
      "[4221]\tvalid_0's mape: 0.0097123\n",
      "params: {'task': 'train', 'boosting_type': 'gbdt', 'objective': 'mse', 'metric': 'mape', 'learning_rate': 0.07, 'max_depth': 16, 'num_leaves': 300, 'feature_fraction': 0.5, 'min_data_in_leaf': 10, 'lambda_l1': 0.01, 'lambda_l2': 0.1}\n",
      "score, mse, mae, mape: 5616.86735, 0.04180, 0.12940, 0.00970.\n",
      "==== Fold 3 ====\n",
      "Training until validation scores don't improve for 1000 rounds.\n",
      "[2000]\tvalid_0's mape: 0.00980517\n",
      "[4000]\tvalid_0's mape: 0.00962068\n",
      "[6000]\tvalid_0's mape: 0.00957016\n",
      "[8000]\tvalid_0's mape: 0.00955058\n",
      "[10000]\tvalid_0's mape: 0.00954035\n",
      "[12000]\tvalid_0's mape: 0.00953489\n",
      "[14000]\tvalid_0's mape: 0.00953185\n",
      "[16000]\tvalid_0's mape: 0.00953028\n",
      "[18000]\tvalid_0's mape: 0.00952923\n",
      "[20000]\tvalid_0's mape: 0.00952855\n",
      "[22000]\tvalid_0's mape: 0.00952808\n",
      "[24000]\tvalid_0's mape: 0.00952784\n",
      "Early stopping, best iteration is:\n",
      "[23941]\tvalid_0's mape: 0.00952784\n",
      "params: {'task': 'train', 'boosting_type': 'gbdt', 'objective': 'mse', 'metric': 'mape', 'learning_rate': 0.01, 'max_depth': 16, 'num_leaves': 300, 'feature_fraction': 0.5, 'min_data_in_leaf': 10, 'lambda_l1': 0.01, 'lambda_l2': 0.1}\n",
      "score, mse, mae, mape: 5711.86726, 0.04006, 0.12667, 0.00952.\n",
      "Training until validation scores don't improve for 1000 rounds.\n",
      "[2000]\tvalid_0's mape: 0.0107465\n",
      "[4000]\tvalid_0's mape: 0.0100499\n",
      "[6000]\tvalid_0's mape: 0.0098199\n",
      "[8000]\tvalid_0's mape: 0.0097116\n",
      "[10000]\tvalid_0's mape: 0.00964848\n",
      "[12000]\tvalid_0's mape: 0.00960805\n",
      "[14000]\tvalid_0's mape: 0.00958137\n",
      "[16000]\tvalid_0's mape: 0.00956251\n",
      "[18000]\tvalid_0's mape: 0.00954941\n",
      "[20000]\tvalid_0's mape: 0.00953977\n",
      "[22000]\tvalid_0's mape: 0.00953239\n",
      "[24000]\tvalid_0's mape: 0.00952732\n",
      "[26000]\tvalid_0's mape: 0.00952258\n",
      "[28000]\tvalid_0's mape: 0.00951877\n",
      "[30000]\tvalid_0's mape: 0.00951575\n",
      "[32000]\tvalid_0's mape: 0.00951329\n",
      "[34000]\tvalid_0's mape: 0.00951129\n",
      "[36000]\tvalid_0's mape: 0.00950957\n",
      "[38000]\tvalid_0's mape: 0.00950848\n",
      "[40000]\tvalid_0's mape: 0.00950729\n",
      "[42000]\tvalid_0's mape: 0.00950615\n",
      "[44000]\tvalid_0's mape: 0.00950528\n",
      "[46000]\tvalid_0's mape: 0.00950436\n",
      "[48000]\tvalid_0's mape: 0.00950375\n",
      "[50000]\tvalid_0's mape: 0.00950321\n",
      "[52000]\tvalid_0's mape: 0.00950264\n",
      "[54000]\tvalid_0's mape: 0.00950214\n",
      "[56000]\tvalid_0's mape: 0.00950175\n",
      "[58000]\tvalid_0's mape: 0.00950141\n",
      "[60000]\tvalid_0's mape: 0.00950113\n",
      "[62000]\tvalid_0's mape: 0.00950087\n",
      "[64000]\tvalid_0's mape: 0.00950062\n",
      "[66000]\tvalid_0's mape: 0.00950044\n",
      "[68000]\tvalid_0's mape: 0.00950027\n",
      "[70000]\tvalid_0's mape: 0.00950016\n",
      "[72000]\tvalid_0's mape: 0.00950001\n",
      "[74000]\tvalid_0's mape: 0.00949991\n",
      "[76000]\tvalid_0's mape: 0.00949982\n",
      "[78000]\tvalid_0's mape: 0.00949975\n",
      "[80000]\tvalid_0's mape: 0.0094997\n",
      "[82000]\tvalid_0's mape: 0.00949967\n",
      "Early stopping, best iteration is:\n",
      "[81284]\tvalid_0's mape: 0.00949967\n",
      "params: {'task': 'train', 'boosting_type': 'gbdt', 'objective': 'mse', 'metric': 'mape', 'learning_rate': 0.003, 'max_depth': 16, 'num_leaves': 300, 'feature_fraction': 0.5, 'min_data_in_leaf': 10, 'lambda_l1': 0.01, 'lambda_l2': 0.1}\n",
      "score, mse, mae, mape: 5743.86761, 0.03987, 0.12629, 0.00950.\n",
      "Training until validation scores don't improve for 1000 rounds.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[2000]\tvalid_0's mape: 0.0096626\n",
      "[4000]\tvalid_0's mape: 0.00963002\n",
      "[6000]\tvalid_0's mape: 0.00962412\n",
      "[8000]\tvalid_0's mape: 0.00962251\n",
      "Early stopping, best iteration is:\n",
      "[8695]\tvalid_0's mape: 0.0096223\n",
      "params: {'task': 'train', 'boosting_type': 'gbdt', 'objective': 'mse', 'metric': 'mape', 'learning_rate': 0.03, 'max_depth': 16, 'num_leaves': 300, 'feature_fraction': 0.5, 'min_data_in_leaf': 10, 'lambda_l1': 0.01, 'lambda_l2': 0.1}\n",
      "score, mse, mae, mape: 5669.86588, 0.04071, 0.12787, 0.00962.\n",
      "Training until validation scores don't improve for 1000 rounds.\n",
      "[2000]\tvalid_0's mape: 0.00996592\n",
      "[4000]\tvalid_0's mape: 0.00968257\n",
      "[6000]\tvalid_0's mape: 0.00960197\n",
      "[8000]\tvalid_0's mape: 0.00956819\n",
      "[10000]\tvalid_0's mape: 0.00954956\n",
      "[12000]\tvalid_0's mape: 0.00953995\n",
      "[14000]\tvalid_0's mape: 0.00953436\n",
      "[16000]\tvalid_0's mape: 0.00953043\n",
      "[18000]\tvalid_0's mape: 0.0095274\n",
      "[20000]\tvalid_0's mape: 0.00952561\n",
      "[22000]\tvalid_0's mape: 0.00952419\n",
      "[24000]\tvalid_0's mape: 0.00952327\n",
      "[26000]\tvalid_0's mape: 0.00952262\n",
      "[28000]\tvalid_0's mape: 0.00952204\n",
      "[30000]\tvalid_0's mape: 0.00952165\n",
      "[32000]\tvalid_0's mape: 0.00952137\n",
      "[34000]\tvalid_0's mape: 0.00952121\n",
      "Early stopping, best iteration is:\n",
      "[34439]\tvalid_0's mape: 0.00952119\n",
      "params: {'task': 'train', 'boosting_type': 'gbdt', 'objective': 'mse', 'metric': 'mape', 'learning_rate': 0.007, 'max_depth': 16, 'num_leaves': 300, 'feature_fraction': 0.5, 'min_data_in_leaf': 10, 'lambda_l1': 0.01, 'lambda_l2': 0.1}\n",
      "score, mse, mae, mape: 5724.86731, 0.04006, 0.12658, 0.00952.\n",
      "Training until validation scores don't improve for 1000 rounds.\n",
      "[2000]\tvalid_0's mape: 0.0102113\n",
      "[4000]\tvalid_0's mape: 0.00979483\n",
      "[6000]\tvalid_0's mape: 0.00966828\n",
      "[8000]\tvalid_0's mape: 0.00960772\n",
      "[10000]\tvalid_0's mape: 0.00957706\n",
      "[12000]\tvalid_0's mape: 0.00955871\n",
      "[14000]\tvalid_0's mape: 0.00954663\n",
      "[16000]\tvalid_0's mape: 0.00953909\n",
      "[18000]\tvalid_0's mape: 0.00953378\n",
      "[20000]\tvalid_0's mape: 0.00952965\n",
      "[22000]\tvalid_0's mape: 0.00952685\n",
      "[24000]\tvalid_0's mape: 0.00952441\n",
      "[26000]\tvalid_0's mape: 0.00952272\n",
      "[28000]\tvalid_0's mape: 0.00952134\n",
      "[30000]\tvalid_0's mape: 0.00952031\n",
      "[32000]\tvalid_0's mape: 0.00951946\n",
      "[34000]\tvalid_0's mape: 0.00951882\n",
      "[36000]\tvalid_0's mape: 0.00951834\n",
      "[38000]\tvalid_0's mape: 0.00951792\n",
      "[40000]\tvalid_0's mape: 0.00951757\n",
      "[42000]\tvalid_0's mape: 0.00951733\n",
      "[44000]\tvalid_0's mape: 0.00951715\n",
      "[46000]\tvalid_0's mape: 0.00951704\n",
      "[48000]\tvalid_0's mape: 0.00951697\n",
      "Early stopping, best iteration is:\n",
      "[48057]\tvalid_0's mape: 0.00951697\n",
      "params: {'task': 'train', 'boosting_type': 'gbdt', 'objective': 'mse', 'metric': 'mape', 'learning_rate': 0.005, 'max_depth': 16, 'num_leaves': 300, 'feature_fraction': 0.5, 'min_data_in_leaf': 10, 'lambda_l1': 0.01, 'lambda_l2': 0.1}\n",
      "score, mse, mae, mape: 5724.86732, 0.04004, 0.12651, 0.00951.\n",
      "Training until validation scores don't improve for 1000 rounds.\n",
      "[2000]\tvalid_0's mape: 0.00969479\n",
      "[4000]\tvalid_0's mape: 0.00968325\n",
      "[6000]\tvalid_0's mape: 0.00968218\n",
      "Early stopping, best iteration is:\n",
      "[5515]\tvalid_0's mape: 0.00968218\n",
      "params: {'task': 'train', 'boosting_type': 'gbdt', 'objective': 'mse', 'metric': 'mape', 'learning_rate': 0.05, 'max_depth': 16, 'num_leaves': 300, 'feature_fraction': 0.5, 'min_data_in_leaf': 10, 'lambda_l1': 0.01, 'lambda_l2': 0.1}\n",
      "score, mse, mae, mape: 5638.86482, 0.04105, 0.12874, 0.00968.\n",
      "Training until validation scores don't improve for 1000 rounds.\n",
      "[2000]\tvalid_0's mape: 0.00978592\n",
      "[4000]\tvalid_0's mape: 0.00978037\n",
      "Early stopping, best iteration is:\n",
      "[4192]\tvalid_0's mape: 0.00978032\n",
      "params: {'task': 'train', 'boosting_type': 'gbdt', 'objective': 'mse', 'metric': 'mape', 'learning_rate': 0.07, 'max_depth': 16, 'num_leaves': 300, 'feature_fraction': 0.5, 'min_data_in_leaf': 10, 'lambda_l1': 0.01, 'lambda_l2': 0.1}\n",
      "score, mse, mae, mape: 5614.86361, 0.04145, 0.13006, 0.00978.\n"
     ]
    }
   ],
   "source": [
    "t0 = time.time()\n",
    "gsearch = {}\n",
    "folds = KFold(n_splits=N_fold, shuffle=True, random_state=123)\n",
    "for i_fold, (itrain, ival) in enumerate(folds.split(df)): # kfold\n",
    "    print('==== Fold', i_fold+1, '====')\n",
    "    \n",
    "    # split train, val\n",
    "    df_train = df.loc[itrain]\n",
    "    df_val = df.loc[ival]\n",
    "    \n",
    "    # feat eng\n",
    "    feat_eng = FeatureEngineering(cols_num, cols_cat, cols_feats, col_target)\n",
    "    df_train = feat_eng.fit_transform(df_train)\n",
    "    df_val = feat_eng.transform(df_val)\n",
    "\n",
    "    # Construct lgb dataset\n",
    "    lgb_train = lgb.Dataset(df_train[cols_feats], df_train[col_target]).construct()\n",
    "    lgb_val = lgb.Dataset(df_val[cols_feats], df_val[col_target], reference=lgb_train).construct()\n",
    "        \n",
    "    # random sample - grid search\n",
    "    rand_params_next = []\n",
    "    for params_rand in rand_params_set:\n",
    "        params = params_fix.copy()\n",
    "        params.update( params_rand )\n",
    "\n",
    "        model = lgb.train(params, lgb_train, valid_sets=lgb_val, keep_training_booster=True,\n",
    "                          **lgb_other_params)\n",
    "\n",
    "        y_pred = model.predict(df_val[cols_feats])\n",
    "        mse = mean_squared_error(df_val[col_target], y_pred)\n",
    "        mae = mean_absolute_error(df_val[col_target], y_pred)\n",
    "        mape = cal_mape(df_val[col_target], y_pred)\n",
    "        \n",
    "        y_pred_final = post_processing(y_pred, df_val)\n",
    "        score = cal_score(df_val['total_price'], y_pred_final)\n",
    "\n",
    "        print('params:', params)\n",
    "        print('score, mse, mae, mape: %.5f, %.5f, %.5f, %.5f.' % (score, mse, mae, mape))\n",
    "\n",
    "        if score < score_low_bound:\n",
    "            continue\n",
    "\n",
    "        rand_params_next.append(params_rand)\n",
    "        tuple_params = tuple(params.items())\n",
    "        if tuple_params not in gsearch:\n",
    "            gsearch[tuple_params] = [[], [], [], [], []]\n",
    "        gsearch[tuple_params][0].append(model.best_iteration)\n",
    "        gsearch[tuple_params][1].append(score)\n",
    "        gsearch[tuple_params][2].append(mse)\n",
    "        gsearch[tuple_params][3].append(mae)\n",
    "        gsearch[tuple_params][4].append(mape)\n",
    "        \n",
    "    rand_params_set = rand_params_next\n",
    "# aggregate, sort gsearch results\n",
    "gsearch_results = [[key, *[np.mean(val_i) for val_i in val[1:]], *val] for key, val in gsearch.items()]\n",
    "gsearch_results.sort(key= lambda x: x[1], reverse=True)\n",
    "t1 = time.time()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total computing time: 16049.273500204086\n"
     ]
    }
   ],
   "source": [
    "print('Total computing time:', t1-t0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "TODO: add lr= 0.001 results here (see how to compare build and predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(('task', 'train'), ('boosting_type', 'gbdt'), ('objective', 'mse'), ('metric', 'mape'), ('learning_rate', 0.003), ('max_depth', 16), ('num_leaves', 300), ('feature_fraction', 0.5), ('min_data_in_leaf', 10), ('lambda_l1', 0.01), ('lambda_l2', 0.1)) 5759.202754425594 0.03935629228200634 0.1256802345112175 0.00944115814180837 [5760.86972615993, 5772.870928085513, 5743.867609031339] [0.03824138609184463, 0.03995908448171842, 0.03986840627245597] [0.12505532226656882, 0.12569174191163468, 0.12629363935544902] [0.00940003315361107, 0.009427576036866952, 0.009495865234947087]\n",
      "(('task', 'train'), ('boosting_type', 'gbdt'), ('objective', 'mse'), ('metric', 'mape'), ('learning_rate', 0.005), ('max_depth', 16), ('num_leaves', 300), ('feature_fraction', 0.5), ('min_data_in_leaf', 10), ('lambda_l1', 0.01), ('lambda_l2', 0.1)) 5751.202629683488 0.03946038448176132 0.1258000685199044 0.009450552656604409 [5764.869723639806, 5763.870840786203, 5724.867324624458] [0.03832758935234754, 0.04001257062937591, 0.04004099346356053] [0.12508459339145825, 0.12580150073126928, 0.1265141114369857] [0.009402448330182177, 0.009436091196589329, 0.009513118443041717]\n",
      "(('task', 'train'), ('boosting_type', 'gbdt'), ('objective', 'mse'), ('metric', 'mape'), ('learning_rate', 0.01), ('max_depth', 16), ('num_leaves', 300), ('feature_fraction', 0.5), ('min_data_in_leaf', 10), ('lambda_l1', 0.01), ('lambda_l2', 0.1)) 5744.86903680761 0.03955470668753225 0.12606760439027134 0.009470130539728958 [5767.869227330606, 5754.870622634057, 5711.86726045817] [0.03853063368951811, 0.04007467974145371, 0.04005880663162492] [0.12548899344325026, 0.12604780984689457, 0.12666600988066917] [0.00943296670105554, 0.009453368113494795, 0.009524056804636537]\n",
      "(('task', 'train'), ('boosting_type', 'gbdt'), ('objective', 'mse'), ('metric', 'mape'), ('learning_rate', 0.007), ('max_depth', 16), ('num_leaves', 300), ('feature_fraction', 0.5), ('min_data_in_leaf', 10), ('lambda_l1', 0.01), ('lambda_l2', 0.1)) 5740.202524692392 0.03950265581374023 0.1259055084574817 0.009458288345624398 [5756.869564111103, 5738.870694974607, 5724.8673149914675] [0.0384083975184408, 0.04003614473864248, 0.04006342518413743] [0.12518139891056254, 0.12595088944856644, 0.12658423701331603] [0.00941061167650365, 0.009446107514660568, 0.009518145845708975]\n",
      "(('task', 'train'), ('boosting_type', 'gbdt'), ('objective', 'mse'), ('metric', 'mape'), ('learning_rate', 0.03), ('max_depth', 16), ('num_leaves', 300), ('feature_fraction', 0.5), ('min_data_in_leaf', 10), ('lambda_l1', 0.01), ('lambda_l2', 0.1)) 5698.201113441514 0.040191575332366815 0.12724324455076158 0.009559747812040173 [5729.868167835901, 5694.869288082981, 5669.865884405659] [0.03907352158511781, 0.04078955042036631, 0.04071165399161632] [0.12649754562155535, 0.1273666820597766, 0.12786550597095275] [0.009509185550019386, 0.00955187034523361, 0.009618187540867525]\n",
      "(('task', 'train'), ('boosting_type', 'gbdt'), ('objective', 'mse'), ('metric', 'mape'), ('learning_rate', 0.05), ('max_depth', 16), ('num_leaves', 300), ('feature_fraction', 0.5), ('min_data_in_leaf', 10), ('lambda_l1', 0.01), ('lambda_l2', 0.1)) 5678.8669385951325 0.04047597297328946 0.12798360239204187 0.009614113372202185 [5704.867412627905, 5692.868581317095, 5638.864821840399] [0.039403938673970865, 0.04097204433871111, 0.04105193590718641] [0.127122343383192, 0.1280897664568802, 0.1287386973360534] [0.00955475841032264, 0.009608143407998957, 0.009679438298284953]\n",
      "(('task', 'train'), ('boosting_type', 'gbdt'), ('objective', 'mse'), ('metric', 'mape'), ('learning_rate', 0.07), ('max_depth', 16), ('num_leaves', 300), ('feature_fraction', 0.5), ('min_data_in_leaf', 10), ('lambda_l1', 0.01), ('lambda_l2', 0.1)) 5619.199039959192 0.041011651972860906 0.1292976781507114 0.009712350845695716 [5625.86616177201, 5616.867352749384, 5614.863605356181] [0.03979322578730114, 0.04179586199592007, 0.04144586813536149] [0.12842956873420117, 0.12940445642916149, 0.1300590092887716] [0.009653736365579093, 0.00970470864168559, 0.009778607529822464]\n"
     ]
    }
   ],
   "source": [
    "for row in gsearch_results:\n",
    "    print(*[x for i, x in enumerate(row) if i != 5])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "metrics = ['score', 'mse', 'mae', 'mape']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "scrolled": true
   },
   "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>learning_rate</th>\n",
       "      <th>max_depth</th>\n",
       "      <th>num_leaves</th>\n",
       "      <th>feature_fraction</th>\n",
       "      <th>min_data_in_leaf</th>\n",
       "      <th>lambda_l1</th>\n",
       "      <th>lambda_l2</th>\n",
       "      <th>score</th>\n",
       "      <th>mse</th>\n",
       "      <th>mae</th>\n",
       "      <th>mape</th>\n",
       "      <th>best_iteration</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.003</td>\n",
       "      <td>16</td>\n",
       "      <td>300</td>\n",
       "      <td>0.5</td>\n",
       "      <td>10</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.1</td>\n",
       "      <td>5759.202754</td>\n",
       "      <td>0.039356</td>\n",
       "      <td>0.125680</td>\n",
       "      <td>0.009441</td>\n",
       "      <td>81504.333333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.005</td>\n",
       "      <td>16</td>\n",
       "      <td>300</td>\n",
       "      <td>0.5</td>\n",
       "      <td>10</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.1</td>\n",
       "      <td>5751.202630</td>\n",
       "      <td>0.039460</td>\n",
       "      <td>0.125800</td>\n",
       "      <td>0.009451</td>\n",
       "      <td>49306.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.010</td>\n",
       "      <td>16</td>\n",
       "      <td>300</td>\n",
       "      <td>0.5</td>\n",
       "      <td>10</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.1</td>\n",
       "      <td>5744.869037</td>\n",
       "      <td>0.039555</td>\n",
       "      <td>0.126068</td>\n",
       "      <td>0.009470</td>\n",
       "      <td>24673.666667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.007</td>\n",
       "      <td>16</td>\n",
       "      <td>300</td>\n",
       "      <td>0.5</td>\n",
       "      <td>10</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.1</td>\n",
       "      <td>5740.202525</td>\n",
       "      <td>0.039503</td>\n",
       "      <td>0.125906</td>\n",
       "      <td>0.009458</td>\n",
       "      <td>35007.666667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.030</td>\n",
       "      <td>16</td>\n",
       "      <td>300</td>\n",
       "      <td>0.5</td>\n",
       "      <td>10</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.1</td>\n",
       "      <td>5698.201113</td>\n",
       "      <td>0.040192</td>\n",
       "      <td>0.127243</td>\n",
       "      <td>0.009560</td>\n",
       "      <td>8774.666667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.050</td>\n",
       "      <td>16</td>\n",
       "      <td>300</td>\n",
       "      <td>0.5</td>\n",
       "      <td>10</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.1</td>\n",
       "      <td>5678.866939</td>\n",
       "      <td>0.040476</td>\n",
       "      <td>0.127984</td>\n",
       "      <td>0.009614</td>\n",
       "      <td>5525.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.070</td>\n",
       "      <td>16</td>\n",
       "      <td>300</td>\n",
       "      <td>0.5</td>\n",
       "      <td>10</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.1</td>\n",
       "      <td>5619.199040</td>\n",
       "      <td>0.041012</td>\n",
       "      <td>0.129298</td>\n",
       "      <td>0.009712</td>\n",
       "      <td>4167.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   learning_rate  max_depth  num_leaves  feature_fraction  min_data_in_leaf  \\\n",
       "0          0.003         16         300               0.5                10   \n",
       "1          0.005         16         300               0.5                10   \n",
       "2          0.010         16         300               0.5                10   \n",
       "3          0.007         16         300               0.5                10   \n",
       "4          0.030         16         300               0.5                10   \n",
       "5          0.050         16         300               0.5                10   \n",
       "6          0.070         16         300               0.5                10   \n",
       "\n",
       "   lambda_l1  lambda_l2        score       mse       mae      mape  \\\n",
       "0       0.01        0.1  5759.202754  0.039356  0.125680  0.009441   \n",
       "1       0.01        0.1  5751.202630  0.039460  0.125800  0.009451   \n",
       "2       0.01        0.1  5744.869037  0.039555  0.126068  0.009470   \n",
       "3       0.01        0.1  5740.202525  0.039503  0.125906  0.009458   \n",
       "4       0.01        0.1  5698.201113  0.040192  0.127243  0.009560   \n",
       "5       0.01        0.1  5678.866939  0.040476  0.127984  0.009614   \n",
       "6       0.01        0.1  5619.199040  0.041012  0.129298  0.009712   \n",
       "\n",
       "   best_iteration  \n",
       "0    81504.333333  \n",
       "1    49306.000000  \n",
       "2    24673.666667  \n",
       "3    35007.666667  \n",
       "4     8774.666667  \n",
       "5     5525.000000  \n",
       "6     4167.000000  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_gsearch_results = []\n",
    "for row in gsearch_results:\n",
    "    result = dict(row[0])\n",
    "    result['score'] = row[1]\n",
    "    result['mse'] = row[2]\n",
    "    result['mae'] = row[3]\n",
    "    result['mape'] = row[4]\n",
    "    result['best_iteration'] = np.mean(row[5])\n",
    "    df_gsearch_results.append(result)\n",
    "df_gsearch_results = pd.DataFrame(df_gsearch_results)\n",
    "df_gsearch_results = df_gsearch_results[list(params_gsearch.keys())+metrics+['best_iteration']]\n",
    "display(df_gsearch_results)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "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": [
    "for name in metrics:\n",
    "    plt.plot(df_gsearch_results[name])\n",
    "    plt.xlabel('rank'); plt.ylabel(name); plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY8AAAELCAYAAAAhuwopAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvIxREBQAAGzBJREFUeJzt3X2YXmVh5/HvD4bXCEXIkEVCGugSXaKC8ABxC62xENFdBaFC1BWwV80uBur2klrclzbV2hUuQytr6kpdEOvyFpcolZcQrYrVJTCjCWEwSMAoM3DlhSqI1iDw2z/OPfAQJ5nnZHLmmQy/z3U913POfe5zzn0bfH5zzn1eZJuIiIg6dut2AyIiYteT8IiIiNoSHhERUVvCIyIiakt4REREbQmPiIioLeERERG1JTwiIqK2hEdERNTW0+0GNGXq1KmeOXNmt5sREbHL6O/v32y7t5O6kzY8Zs6cSV9fX7ebERGxy5D0o07r5rRVRETUlvCIiIjaEh4REVFbwiMiImprdMBc0nrgZ8CzwDO2W5JuAF5ZqhwA/NT2MaX+a4HPAPsDzwHH2/6lpOOAzwH7ALcCH3BeRBIR0TXjcbXVXNubh2dsnzM8LWkx8ESZ7gG+ALzH9mpJBwG/KlU/DbwPWEkVHqcBt41D2yMiYgRdO20lScDZwHWlaB5wr+3VALYft/2spEOA/W3fVY42Pg+c0ZVGR0QE0Hx4GLhDUr+kBVstOxnYYPvBMj8LsKTlkr4r6UOl/FBgsG29wVIWERFd0nR4nGT7WODNwEJJv9O27J28cNQB1Sm0k4B3l++3S/q9OjuTtEBSn6S+TZs2jbHpETvfyqWL2bDoCFYuXdztpkSMSaPhYXuofG8ElgEnwPPjG2cCN7RVHwTutL3Z9i+oxjaOBYaA6W31ppeykfZ3pe2W7VZvb0d32EeMq5kDS5jG48wcWNLtpkSMSWPhIWmKpP2Gp6nGNO4ri08B1tpuPx21HHiNpH1LuPwucL/tx4AnJc0p4yTnAl9uqt0RTVo/eyEbOIj1sxd2uykRY9Lk1VbTgGXV7z09wLW2by/L5vPiU1bY/omky4F7qMZKbrV9S1n8fl64VPc2cqVV7KJOfMcH4R0fZFq3GxIxRpqst0u0Wi3nwYgREZ2T1G+71Und3GEeERG1JTwiIqK2hEdERNSW8IiIiNoSHhERUVvCIyIiakt4RIyjPJ4kJouER8Q4yuNJYrJIeESMozyeJCaL3GEeERFA7jCPiIiGJTwiIqK2hEdERNSW8IiIiNoSHhERUVvCI2Ic5SbBmCwSHhHjKDcJxmSR8IgYR7lJMCaL3CQYERHABLpJUNJ6SWskrZLUV8puKPOryvJVW60zQ9JTki5uKztN0gOS1km6pMk2R0TE6HrGYR9zbW8enrF9zvC0pMXAE1vVvxy4ra3O7sAS4FRgELhH0s2272+01RERsU3jER4jkiTgbOCNbWVnAD8Eft5W9QRgne2HS53rgdOBhEdERJc0PWBu4A5J/ZIWbLXsZGCD7QcBJL0M+FPgL7aqdyjwSNv8YCmL2OXkUt2YLJoOj5NsHwu8GVgo6Xfalr0TuK5tfhHw17af2tGdSVogqU9S36ZNm3Z0MxGNyaW6MVk0Gh62h8r3RmAZ1SkoJPUAZwI3tFU/EbhM0nrgPwP/RdKFwBBwWFu96aVspP1dabtlu9Xb27uTexMxdrlUNyaLxsY8JE0BdrP9szI9D/hIWXwKsNb24HB92ye3rbsIeMr2p0rQHCnpcKrQmA+8q6l2RzTpxHd8EN7xQaZ1uyERY9TkgPk0YFk1Lk4PcK3t28uy+bz4lNU22X6mHIEsB3YHrrI90EB7IyKiQ7lJMCIigAl0k2BERExOCY+IiKgt4REREbUlPCIioraER0RE1JbwiIiI2hIeERFRW8IjIiJqS3hERERtCY+IiKgt4REREbUlPCIioraER0RE1JbwiIiI2hIeERFRW8IjIiJqS3hERERtCY+IiKgt4REREbU1Gh6S1ktaI2mVpL5SdkOZX1WWryrlp0rqL/X7Jb2xbTvHlfJ1kq6QpCbbHRER29czDvuYa3vz8Iztc4anJS0Gniizm4G32n5U0quB5cChZdmngfcBK4FbgdOA28ah7RERMYKunbYqRw9nA9cB2P6e7UfL4gFgH0l7SToE2N/2XbYNfB44oyuNjogIoPnwMHBHOQ21YKtlJwMbbD84wnpnAd+1vYXq6GOwbdkgLxyRREREFzR92uok20OSDgZWSFpr+86y7J2Uo452kmYDlwLz6u6sBNQCgBkzZux4qyMiYrsaPfKwPVS+NwLLgBMAJPUAZwI3tNeXNL3UO9f2Q6V4CJjeVm16KRtpf1fabtlu9fb27syuREREm8bCQ9IUSfsNT1MdSdxXFp8CrLU92Fb/AOAW4BLb3x4ut/0Y8KSkOWWc5Fzgy021OyIiRtfkkcc04J8krQbuBm6xfXtZNp9fP2V1IfCvgT9ru5T34LLs/cBngXXAQ+RKq4iIrlJ1AdPk02q13NfX1+1mRETsMiT12251Ujd3mEdERG0Jj4iIqC3hERERtSU8IiKitoRHRETUlvCIiIjaEh4REVFbwiMiImpLeERERG0Jj4iIqC3hERERtSU8IiKitoRHRETUlvCIiIjaEh4REVFbwiMiImpLeERERG0Jj4iIqC3hERERtTUaHpLWS1ojaZWkvlJ2Q5lfVZavaqv/YUnrJD0g6U1t5aeVsnWSLmmyzRERMbqecdjHXNubh2dsnzM8LWkx8ESZPgqYD8wGXgF8VdKsUnUJcCowCNwj6Wbb949D2yMiYgTjER4jkiTgbOCNpeh04HrbW4AfSloHnFCWrbP9cFnv+lI34RER0SVNj3kYuENSv6QFWy07Gdhg+8EyfyjwSNvywVK2rfKIiOiSpo88TrI9JOlgYIWktbbvLMveCVy3M3dWAmoBwIwZM3bmpiMiok2jRx62h8r3RmAZ5TSUpB7gTOCGtupDwGFt89NL2bbKR9rflbZbtlu9vb07qxsREbGVxsJD0hRJ+w1PA/OA+8riU4C1tgfbVrkZmC9pL0mHA0cCdwP3AEdKOlzSnlSD6jc31e6IJq1cupgNi45g5dLF3W5KxJg0eeQxDfgnSaupQuAW27eXZfPZ6pSV7QHgRqqB8NuBhbaftf0McCGwHPg+cGOpG7HLmTmwhGk8zsyBJd1uSsSYdDzmIekk4EjbV0vqBV5m+4fbql+ujjp6G8vO30b5x4CPjVB+K3Brp22NmKjWz14IA0tYP3sh07rdmIgxkO3RK0l/DrSAV9qeJekVwFLbv910A3dUq9VyX19ft5sREbHLkNRvu9VJ3U5PW70deBvwcwDbjwL77VjzIiJiV9dpeDzt6hDF8PwAeEREvER1Gh43SvoMcICk9wFfBf6uuWZFRMRE1tGAue1PSDoVeBJ4JfBntlc02rKIiJiwRg0PSbsDX7U9F0hgRETE6KetbD8LPCfpN8ahPRERsQvo9D6Pp4A1klZQrrgCsP1HjbQqIiImtE7D46byiYiI6HjA/JryXKnhlzM9YPtXzTUrIiImso7CQ9IbgGuA9YCAwySd1/Z49YiIeAnp9LTVYmCe7QcAyuthrwOOa6phERExcXV6k+Aew8EBYPsHwB7NNCkiIia6To88+iR9FvhCmX83kKcORkS8RHUaHhcAC4HhS3O/BfxtIy2KiIgJr9Pw6AE+aftyeP6u870aa1VERExonY55fA3Yp21+H6qHI0ZExEtQp+Gxt+2nhmfK9L7NNCkiIia6TsPj55KOHZ6R1AL+pZkmRUTERNdpeHwAWCrpW5K+BVwPXDjaSpLWS1ojaZWkvrbyiyStlTQg6bJStoeka0r970v6cFv90yQ9IGmdpEvqdTEiIna2TgfMDwdeB8wAzgROpLxVsANzbW8enpE0FzgdONr2FkkHl0XvAPay/RpJ+wL3S7oOeARYApwKDAL3SLrZ9v0d7j8iInayTo88/rvtJ4EDgLlUl+l+egf3eQHwcdtbAGxvLOUGpkjqoRqQf5rq5VMnAOtsP2z7aaqjntN3cN8REbETdBoez5bvfwf8ne1bgD07WM/AHZL6JS0oZbOAkyWtlPRNSceX8i9SPe79MeDHwCds/zNwKNXRx7DBUhYREV3S6WmrofIO81OBSyXtRWfBc5LtoXJqaoWktWWfBwJzgOOp3o9+BNURxrPAK4CXA9+SVOty4BJQCwBmzJhRZ9WIiKih0yOPs4HlwJts/5Tqx/9PRlvJ9lD53ggsowqIQeAmV+4GngOmAu8Cbrf9q1L/20ALGAIOa9vs9FI20v6utN2y3ert7e2waxHjZ+XSxWxYdAQrly7udlMixqSj8LD9C9s32X6wzD9m+47trSNpiqT9hqeBecB9wJeoxk2Gn867J7CZ6lTVG9vqzwHWAvcAR0o6vLxTZD5wc92ORkwEMweWMI3HmTmwpNtNiRiTTk9b7YhpwDJJw/u51vbtJQCuknQf1aD4ebYtaQlwtaQBqneGXG37XgBJF1Id+ewOXGV7oMF2RzRm/eyFMLCE9bMXMq3bjYkYA9mdXnG7a2m1Wu7ry4N/IyI6JanfdquTup2OeURERDwv4REREbUlPCIioraER0RE1JbwiIiI2hIeERFRW8IjIiJqS3hERERtCY+IiKgt4REREbUlPCIioraER0RE1JbwiIiI2hIeERFRW8IjIiJqS3hERERtCY+IiKgt4REREbUlPCIiorZGw0PSeklrJK2S1NdWfpGktZIGJF3WVv5aSf+vlK+RtHcpP67Mr5N0hSQ12e6IiNi+nnHYx1zbm4dnJM0FTgeOtr1F0sGlvAf4AvAe26slHQT8qqz2aeB9wErgVuA04LZxaHtERIygG6etLgA+bnsLgO2NpXwecK/t1aX8cdvPSjoE2N/2XbYNfB44owvtjoiIounwMHCHpH5JC0rZLOBkSSslfVPS8W3llrRc0nclfaiUHwoMtm1zsJRFRESXNH3a6iTbQ+XU1ApJa8s+DwTmAMcDN0o6opSfVMp+AXxNUj/wRKc7KwG1AGDGjBk7tSMREfGCRo88bA+V743AMuAEqiOHm1y5G3gOmFrK77S92fYvqMY2jgWGgOltm51eykba35W2W7Zbvb29TXUrIuIlr7HwkDRF0n7D01RjGvcBXwLmlvJZwJ7AZmA58BpJ+5bB898F7rf9GPCkpDnlKqtzgS831e6IiBhdk6etpgHLylW1PcC1tm+XtCdwlaT7gKeB88pA+E8kXQ7cQzVWcqvtW8q23g98DtiH6iqrXGkVEdFFqn63J59Wq+W+vr7RK0ZEBACS+m23OqmbO8wjIqK2hEdERNSW8IiIiNoSHhERUVvCIyIiakt4REREbQmPiIioLeERERG1JTwiIqK2hEdERNSW8IiIiNoSHhERUVvCIyIiakt4REREbQmPiIioLeERERG1JTwiIqK2hEdERNSW8IiIiNoaDQ9J6yWtkbRKUl9b+UWS1koakHTZVuvMkPSUpIvbyk6T9ICkdZIuabLNERExup5x2Mdc25uHZyTNBU4Hjra9RdLBW9W/HLitrf7uwBLgVGAQuEfSzbbvb77pERExkm6ctroA+LjtLQC2Nw4vkHQG8ENgoK3+CcA62w/bfhq4nip8IiKiS5oODwN3SOqXtKCUzQJOlrRS0jclHQ8g6WXAnwJ/sdU2DgUeaZsfLGUREdElTZ+2Osn2UDk1tULS2rLPA4E5wPHAjZKOABYBf237KUk7tLMSUAsAZsyYsROaHxERI2k0PGwPle+NkpZRnYIaBG6ybeBuSc8BU4ETgd8vA+gHAM9J+iXQDxzWttnpwNA29nclcCVAq9VyM72KiIjGwkPSFGA32z8r0/OAjwBPAXOBr0uaBewJbLZ9ctu6i4CnbH9KUg9wpKTDqUJjPvCuptodERGja/LIYxqwrJyC6gGutX27pD2BqyTdBzwNnFeOQkZk+xlJFwLLgd2Bq2wPbKt+REQ0T9v53d6ltVot9/X1jV4xIiIAkNRvu9VJ3dxhHhERtSU8IiKitoRHRETUlvCIiIjaEh4REVFbwiMiImpLeERERG0Jj4iIqC3hERERtSU8IiKitoRHRETUlvCIiIjaEh4REVFbwiMiImpLeERERG0Jj4iIqC3hERERtSU8IiKitoRHRETU1mh4SFovaY2kVZL62sovkrRW0oCky0rZqZL6S/1+SW9sq39cKV8n6QpJarLdERGxfT3jsI+5tjcPz0iaC5wOHG17i6SDy6LNwFttPyrp1cBy4NCy7NPA+4CVwK3AacBt49D2iIgYQTdOW10AfNz2FgDbG8v392w/WuoMAPtI2kvSIcD+tu+ybeDzwBldaHdERBRNh4eBO8ppqAWlbBZwsqSVkr4p6fgR1jsL+G4JmEOBwbZlg7xwRBIREV3QdHicZPtY4M3AQkm/Q3Wq7EBgDvAnwI3tYxiSZgOXAv+x7s4kLZDUJ6lv06ZNO6UDETvTyqWL2bDoCFYuXdztpkSMSaPhYXuofG8ElgEnUB053OTK3cBzwFQASdNLvXNtP1Q2MwRMb9vs9FI20v6utN2y3ert7W2iSxFjMnNgCdN4nJkDS7rdlIgxaSw8JE2RtN/wNDAPuA/4EjC3lM8C9gQ2SzoAuAW4xPa3h7dj+zHgSUlzyhHKucCXm2p3RJPWz17IBg5i/eyF3W5KxJioGoNuYMPSEVRHEVCdqrrW9sck7QlcBRwDPA1cbPsfJf034MPAg22bmWd7o6QW8DlgH6qrrC7yKA1vtVru6+vbXpWIiGgjqd92q6O6TYVHtyU8IiLqqRMeucM8IiJqS3hERERtCY+IiKgt4REREbUlPCIioraER0RE1JbwiIiI2ibtfR6SNgE/6nY7xmgq1aPqXwrS18kpfd21/Kbtjp7tNGnDYzKQ1NfpDTu7uvR1ckpfJ6+ctoqIiNoSHhERUVvCY2K7stsNGEfp6+SUvk5SGfOIiIjacuQRERG1JTzGkaTTJD0gaZ2kS7ZT7yxJLu8xQdKekq6WtEbSaklvaKv7zlJ+r6TbJU0dh66MqqG+nlP6OSDp0nHoxqhG66ek8yVtkrSqfP6wbdl5kh4sn/Payo8r/V8n6Yr21zR3U0N9/ZikRyQ9NV796MTO7qukfSXdImlt+e/34+PZn0bYzmccPsDuwEPAEVRvT1wNHDVCvf2AO4G7gFYpWwhcXaYPBvqpgr8H2AhMLcsuAxZN0r4eBPwY6C3LrgF+b6L3Ezgf+NQI6x4IPFy+X16mX16W3Q3MAUT18rM37wr/pjvY1znAIcBT3e5jk30F9gXmljp7At+aCP+uY/nkyGP8nACss/2w7aeB64HTR6j3UeBS4JdtZUcB/wjPvw/+p0CL6sdFwJTy1+n+wKON9aBzTfT1COBB25tKva8CZzXT/I512s+RvAlYYfufbf8EWAGcJukQYH/bd7n6pfk8cEYTja9pp/cVoPTzsUZavON2el9t/8L21wHKNr8LTG+g7eMm4TF+DgUeaZsfLGXPk3QscJjtW7ZadzXwNkk9kg4Hjiv1fgVcAKyhCo2jgP/dUPvr2Ol9BdYBr5Q0U1IP1Q/qYU11oEOj9rM4q5xu+6Kk4TZva91Dy/Ro2xxvTfR1omq0r5IOAN4KfG3nNXn8JTwmCEm7AZcDHxxh8VVU/xH2AX8DfAd4VtIeVOHxOuAVwL1U74Gf0Hakr+WvuAuAG6gO+dcDz45He8foH4CZtl9L9VfoNV1uT5PS11GUP3yuA66w/XCD7WtcwmP8DPHiv5Snl7Jh+wGvBr4haT3VueCbJbVsP2P7j20fY/t04ADgB8AxALYfKqc4bgT+bfNdGVUTfcX2P9g+0fbrgQeGy7totH5i+3HbW8rsZ6mOpLa37hAvPp3xa9vskib6OlE12dcrqU6//s1ObXE3dHvQ5aXyoRrcfhg4nBcG4WZvp/43eGEQeV9gSpk+FbizTL8CeIwXBpE/CiyejH0t8weX75cDq4BZE72fwCFt028H7irTBwI/LH15eZk+sCzbesD8LbvCv+mO9LWt/kQaMG/q3/Uvgf8L7NbtPu6U/5263YCX0gd4C9Vfyw8B/7WUfQR42wh1239QZ1L9pf19qoHi32yr959K+b1Uh9IHdbufDfb1OuD+8pnf7T520k/gfwAD5Qfo68Cr2tb9A6qxnHXAe9vKW8B9ZZufotzM2+1PQ329jOo05XPle1G3+9lEX6mOQFz+u15VPn/Y7X6O5ZM7zCMioraMeURERG0Jj4iIqC3hERERtSU8IiKitoRHRETUlvCIiIjaEh4xKUn6I0nfl/R/aq43U9K7mmpX235eVR7l/T1JvzXGbR0j6S1t82/b3mPwI3aG3OcRk5KktcAptgdHrfzi9d4AXGz739dcb3fbHT9rq/y499j+y63KRfX/y+dqbOt8qpssL+x0nYixSnjEpCPpf1Hd5fsA1eO0f4vqWVp7UN3B/GVJM4G/B6aU1S60/R1JdwH/huqxEtcAP6Hth1nSV4BP2P5GeYHRZ4BTqN5D8i9UD3x8GbAZON8jPG68HCVcRfVgxx8A7wWWAyupnpH0FuAS4HhgH+CLtv+8rHs88MnS7i1Uj3BZU+oNUd35vM9wm0s/rwKmApuo7nj+saTPAU9S3c3+r4AP2f5i7f+x46Wr27e455NPEx+qp+5OBf4K+A+lbPghi1OonqG1dyk/Eugr028AvtK2nfNpe+kP8BXgDWXawNlleg+qJwAPP2fsHOCq7bRvEdURDlSPZHkOmNO2fPh5SLtTPb7ltVTPWXoYOL4s25/qOUxbt/H5eapH1pxXpv8A+FKZ/hywlOrU9VFU76/o+r9bPrvOp2dHAidiFzKP6v0gF5f5vYEZVO8/+ZSkY6iOAGbtwLafpXrQHcArqY5uVpS3xu5O9dDKTv3I9l1t82dLWkAVDodQ/cAbeMz2PQC2nwQY5S21rwfOLNN/T/UsqWFfcnV67H5J02q0NSLhEZOegLNsP/CiQmkRsAE4muqv71/++qoAPMOLLyzZu236l35hnEPAgKvHxe+In7e17XDgYqojjJ+UU0x7b2vFMdjSNj0h3pMeu45cbRWT3XLgojIQjaTXlfLfoPor/jngPVRHCgA/o3rfyLD1wDGSditvizthG/t5AOiV9Pqynz0kzd7BNu9PFSZPlCOCN7ft45Ay7oGk/crLhbZuc7vvAPPL9LupXqQVMWYJj5jsPko1HnGvpIEyD/C3wHmSVgOv4oW//O+lekvjakl/DHybavD8fuAKqndP/xpX76X+feDSss1V7OCLuWyvBr4HrAWuLW0Y3sc5wP8s+1hBdUTydeCocunvOVtt7iLgvZLupQrJD+xImyK2lqutIiKithx5REREbRkwj2iQpCXAb29V/EnbV3ejPRE7S05bRUREbTltFRERtSU8IiKitoRHRETUlvCIiIjaEh4REVHb/wdZju2lf9bTFwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "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": [
    "for p in params_gsearch.keys():\n",
    "    plt.scatter(df_gsearch_results[p], df_gsearch_results['score'], s=3)\n",
    "    plt.scatter(df_gsearch_results[p][:10], df_gsearch_results['score'][:10], s=3)\n",
    "    plt.xlabel(p); plt.ylabel('score'); plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZIAAAELCAYAAADz6wBxAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvIxREBQAAFkRJREFUeJzt3XGwnXV95/H3p8Qo1A1QuLJIkEChYrKlUS9Rq83KWm1ot6SdgoBdMe4iq4Xptmu3w5Ydy7DtVrtlh2VMK2i1i4IIbKk4qExFR90tIJcMBC8IBBpIwJXo1jourTTmu3+c59rjnZB7bn55cnKT92vmDOf8fs/vyfcL4X7u8zznPCdVhSRJu+tHxl2AJGlhM0gkSU0MEklSE4NEktTEIJEkNTFIJElNDBJJUhODRJLUxCCRJDVZNO4C9oYjjzyyli1bNu4yJGlBueeee75ZVRNzbXdABMmyZcuYmpoadxmStKAkeXyU7Ty1JUlqYpBIkpoYJJKkJgaJJKmJQSJJamKQSJKaGCSSpCYGiSSpiUEiSWpikEiSmhgkkqQmBokkqYlBIklqYpBIkpoYJJKkJgaJJKmJQSJJamKQSJKaGCSSpCYGiSSpiUEiSWpikEiSmhgkkqQmBokkqYlBIklqYpBIkpr0GiRJ1iR5KMmmJBfvZH51kg1Jtic5c2j8uG783iTTSd7ZjR+S5NYkX+vG39tn/ZKkufUWJEkOAtYDpwPLgXOTLJ+12RPAOuC6WeNfB15TVSuBVwEXJ3lxN/dHVXUy8HLgtUlO76kFSdIIFvW471XApqp6DCDJ9cBa4IGZDapqcze3Y3hhVT079PL5dIFXVc8AX5jZJskGYGl/LUiS5tLnqa1jgC1Dr7d2YyNJcmySjd0+3ldVT82aPwz4ReD2PVCrJGk37bMX26tqS1WdApwIvC3JUTNzSRYBHweunDnimS3JBUmmkkxt27Zt7xQtSQegPoPkSeDYoddLu7F56Y5Evgr8zNDw1cAjVXXFLtZdXVWTVTU5MTEx3z9WkjSiPoPkbuCkJMcnWQycA9wyysIkS5Mc3D0/HHgd8FD3+veAQ4Hf6KVqSdK89BYkVbUduAi4DXgQuKGqppNcluQMgCSnJtkKnAVclWS6W/4y4K4k9wFfZPBOrfuTLAUuYfAusJm3B5/fVw+SpLmlqsZdQ+8mJydrampq3GVI0oKS5J6qmpxru332YrskaWEwSCRJTQwSSVITg0SS1MQgkSQ1MUgkSU0MEklSE4NEktTEIJEkNTFIJElNDBJJUhODRJLUxCCRJDUxSCRJTQwSSVITg0SS1MQgkSQ1MUgkSU0MEklSE4NEktTEIJEkNTFIJElNDBJJUhODRJLUxCCRJDUxSCRJTXoNkiRrkjyUZFOSi3cyvzrJhiTbk5w5NH5cN35vkukk7xyae2WS+7t9XpkkffYgSdq13oIkyUHAeuB0YDlwbpLlszZ7AlgHXDdr/OvAa6pqJfAq4OIkL+7m/gR4B3BS91jTSwOSpJH0eUSyCthUVY9V1bPA9cDa4Q2qanNVbQR2zBp/tqq+1718/kydSY4GllTVnVVVwDXAL/XYgyRpDn0GyTHAlqHXW7uxkSQ5NsnGbh/vq6qnuvVbR9lnkguSTCWZ2rZt27yLlySNZp+92F5VW6rqFOBE4G1Jjprn+qurarKqJicmJvopUpLUa5A8CRw79HppNzYv3ZHIV4Gf6dYvbd2nJGnP6TNI7gZOSnJ8ksXAOcAtoyxMsjTJwd3zw4HXAQ9V1deB7yR5dfdurfOAT/ZTviRpFL0FSVVtBy4CbgMeBG6oqukklyU5AyDJqUm2AmcBVyWZ7pa/DLgryX3AF4E/qqr7u7lfAz4EbAIeBT7TVw+SpLll8Oan/dvk5GRNTU2NuwxJWlCS3FNVk3Ntt89ebJckLQwGiSSpiUEiSWpikEiSmhgkkqQmBok0JnfdeDnfuPQE7rrx8nGXIjUxSKQxWTa9nqP4Fsum14+7FKmJQSKNyeYVF/INjmDzigvHXYrUxA8kSpJ2yg8kSpL2CoNEktTEIJEkNTFIJElNDBJJUhODRJLUxCCRJDUxSCRJTQwSSVITg0SS1MQgkSQ1MUgkSU0MEklSE4NEktTEIJEkNTFIJElNeg2SJGuSPJRkU5KLdzK/OsmGJNuTnDk0vjLJHUmmk2xMcvbQ3Bu6Nfcm+V9JTuyzB0nSrvUWJEkOAtYDpwPLgXOTLJ+12RPAOuC6WePPAOdV1QpgDXBFksO6uT8BfrWqVnbr/lM/HUiSRrGox32vAjZV1WMASa4H1gIPzGxQVZu7uR3DC6vq4aHnTyV5GpgAvg0UsKSbPhR4qr8WJElzGfmIJMlxSX62e35wkn8yx5JjgC1Dr7d2Y/OSZBWwGHi0Gzof+HSSrcBbgfc+x7oLkkwlmdq2bdt8/1ipd3fdeDnfuPQE7rrx8nGXIjUZKUiSvAO4CbiqG1oK/EVfRQ39uUcDHwXeXlUzRy2/Cfx8VS0FPgL8t52traqrq2qyqiYnJib6LlWat2XT6zmKb7Fsev24S5GajHpEciHwWuA7AFX1CPCiOdY8CRw79HppNzaSJEuAW4FLqurObmwC+Kmquqvb7BPAT4+6T2lfsnnFhXyDI9i84sJxlyI1GfUayfeq6tkkACRZxOBaxa7cDZyU5HgGAXIO8JZR/rAki4GbgWuq6qahqb8BDk3yE911lDcCD47Yg7RPedVZ74az3s1R4y5EajTqEckXk/wOcHCSNwI3Ap/a1YKq2g5cBNzG4If9DVU1neSyJGcAJDm1u9ZxFnBVkulu+ZuB1cC67m2+9yZZ2e3zHcD/THIfg2sk/2FeHUuS9qhUzXVgAUl+BPg3wJuAMAiHD9Uoi/cBk5OTNTU1Ne4yJGlBSXJPVU3Otd1Ip7a6C90f7B6SJP3ASEGS5CTgDxh8sPAFM+NVdUJPdUmSFohRr5F8hMEnyrcDpwHXAB/rqyhJ0sIxapAcXFW3M7im8nhVXQr8Qn9lSZIWipHf/ttdcH8kyUUM3s77wv7KkiQtFKMekfw74BDg14FXAv8KOK+voiRJC8eoQVIMblVyCzAJ/AS+g0tq4r22tL8Y9dTWtQw++Hc/sGOObSWNYOZeW0yvh7PePe5ypN02apBsq6pbeq1EOsBsXnEhTK9n84oLvU2KFrRRP9n+BuBc4HbgezPjVfXn/ZW25/jJdkmavz36yXbg7cDJwPP4x1NbBSyIIJEk9WfUIDm1ql7aayWSpAVp1Hdt/dVOvm9dkqSRj0heDdyb5K8ZXCMJUFV1Sm+VSZIWhFGDZE2vVUiSFqxRbyP/eN+FSJIWplGvkUiStFMGiSSpiUEiSWpikEiSmhgkkqQmBokkqYlBIklqYpBIkpoYJJKkJr0GSZI1SR5KsinJxTuZX51kQ5LtSc4cGl+Z5I4k00k2Jjl7aC5Jfj/Jw0keTPLrffYgSdq13oIkyUHAeuB0YDlw7k7uIPwEsA64btb4M8B5VbWCwX2+rkhyWDe3DjgWOLmqXgZc30sDUs/8znbtL/o8IlkFbKqqx6rqWQY/8NcOb1BVm6tqI7O+B76qHq6qR7rnTwFPAxPd9LuAy6pqRzf/dI89SL2Z+c72ZdPrx12K1KTPIDkG2DL0ems3Ni9JVgGLgUe7oR8Hzk4yleQzSU5qrlQag80rLuQbHDH47nZpAdunL7YnORr4KPD2mSMQ4PnA33ffI/xB4MPPsfaCLmymtm3btncKlqQDUJ9B8iSDaxkzlnZjI0myBLgVuKSq7hya2so/flf8zcBOv1yrqq6uqsmqmpyYmNjZJtJYeWpL+4s+g+Ru4KQkxydZDJwD3DLKwm77m4FrquqmWdN/AZzWPf/nwMN7qF5pr3p8ySv4foXHl7xi3KVITXoLkqraDlwE3AY8CNxQVdNJLktyBkCSU5NsBc4Crkoy3S1/M7AaWJfk3u6xspt7L/ArSe4H/gA4v68epD4d950NHJTiuO9sGHcpUpNRv2p3t1TVp4FPzxp7z9Dzuxmc8pq97mPAx55jn98GfmHPVirtfY8veQVH/u3nePzQV3DUuIuRGuzTF9ul/ZlHJNpfGCTSmPj2X+0vUlXjrqF3k5OTNTU1Ne4yJGlBSXJP91GLXfKIRJLUxCCRJDUxSCRJTQwSSVITg0SS1MQgkSQ1MUgkSU0MEklSE4NEktTEIJEkNTFIJElNDBJJUhODRJLUxCCRJDUxSCRJTQwSSVITg0SS1MQgkSQ1MUgkSU0MEklSE4NEktTEIJEkNek1SJKsSfJQkk1JLt7J/OokG5JsT3Lm0PjKJHckmU6yMcnZO1l7ZZLv9lm/JGluvQVJkoOA9cDpwHLg3CTLZ232BLAOuG7W+DPAeVW1AlgDXJHksKF9TwKH91S6JGke+jwiWQVsqqrHqupZ4Hpg7fAGVbW5qjYCO2aNP1xVj3TPnwKeBibgBwH1X4Hf7rF2SdKI+gySY4AtQ6+3dmPzkmQVsBh4tBu6CLilqr7eXKEkqdmicRewK0mOBj4KvK2qdiR5MXAW8PoR1l4AXADwkpe8pM8yJemA1ucRyZPAsUOvl3ZjI0myBLgVuKSq7uyGXw6cCGxKshk4JMmmna2vqqurarKqJicmJnanfknSCPo8IrkbOCnJ8QwC5BzgLaMsTLIYuBm4pqpumhmvqluBfzq03Xer6sQ9WrUkaV56OyKpqu0MrmfcBjwI3FBV00kuS3IGQJJTk2xlcLrqqiTT3fI3A6uBdUnu7R4r+6pVkrT7UlXjrqF3k5OTNTU1Ne4yJGlBSXJPVU3OtZ2fbJckNTFIJElNDBJJUhODRJLUxCCRJDUxSCRJTQwSSVITg0SS1MQgkSQ1MUgkSU0MEklSE4NEktTEIJEkNTFIJElNDBJJUhODRJLUxCCRJDUxSCRJTQwSSVITg0SS1MQgkSQ1MUgkSU0MEklSE4NEktTEIJEkNek1SJKsSfJQkk1JLt7J/OokG5JsT3Lm0PjKJHckmU6yMcnZQ3PXdvv8apIPJ3lenz1IknattyBJchCwHjgdWA6cm2T5rM2eANYB180afwY4r6pWAGuAK5Ic1s1dC5wM/CRwMHB+Lw1IkkayqMd9rwI2VdVjAEmuB9YCD8xsUFWbu7kdwwur6uGh508leRqYAL5dVZ+emUvyFWBpjz1IkubQ56mtY4AtQ6+3dmPzkmQVsBh4dNb484C3Ap9tqFGS1Gifvtie5Gjgo8Dbq2rHrOk/Br5UVV9+jrUXJJlKMrVt27a+S5WkA1afQfIkcOzQ66Xd2EiSLAFuBS6pqjtnzf0ug1Nd//651lfV1VU1WVWTExMT8ypckjS6PoPkbuCkJMcnWQycA9wyysJu+5uBa6rqpllz5wM/B5y7k6MUSdJe1luQVNV24CLgNuBB4Iaqmk5yWZIzAJKcmmQrcBZwVZLpbvmbgdXAuiT3do+V3dwHgKOAO7rx9/TVgyRpbqmqcdfQu8nJyZqamhp3GZK0oCS5p6om59pun77YLkna9xkkkqQmBokkqYlBIklqYpBIkpoYJJKkJgaJJKmJQSJJamKQSJKaGCSSpCYHxC1SkmwDHh93HXvAkcA3x13EXnCg9An2ur/aX3o9rqrmvH36AREk+4skU6Pc92ahO1D6BHvdXx1IvYKntiRJjQwSSVITg2RhuXrcBewlB0qfYK/7qwOpV6+RSJLaeEQiSWpikIxJkjVJHkqyKcnFO5l/fpJPdPN3JVnWjS9O8pEk9ye5L8nrh9ac241vTPLZJEfutYZ2oadez+76nE7yvr3WzBxG6HV1kg1Jtic5c9bc25I80j3eNjT+yu7fwaYkVybJ3uhlV3rq8/eTbEny3b3Rw6j2dK9JDklya5KvdX9/37u3eulNVfnYyw/gIOBR4ARgMXAfsHzWNr8GfKB7fg7wie75hcBHuucvAu5h8AvBIuBp4Mhu7g+BS/fTXo8AngAmurn/AbxhgfS6DDgFuAY4c2j8x4DHun8e3j0/vJv7CvBqIMBngNP30z5fDRwNfHfc/y377BU4BDit22Yx8OVx/zdtfXhEMh6rgE1V9VhVPQtcD6ydtc1aBj8gAW4C3tD9Jroc+DxAVT0NfBuYZPBDJsCPdtstAZ7qu5ER9NHrCcAjVbWtW/M54Fd67WI0c/ZaVZuraiOwY9banwP+sqr+b1X9DfCXwJokRwNLqurOGvzkuQb4pd472bU93me35s6q+nr/5c/LHu+1qp6pqi90a58FNgBL+26kTwbJeBwDbBl6vbUb2+k2VbUd+FsGv4nfB5yRZFGS44FXAsdW1T8A7wLuZxAgy4E/7bOJEe3xXoFNwEuTLEuyiMEP1mN77WI0o/Q637XHdM93Z5996aPPfVWvvSY5DPhF4PaGGsfOIFl4PszgL+QUcAXwV8D3kzyPQZC8HHgxsBH4j+Mqcg/Zaa/db3fvAj7B4LTAZuD7Y6pR2i3dL0EfB66sqsfGXU8Lg2Q8nuSHf4Ne2o3tdJvuL9yhwLeqantV/WZVrayqtcBhwMPASoCqerQ7BXID8NP9tjGSPnqlqj5VVa+qqtcAD82Mj9kovc537ZP88GmP+eyzL330ua/qs9erGZyivaKpwn2AQTIedwMnJTk+yWIGF5hvmbXNLcDMO1rOBD5fVdW94+NHAZK8EdheVQ8w+Au6PMnMDdbeCDzYdyMj6KNXkryo++fhDC7Wf6j/VuY0Sq/P5TbgTUkO73p6E3Bbd83gO0le3V03Og/4ZB/Fz8Me77OnOveEXnpN8nsMfmH6jR5q3vvGfbX/QH0AP8/gt+hHgUu6scuAM7rnLwBuZHA94CvACd34Mga/gT/I4CLzcUP7fGc3vhH4FHDEuPvssdePAw90j3PG3eM8ej2Vwem6/wd8C5geWvuvu38Hm4C3D41PAl/t9vl+ug8S74d9/mG3Zkf3z0vH3WcfvTI4Mqnu7/W93eP8cffZ8vCT7ZKkJp7akiQ1MUgkSU0MEklSE4NEktTEIJEkNTFIJElNDBLpOeyp25knuTTJb42w3Z/Nvg35iPv/wbok13a3PP9qkg93t86RemWQSPuXa4GTgZ8EDgbOH285OhAYJNIckrwwye3dlxfdn2RtN76s+3KiP0vycHc08LNJ/nf3RUarhnbzU0nu6Mbf0a1Pkvd3RxCfY/CdKzN/5nuS3N0dWVw96pdZVdWnq8PgLgEL+vbkWhgMEmlufw/8clW9AjgNuHzoB/uJwOUMjgJOBt4CvA74LeB3hvZxCvAvgNcA70nyYuCXgZcyuOX/efzwTTbfX1WnVtU/Y3Bk8S/nU3B3SuutwGfns07aHYvGXYC0AAT4L0lWM7gP1DHAUd3cX1fV/QBJpoHbq6qS3M/gXmEzPllVfwf8XZIvMPjCpNXAx6vq+8BTST4/tP1pSX6bwbfp/RgwzeD+aaP6Y+BLVfXlefYqzZtBIs3tV4EJ4JVV9Q9JNjO40STA94a22zH0egc//P/X7JvaPedN7pK8gEEQTFbVliSXDv15c0ryu129/3bUNVILT21JczsUeLoLkdOA43ZjH2uTvCDJEcDrGdye/EvA2UkO6r5S97Ru25nQ+GaSFzK4tf5IkpzP4Ctez62q2V/9KvXCIxJpbtcCn+pOV00BX9uNfWwEvgAcCfznqnoqyc0Mrps8ADwB3AFQVd9O8kEGt47/PwxCZ1QfAB4H7ugu4/x5VV22G/VKI/M28pKkJp7akiQ18dSWtEAkWQ+8dtbwf6+qj4yjHmmGp7YkSU08tSVJamKQSJKaGCSSpCYGiSSpiUEiSWry/wErsuABGJgTdwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for p in params_gsearch.keys():\n",
    "    plt.scatter(df_gsearch_results[p], df_gsearch_results['mae'], s=3)\n",
    "    plt.scatter(df_gsearch_results[p][:10], df_gsearch_results['mae'][:10], s=3)\n",
    "    plt.xlabel(p); plt.ylabel('mae'); plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['learning_rate', 'max_depth', 'num_leaves', 'feature_fraction',\n",
       "       'min_data_in_leaf', 'lambda_l1', 'lambda_l2', 'score', 'mse', 'mae',\n",
       "       'mape', 'best_iteration'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_gsearch_results.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZsAAAELCAYAAAAP/iu7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvIxREBQAAGvZJREFUeJzt3XmUZnV95/H3x24aiQtI0+YgDenOQExAM6g1BB2TMSFou7YZydiOR9BgiKM4STzHBMZxCZnMCZrEjEo0qEQ0GiDEpccsHRXNYsaWakVZtLVYDN062rK5DFGafOeP+yt4urrWrrr1PFa/X+c8p+7zu7977/fyFP2puzy/m6pCkqQ+PWDYBUiSVj7DRpLUO8NGktQ7w0aS1DvDRpLUO8NGktQ7w0aS1Luhh02STUl2JplIct408w9Ncnmbvz3JhtZ+epIdSa5tP39uYJlPtHVe014PX749kiRNtXqYG0+yCrgIOB3YBVydZGtV3TDQ7Wzgjqo6PskW4ELgucA3gWdW1VeTPArYBhwzsNzzq2p8WXZEkjSroYYNcAowUVU3ASS5DNgMDIbNZuB1bfpK4C1JUlWfHehzPXBYkkOr6nsHUshRRx1VGzZsOJBFJemgtWPHjm9W1bq5+g07bI4Bbh14vwv4qZn6VNXeJHcBa+mObCY9B/jMlKD5kyT3An8B/I+aY1yeDRs2MD7ugZAkLUSSr8yn39Cv2SxWkpPoTq39ykDz86vq0cBPt9cLZlj2nCTjScb37NnTf7GSdJAadtjsBo4deL++tU3bJ8lq4HDgtvZ+PfAB4MyqunFygara3X5+G3gf3em6/VTVxVU1VlVj69bNeRQoSTpAww6bq4ETkmxMsgbYAmyd0mcrcFabPgO4qqoqyRHAXwLnVdUnJzsnWZ3kqDZ9CPAM4Lqe90OSNIuhhk1V7QXOpbuT7AvAFVV1fZILkjyrdXsnsDbJBPAKYPL26HOB44HXTLnF+VBgW5LPA9fQHRm9ffn2SpI0VXyeTWdsbKy8QUCSFibJjqoam6vfsE+jSZIOAoaNJKl3ho0kqXeGjSSpd4aNJKl3ho0kqXeGjSSpd4aNJKl3ho0kqXeGjSSpd4aNJKl3ho0kqXeGjSSpd4aNJKl3ho0kqXeGjSSpd4aNJKl3ho0kqXeGjSSpd4aNJKl3ho0kqXeGjSSpd4aNJKl3ho0kqXeGjSSpd4aNJKl3ho0kqXdDD5skm5LsTDKR5Lxp5h+a5PI2f3uSDa399CQ7klzbfv7cwDKPa+0TSd6UJMu3R5KkqYYaNklWARcBTwVOBJ6X5MQp3c4G7qiq44E3Ahe29m8Cz6yqRwNnAe8ZWOatwC8DJ7TXpt52QpI0p2Ef2ZwCTFTVTVX1feAyYPOUPpuBS9v0lcBpSVJVn62qr7b264HD2lHQ0cBDq+pTVVXAu4Fn978rkqSZDDtsjgFuHXi/q7VN26eq9gJ3AWun9HkO8Jmq+l7rv2uOdUqSltHqYRewWElOoju19uQDWPYc4ByA4447bokrkyRNGvaRzW7g2IH361vbtH2SrAYOB25r79cDHwDOrKobB/qvn2OdAFTVxVU1VlVj69atW+SuSJJmMuywuRo4IcnGJGuALcDWKX220t0AAHAGcFVVVZIjgL8EzquqT052rqqvAd9Kcmq7C+1M4EN974gkaWZDDZt2DeZcYBvwBeCKqro+yQVJntW6vRNYm2QCeAUweXv0ucDxwGuSXNNeD2/zXgq8A5gAbgT+enn2SJI0nXQ3bGlsbKzGx8eHXYYk/UBJsqOqxubqN+zTaJKkg4BhI0nqnWEjSeqdYSNJ6p1hI0nqnWEjSeqdYSNJ6p1hI0nqnWEjSeqdYSNJ6p1hI0nqnWEjSeqdYSNJ6p1hI0nqnWEjSeqdYSNJ6p1hI0nqnWEjSeqdYSNJ6p1hI0nqnWEjSeqdYSNJ6p1hI0nqnWEjSeqdYSNJ6p1hI0nqnWEjSerd0MMmyaYkO5NMJDlvmvmHJrm8zd+eZENrX5vk40m+k+QtU5b5RFvnNe318OXZG0nSdFYPc+NJVgEXAacDu4Crk2ytqhsGup0N3FFVxyfZAlwIPBf4F+DVwKPaa6rnV9V4rzsgSZqXYR/ZnAJMVNVNVfV94DJg85Q+m4FL2/SVwGlJUlXfrap/pAsdSdIIG3bYHAPcOvB+V2ubtk9V7QXuAtbOY91/0k6hvTpJpuuQ5Jwk40nG9+zZs/DqJUnzMuyw6cvzq+rRwE+31wum61RVF1fVWFWNrVu3blkLlKSDybDDZjdw7MD79a1t2j5JVgOHA7fNttKq2t1+fht4H93pOknSkAw7bK4GTkiyMckaYAuwdUqfrcBZbfoM4KqqqplWmGR1kqPa9CHAM4DrlrxySdK8DfVutKram+RcYBuwCrikqq5PcgEwXlVbgXcC70kyAdxOF0gAJLkFeCiwJsmzgScDXwG2taBZBXwUePsy7pYkaYrMcpBwUBkbG6vxce+UlqSFSLKjqsbm6jfs02iSpIOAYSNJ6p1hI0nqnWEjSeqdYSNJ6p1hI0nqnWEjSerdosImycVLVYgkaeWacwSBJEfONAt42tKWI0laieYzXM0euiFgBofpr/beJ2BKkuY0n7C5CTitqv556owkt07TX5Kkfcznms0fAg+bYd7rl7AWSdIKNeeRTVVdNMu8Ny9tOZKklWg+Nwj8x9nmV9X7l64cSdJKNJ9rNs+cZV4Bho0kaVbzOY32ovmsKMlZVXXp4kuSJK00SzmCwK8u4bokSSvIUoZN5u4iSToYLWXY+HxpSdK0PLKRJPVuKcPmk0u4LknSCjKfW5/vk+TpwEnAAyfbquqC9vPcpS1NkrRSzPvIJsnbgOcCL6c7ZfaLwI/0VJckaQVZyGm0J1TVmcAdVfVbwOOBH+unLEnSSrKQsLm7/fx/SR4B3AMcvfQlSZJWmoVcs/lwkiOANwCfobvV+R29VCVJWlHmfWRTVb9dVXdW1V/QXav58ap69WILSLIpyc4kE0nOm2b+oUkub/O3J9nQ2tcm+XiS7yR5y5RlHpfk2rbMm5J4W7YkDdGiRn1OsqhRn5OsAi4CTgd2AVcn2VpVNwx0O5vuOtHxSbYAF9LdqPAvwKuBR7XXoLcCvwxsB/4K2AT89YHWOZMPfnY3b9i2k6/eeTePOOIwXvmURwLs1/bsxxyz4PXMtcxKMer7Pur1LYWDYR+1v+X+3Bcy6vPDgScAV7X3Pwv8E4sb9fkUYKKqbgJIchmwGRgMm83A69r0lcBbkqSqvgv8Y5LjB1eY5GjgoVX1qfb+3cCzWeKw+eBnd3P++6/l7nvuBWD3nXfzyis/BwX3/Gvd13b++68FmPFDnG49cy2zUoz6vo96fUvhYNhH7W8Yn/ucp9Gq6kVt5OdDgBOr6jlV9Ry679scssjtHwMMPlp6V2ubtk9V7QXuAtbOsc5dc6xz0d6wbed9H9Ske+6t+4Jm0t333Msbtu1c0HrmWmalGPV9H/X6lsLBsI/a3zA+94XcjXZsVX1t4P3XgeOWuJ5lleScJONJxvfs2bOgZb96591zd5pH35nmLWT9P6hGfd9Hvb6lcDDso/Y3jM99IWHzsSTbkrwwyQuBvwQ+usjt7waOHXi/vrVN2yfJauBw4LY51rl+jnUCUFUXV9VYVY2tW7duQYU/4ojDlqTvTPMWsv4fVKO+76Ne31I4GPZR+xvG576Qu9HOBd4G/Nv2uriqXr7I7V8NnJBkY5I1wBZg65Q+W4Gz2vQZwFVVNeMI0+3o61tJTm13oZ0JfGiRde7nlU95JIcdsmqftkNWhUMesO+Nb4cdsuq+Gwfmu565llkpRn3fR72+pXAw7KP2N4zPfUFjowH/B/jX9rp6sRuvqr1JzgW2AauAS6rq+iQXAONVtRV4J/CeJBPA7XSBBECSW4CHAmuSPBt4cruT7aXAu4DD6G4MWPI70SYvoi32brSZ1nMwXJwd9X0f9fqWwsGwj9rfMD73zHKQsG/H5MXAa+juRgvwH4ALquqS3qpbRmNjYzU+Pj7sMiTpB0qSHVU1Nle/hRzZvBJ4TFXd1jawlu7W5xURNpKk/izkBoHbgG8PvP82s1+olyQJmN8IAq9okxPA9iQfohsXbTPw+R5rkyStEPM5jfaQ9vPG9pq05Hd4SZJWpjnDpj27RpKkAzbvGwSSjAGvohvx+b7lquone6hLkrSCLORutPfS3ZF2Ld33bCRJmpeFhM2e9iVLSZIWZCFh89ok7wA+BnxvsnExz7ORJB0cFhI2LwJ+nO6xApOn0YrFPc9GknQQWEjY/LuqcnQ+SdKCLWQEgX9KcmJvlUiSVqyFHNmcClyT5Ga6azYBylufJUlzWUjYbOqtCknSijbvsKmqrwAkeTjwwN4qkiStOPO+ZpPkWUm+DNwM/B1wCz08lEyStPIs5AaB36a7bvOlqtoInAZ8qpeqJEkrykLC5p724LQHJHlAVX0cmPPpbJIkLeQGgTuTPBj4e+C9Sb4BfLefsiRJK8lCjmw2A3cDvw78Dd2zbZ7ZR1GSpJVlIXejDR7FXNpDLZKkFWo+j4X+Nt0YaPvNovtS50OXvCpJ0ooynyd1PmSuPpIkzWYh12wkSTogho0kqXeGjSSpd4aNJKl3Qw+bJJuS7EwykeS8aeYfmuTyNn97kg0D885v7TuTPGWg/ZYk1ya5Jsn48uyJJGkmCxlBYMklWQVcBJwO7AKuTrK1qm4Y6HY2cEdVHZ9kC3Ah8Nz2ILctwEnAI4CPJvmxqrq3LfezVfXNZdsZSdKMhn1kcwowUVU3VdX3gcvoRioYtJn7v0R6JXBakrT2y6rqe1V1MzDR1idJGjHDDptjgFsH3u9qbdP2qaq9wF3A2jmWLeBvk+xIck4PdUuSFmCop9F69MSq2t0e9PaRJF+sqr+f2qkF0TkAxx133HLXKEkHjWEf2ewGjh14v761TdsnyWrgcOC22Zatqsmf3wA+wAyn16rq4qoaq6qxdevWLXpnJEnTG3bYXA2ckGRjkjV0F/y3TumzFTirTZ8BXFVV1dq3tLvVNgInAJ9O8qAkDwFI8iDgycB1y7AvkqQZDPU0WlXtTXIusA1YBVxSVdcnuQAYr6qtwDuB9ySZAG6nCyRavyuAG4C9wMuq6t4kPwx8oLuHgNXA+6rqb5Z95yRJ90l3kKCxsbEaH/crOZK0EEl2VNWcT20e9mk0SdJBwLCRJPXOsJEk9c6wkST1zrCRJPXOsJEk9c6wkST1zrCRJPXOsJEk9c6wkST1zrCRJPXOsJEk9c6wkST1zrCRJPXOsJEk9c6wkST1zrCRJPXOsJEk9c6wkST1zrCRJPXOsJEk9c6wkST1zrCRJPXOsJEk9c6wkST1zrCRJPXOsJEk9W7oYZNkU5KdSSaSnDfN/EOTXN7mb0+yYWDe+a19Z5KnzHedkqTlNdSwSbIKuAh4KnAi8LwkJ07pdjZwR1UdD7wRuLAteyKwBTgJ2AT8UZJV81ynJGkZDfvI5hRgoqpuqqrvA5cBm6f02Qxc2qavBE5LktZ+WVV9r6puBiba+uazTknSMhp22BwD3Drwfldrm7ZPVe0F7gLWzrLsfNYpSVpGww6boUpyTpLxJON79uwZdjmStGINO2x2A8cOvF/f2qbtk2Q1cDhw2yzLzmedAFTVxVU1VlVj69atW8RuSJJmM+ywuRo4IcnGJGvoLvhvndJnK3BWmz4DuKqqqrVvaXerbQROAD49z3VKkpbR6mFuvKr2JjkX2AasAi6pquuTXACMV9VW4J3Ae5JMALfThQet3xXADcBe4GVVdS/AdOtc7n2TJN0v3UGCxsbGanx8fNhlSNIPlCQ7qmpsrn7DPo0mSToIGDaSpN4ZNpKk3hk2kqTeGTaSpN4ZNpKk3hk2kqTeGTaSpN4ZNpKk3hk2kqTeGTaSpN4ZNpKk3hk2kqTeGTaSpN4ZNpKk3hk2kqTeGTaSpN4ZNpKk3hk2kqTeGTaSpN4ZNpKk3hk2kqTeGTaSpN4ZNpKk3hk2kqTeGTaSpN4ZNpKk3g0tbJIcmeQjSb7cfj5shn5ntT5fTnLWQPvjklybZCLJm5Kktb8uye4k17TX05ZrnyRJ0xvmkc15wMeq6gTgY+39PpIcCbwW+CngFOC1A6H0VuCXgRPaa9PAom+sqpPb66963AdJ0jwMM2w2A5e26UuBZ0/T5ynAR6rq9qq6A/gIsCnJ0cBDq+pTVVXAu2dYXpI0AoYZNj9cVV9r0/8X+OFp+hwD3DrwfldrO6ZNT22fdG6Szye5ZKbTc5Kk5dNr2CT5aJLrpnltHuzXjk5qiTb7VuDfACcDXwN+f5b6zkkynmR8z549S7R5SdJUq/tceVX9/Ezzknw9ydFV9bV2Wuwb03TbDTxp4P164BOtff2U9t1tm18f2MbbgQ/PUt/FwMWt754kX5ljl4bhKOCbwy5iBqNcG4x2fdZ24Ea5vlGuDfqp70fm06nXsJnDVuAs4Hfbzw9N02cb8D8HToU9GTi/qm5P8q0kpwLbgTOBNwNMBljr/wvAdfMppqrWHfCe9CjJeFWNDbuO6YxybTDa9VnbgRvl+ka5NhhufcMMm98FrkhyNvAV4D8BJBkDXlJVL26h8tvA1W2ZC6rq9jb9UuBdwGHAX7cXwOuTnEx3Wu4W4FeWYV8kSbMYWthU1W3AadO0jwMvHnh/CXDJDP0eNU37C5a2UknSYjmCwOi7eNgFzGKUa4PRrs/aDtwo1zfKtcEQ60t3I5gkSf3xyEaS1DvDZhm0L5d+I8l1A23Tjg2XzpvamG+fT/LYgWUWNE7cPGs7NsnHk9yQ5Pokvzpi9T0wyaeTfK7V91utfWOS7W2dlydZ09oPbe8n2vwNA+s6v7XvTPKUgfZNrW0iyX7DJs2jxlVJPpvkwyNY2y3tv/01ScZb26h8tkckuTLJF5N8IcnjR6G2JI/M/WMrXpPuztdfG4XaBpb/9fb/w3VJ/izd/ycj83s3rary1fML+BngscB1A22vB85r0+cBF7bpp9HdWRfgVGB7az8SuKn9fFibflib9+nWN23Zpy6gtqOBx7bphwBfAk4cofoCPLhNH0J3q/upwBXAltb+NuC/tOmXAm9r01uAy9v0icDngEOBjcCNwKr2uhH4UWBN63PiAj/fVwDvAz7c3o9SbbcAR01pG5XP9lLgxW16DXDEqNQ2UOMquhFOfmRUaqMbLeVm4LCB37cXjtLv3bR1L3YFvub9C7KBfcNmJ3B0mz4a2Nmm/xh43tR+wPOAPx5o/+PWdjTwxYH2ffodQJ0fAk4fxfqAHwI+Qzcw6zeB1a398cC2Nr0NeHybXt36BTif7jtaDPYbXLa179NvHjWtpxtI9ufovkCcUamtLXML+4fN0D9b4HC6fzAzarVNqefJwCdHqTbuH8bryPZ79GG6cSRH5vduupen0YZnprHhZhsP7kDGiZu3dnj9GLqjh5GpL91pqmvoRpn4CN1fXXdW1d5p1nlfHW3+XcDaA6h7vv4Q+A3gX9v7tSNUG3TfN/vbJDuSnNPaRuGz3QjsAf4k3SnIdyR50IjUNmgL8GdteiRqq6rdwO8B/0w3JNddwA5G6/duP4bNCKjuz4eh3haY5MHAXwC/VlXfGpw37Pqq6t6qOpnuKOIU4MeHVcugJM8AvlFVO4ZdyyyeWFWPBZ4KvCzJzwzOHOJnu5ru1PJbq+oxwHeZ8piRYf/etWsezwL+fOq8YdbWrhVtpgvsRwAPYt9HrIwkw2Z4vp5uTDiy79hwu4FjB/pNjvs2W/u048TNV5JD6ILmvVX1/lGrb1JV3Ql8nO4w/4gkk19KHlznfXW0+YcDtx1A3fPx74FnJbkFuIzuVNr/GpHagPv+CqaqvgF8gC6sR+Gz3QXsqqrt7f2VdOEzCrVNeirwmbp/vMVRqe3ngZurak9V3QO8n+53cWR+76a12PNwvuZ9nnUD+16zeQP7Xmx8fZt+OvtebPx0az+S7hz3w9rrZuDINm/qxcanLaCu0D0P6A+ntI9KfeuAI9r0YcA/AM+g+2tz8GLoS9v0y9j3YugVbfok9r0YehPdhdDVbXoj918MPekAPt8ncf8NAiNRG91fvA8ZmP4nur+AR+Wz/QfgkW36da2ukaitLX8Z8KIR/H/ip4Dr6a5hhu5Gi5ePyu/djHUvdgW+5vXL8Wd051bvofuL7my6c6YfA74MfHTglzDARXTXJa4FxgbW80vARHsN/k8wRjfg6I3AW5hy0XWO2p5Idzrg88A17fW0EarvJ4HPtvquA17T2n+0/Q870f4nO7S1P7C9n2jzf3RgXa9qNexk4O6ftr9favNedYCf8ZO4P2xGorZWx+fa6/rJ5Ufosz0ZGG+f7Qfp/kEeldoeRPfX/+EDbSNRW1v+t4AvtnW8hy4wRuL3bqaXIwhIknrnNRtJUu8MG0lS7wwbSVLvDBtJUu8MG0lS7wwbSVLvDBtpFkk2ZODREAe4jiclecIcfV6S5Mw2/cIkj1jMNmfb/uC2pOWyeu4ukhbpScB36L7BP62qetvA2xfSfVnvq/PdQJLVdf8gjLNuf8q2pGXhlzqlWbSRsP+GblTdx9J9E/9M4CeAPwAeTDdk+wur6mtJ/ivwEmAvcAPdsCafAu6lG+X45VX1D9Ns53V0gXAL8C66sajuphsH7sQZtvUJuhEfnkg3SsWXgP9ON8TIbcDz6Yb42Wf7wGnAd6rq95KcTDe0yQ/RfVv8l6rqjrbu7cDP0j1n5uzp6pbmy9No0tweCfxRVf0E8C26sabeDJxRVY8DLgF+p/U9D3hMVf0k8JKquoXuH/M3VtXJc/2DXVVX0g3h8vzqRrreO8u2ANZU1VhV/T7wj8Cp1Y2ifBnwG/PY/ruB32z1Xgu8dmDe6qo6Bfi1Ke3SgnkaTZrbrVX1yTb9p8B/Ax4FfKQ9zXcV3dh30I3z9d4kH6Qb72uxHjnLtgAuH5heD1zeRiReQzfw44ySHE43yOnftaZL2Xc4/ckRwHfQDSQrHTDDRprb1HPN3waur6rHT9P36XSPAX8m8Kokj17ktjPLtqB7DsykNwN/UFVbkzyJbiTlxfhe+3kv/luhRfI0mjS345JM/mP/n+mugaybbEtySJKTkjwAOLaqPg78Jt1zQx5MF04PWcD2BvvvnG5bMyx3OPc/d+SsGdZ3n6q6C7gjyU+3phcAfze1n7QUDBtpbjvpnnL5Bbph8N8MnAFcmORzdBfpn0B3iutPk1xL91iEN1X3wLf/DfxCkmsG/mGfzbuAt7VHYa+aYVvTeR3w50l20N1IMGm27Z8FvCHJ5+mG/L9gHvVJC+bdaJKk3nlkI0nqnRf9pGWU5FXAL05p/vOq+p3p+ksrhafRJEm98zSaJKl3ho0kqXeGjSSpd4aNJKl3ho0kqXf/H4XJ5cEnaKTNAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "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": [
    "# best iteration\n",
    "for c in df_gsearch_results.columns:\n",
    "    if c != 'best_iteration':\n",
    "        plt.scatter(df_gsearch_results['best_iteration'], df_gsearch_results[c])\n",
    "        plt.ylabel(c)\n",
    "    else:\n",
    "        plt.hist(df_gsearch_results['best_iteration'], bins=30)\n",
    "        plt.ylabel('N')        \n",
    "    plt.xlabel('best_iteration')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Feature importance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "ename": "TypeError",
     "evalue": "booster must be Booster or LGBMModel.",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mTypeError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-29-9312abb43a65>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mlgb\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mplot_importance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mgsearch_results\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m5\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfigsize\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m12\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m30\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mheight\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m0.5\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;31m# 050, 051, 052\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      2\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshow\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/lightgbm/plotting.py\u001b[0m in \u001b[0;36mplot_importance\u001b[0;34m(booster, ax, height, xlim, ylim, title, xlabel, ylabel, importance_type, max_num_features, ignore_zero, figsize, grid, **kwargs)\u001b[0m\n\u001b[1;32m     79\u001b[0m         \u001b[0mbooster\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mbooster\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mbooster_\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     80\u001b[0m     \u001b[0;32melif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0misinstance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mbooster\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mBooster\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 81\u001b[0;31m         \u001b[0;32mraise\u001b[0m \u001b[0mTypeError\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'booster must be Booster or LGBMModel.'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     82\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     83\u001b[0m     \u001b[0mimportance\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mbooster\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfeature_importance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mimportance_type\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mimportance_type\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mTypeError\u001b[0m: booster must be Booster or LGBMModel."
     ]
    }
   ],
   "source": [
    "lgb.plot_importance(gsearch_results[0][5][0], figsize=(12, 30), height=0.5) # 050, 051, 052\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "ename": "TypeError",
     "evalue": "booster must be Booster or LGBMModel.",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mTypeError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-30-d98df6a9b2c1>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mlgb\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mplot_importance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mgsearch_results\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m5\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfigsize\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m12\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m30\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mheight\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m0.5\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mimportance_type\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'gain'\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;31m# 050, 051, 052\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      2\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshow\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/lightgbm/plotting.py\u001b[0m in \u001b[0;36mplot_importance\u001b[0;34m(booster, ax, height, xlim, ylim, title, xlabel, ylabel, importance_type, max_num_features, ignore_zero, figsize, grid, **kwargs)\u001b[0m\n\u001b[1;32m     79\u001b[0m         \u001b[0mbooster\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mbooster\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mbooster_\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     80\u001b[0m     \u001b[0;32melif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0misinstance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mbooster\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mBooster\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 81\u001b[0;31m         \u001b[0;32mraise\u001b[0m \u001b[0mTypeError\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'booster must be Booster or LGBMModel.'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     82\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     83\u001b[0m     \u001b[0mimportance\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mbooster\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfeature_importance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mimportance_type\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mimportance_type\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mTypeError\u001b[0m: booster must be Booster or LGBMModel."
     ]
    }
   ],
   "source": [
    "lgb.plot_importance(gsearch_results[0][5][0], figsize=(12, 30), height=0.5, importance_type='gain') # 050, 051, 052\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "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": "3.6.6"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": true,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {
    "height": "calc(100% - 180px)",
    "left": "10px",
    "top": "150px",
    "width": "193.797px"
   },
   "toc_section_display": true,
   "toc_window_display": true
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}