Python_scripts/Averages_Datasets_Light_copy.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-10-15T16:37:36.446675733Z",
     "start_time": "2023-10-15T16:37:36.338837792Z"
    }
   },
   "outputs": [],
   "source": [
    "import os\n",
    "from IPython.display import display, HTML\n",
    "from bs4 import BeautifulSoup\n",
    "import pandas as pd\n",
    "import eli5\n",
    "from eli5.sklearn import PermutationImportance\n",
    "import pandas as pd\n",
    "from scipy.stats import zscore\n",
    "import numpy as np\n",
    "from numpy import mean\n",
    "from numpy import std\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.model_selection import RepeatedStratifiedKFold\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.model_selection import KFold\n",
    "from itertools import combinations\n",
    "from pandas import DataFrame\n",
    "from sklearn.preprocessing import PolynomialFeatures\n",
    "from sklearn.ensemble import GradientBoostingRegressor\n",
    "import lightgbm as xgb\n",
    "from lightgbm import LGBMRegressor\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.model_selection import cross_val_score, KFold\n",
    "from sklearn.metrics import mean_squared_error\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.model_selection import RepeatedKFold\n",
    "from numpy import absolute\n",
    "from sklearn.model_selection import RandomizedSearchCV\n",
    "from sklearn.metrics import mean_absolute_error\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "import pandas as pd\n",
    "from sklearn.model_selection import train_test_split\n",
    "from lightgbm import LGBMRegressor\n",
    "from sklearn.metrics import mean_absolute_error, mean_squared_error, r2_score\n",
    "import random\n",
    "import shap\n",
    "import re\n",
    "random_seed = 3031\n",
    "\n",
    "np.random.seed(random_seed)\n",
    "random.seed(random_seed)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Prepare the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-10-15T16:37:37.863551893Z",
     "start_time": "2023-10-15T16:37:36.360388600Z"
    }
   },
   "outputs": [],
   "source": [
    "v1 = pd.read_excel(\"/home/nick/PycharmProjects/Engineering_EKB/data/new_dataset_1.xlsx\")\n",
    "v3 = pd.read_excel(\"/home/nick/PycharmProjects/Engineering_EKB/data/new_dataset_3.xlsx\")\n",
    "v4 = pd.read_excel(\"/home/nick/PycharmProjects/Engineering_EKB/data/new_dataset_4.xlsx\")\n",
    "v5 = pd.read_excel(\"/home/nick/PycharmProjects/Engineering_EKB/data/new_dataset_5.xlsx\")\n",
    "v6 = pd.read_excel(\"/home/nick/PycharmProjects/Engineering_EKB/data/new_dataset_6.xlsx\")\n",
    "\n",
    "vavg = pd.concat([v1, v3, v4, v5, v6]).groupby(level=0).mean()\n",
    "vavg = vavg.drop(['DELTA30'], axis=1)\n",
    "df = pd.read_excel('/home/nick/PycharmProjects/Engineering_EKB/data/new_df_66.xlsx', engine='openpyxl')\n",
    "seq_and_y = df.iloc[:, 126:]\n",
    "vavg = pd.concat([vavg, seq_and_y], axis=1)\n",
    "\n",
    "#\n",
    "# Q1 = np.percentile(vavg['DELTA30'], 5)\n",
    "# Q3 = np.percentile(vavg['DELTA30'], 95)\n",
    "# IQR = Q3 - Q1\n",
    "#\n",
    "# # Define the lower and upper bounds for outlier detection\n",
    "# lower_bound = Q1 - 1.5 * IQR\n",
    "# upper_bound = Q3 + 1.5 * IQR\n",
    "# # Filter out the outliers\n",
    "# data_filtered = vavg[(vavg['DELTA30'] >= lower_bound) & (v1['DELTA30'] <= upper_bound)]\n",
    "#\n",
    "# # Separate the filtered X and Y\n",
    "# X = data_filtered.drop('DELTA30', axis=1)\n",
    "# Y = data_filtered['DELTA30']\n",
    "# #\n",
    "X = vavg.drop(['DELTA30'], axis=1)\n",
    "Y = vavg['DELTA30']\n",
    "\n",
    "\n",
    "scaler = MinMaxScaler()\n",
    "X_scaled = scaler.fit_transform(X)\n",
    "X = pd.DataFrame(X_scaled, index=X.index, columns=X.columns)\n",
    "# if hte feature ends with .1 remove the .1\n",
    "X.columns = [col.split('.1')[0] if col != 'h_3.10' else col for col in X.columns]\n",
    "data_clean = pd.concat([X, Y], axis=1)\n",
    "\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-10-15T16:37:37.866205148Z",
     "start_time": "2023-10-15T16:37:37.863885505Z"
    }
   },
   "outputs": [],
   "source": [
    "#get all the columns up to val_st\n",
    "Seq_MD_MDpocket = X\n",
    "MDpocket = X.iloc[:, :80]\n",
    "MD = X.iloc[:, 80:112]\n",
    "Seq = X.iloc[:, 112:]\n",
    "Seq_MD = X.iloc[:, 80:]\n",
    "MD_MDpocket = X.iloc[:, :112]\n",
    "Seq_MDpocket = pd.concat([MDpocket, Seq], axis =1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Split train and test sets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-10-15T16:37:37.866381871Z",
     "start_time": "2023-10-15T16:37:37.864179168Z"
    }
   },
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, Y, test_size=0.2, random_state=random_seed)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-10-15T16:37:38.056006858Z",
     "start_time": "2023-10-15T16:37:37.864405928Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": "<Figure size 640x480 with 1 Axes>",
      "image/png": "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"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot the distribution of y_test on top of the distribution of y_train\n",
    "fig, ax = plt.subplots()\n",
    "ax.hist(y_train, bins=20, alpha=0.5, label='FCA training set')\n",
    "ax.hist(y_test, bins=20, alpha=0.5, label='FCA test set')\n",
    "# label x axis as 'FAC'\n",
    "ax.set_xlabel('FCA')\n",
    "# label y axis as 'count'\n",
    "ax.set_ylabel('Count')\n",
    "ax.legend(loc='upper right')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-10-15T16:37:38.056345515Z",
     "start_time": "2023-10-15T16:37:38.051649991Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "y_train negative values:  77\n",
      "y_test negative values:  21\n",
      "y_train positive values:  172\n",
      "y_test positive values:  42\n"
     ]
    }
   ],
   "source": [
    "# count how many negative y are in the train and test sets\n",
    "print('y_train negative values: ', sum(y_train < 0))\n",
    "print('y_test negative values: ', sum(y_test < 0))\n",
    "print('y_train positive values: ', sum(y_train >= 0))\n",
    "print('y_test positive values: ', sum(y_test >= 0))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-10-15T16:37:38.056512715Z",
     "start_time": "2023-10-15T16:37:38.051896720Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "y_train average:  0.6895582329317269\n",
      "y_test average:  0.6968253968253968\n"
     ]
    }
   ],
   "source": [
    "# print average of y_train and y_test\n",
    "print('y_train average: ', np.mean(y_train))\n",
    "print('y_test average: ', np.mean(y_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-10-15T16:37:38.056652075Z",
     "start_time": "2023-10-15T16:37:38.052093816Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "y_train standard deviation:  1.478592692740836\n"
     ]
    }
   ],
   "source": [
    "# print y_train standard deviation\n",
    "print('y_train standard deviation: ', np.std(y_train))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Parameter search space"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-10-15T16:37:38.056739780Z",
     "start_time": "2023-10-15T16:37:38.052284588Z"
    }
   },
   "outputs": [],
   "source": [
    "parameters = {'learning_rate': [0.001,0.004,0.008,0.01,0.02,0.03,0.04,0.05,0.06,0.07,\n",
    "                                0.08,0.09,0.1,0.2,0.4,0.6,0.8,1],\n",
    "              'max_depth': range(1,100,10),\n",
    "              'num_leaves': range(2,100,10),\n",
    "              'colsample_bytree': [0.1,0.2,0.4,0.8,1],\n",
    "              'min_split_gain': range(0,20,1),\n",
    "              'n_estimators': range(10,1000,10),\n",
    "              \"reg_alpha\"   : [0.1,0.2,0.4,0.6,0.8,1],\n",
    "              \"reg_lambda\"  : range(1,10,1)\n",
    "          }"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Randomized search"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "######## COMPARING DATASET ORIGIN #################"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-10-15T16:43:00.028872720Z",
     "start_time": "2023-10-15T16:37:38.052443425Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "/usr/ulocal/anaconda3/envs/Engineering_EKB/lib/python3.8/site-packages/lightgbm/sklearn.py:736: UserWarning: 'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'verbose' argument is deprecated and will be removed in a future release of LightGBM. \"\n",
      "'verbose' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n"
     ]
    }
   ],
   "source": [
    "# Create an empty DataFrame\n",
    "datasets = [Seq, Seq_MD, Seq_MD_MDpocket, Seq_MDpocket, MD_MDpocket, MD, MDpocket]\n",
    "best_params = []\n",
    "best_estimators = []\n",
    "all_results = []\n",
    "results_table = pd.DataFrame(index=['Seq', 'Seq_MD', 'Seq_MD_MDpocket', 'Seq_MDpocket', 'MD_MDpocket', 'MD', 'MDPocket'])\n",
    "r2_values = []\n",
    "rmse_values = []\n",
    "mae_values = []\n",
    "\n",
    "def mape(actual, pred):\n",
    "    actual, pred = np.array(actual), np.array(pred)\n",
    "    return np.mean(np.abs((actual - pred) / actual)) * 100\n",
    "# Iterate over datasets\n",
    "for i, dataset in enumerate(datasets):\n",
    "    X_train, X_test, y_train, y_test = train_test_split(dataset, Y, test_size=0.2, random_state=random_seed)\n",
    "    xgb_model = LGBMRegressor(random_state=random_seed)\n",
    "\n",
    "    grid_obj_xgb = RandomizedSearchCV(xgb_model, parameters, cv=5, n_iter=15,\n",
    "                                      scoring='neg_mean_absolute_error', verbose=0, n_jobs=12, random_state=random_seed) #random_state=7)\n",
    "    grid_obj_xgb.fit(X_train, y_train, verbose=1)\n",
    "\n",
    "    y_pred_train = grid_obj_xgb.predict(X_train)\n",
    "    y_pred_test = grid_obj_xgb.predict(X_test)\n",
    "\n",
    "    err_xgb_train = mean_absolute_error(y_train, y_pred_train, multioutput='raw_values')\n",
    "    best_estimators.append(grid_obj_xgb.best_estimator_)\n",
    "    best_params.append(grid_obj_xgb.best_params_)\n",
    "\n",
    "    # Evaluation\n",
    "    results = {}\n",
    "    results['best_estimator'] = grid_obj_xgb.best_estimator_\n",
    "\n",
    "    # Predict on test set\n",
    "    y_pred = grid_obj_xgb.predict(X_test)\n",
    "\n",
    "    # Calculate metrics\n",
    "    results['r2_score'] = r2_score(y_test, y_pred)\n",
    "    r2_values.append(r2_score(y_test, y_pred))\n",
    "    #results['mean_absolute_perc_error'] = mape(y_test, y_pred)\n",
    "    #results['std_absolute_perc_error'] = np.std(np.abs((y_test - y_pred) / y_pred))\n",
    "    results['mean_absolute_error'] = mean_absolute_error(y_test, y_pred)\n",
    "    # get the standard deviation of the MAEs\n",
    "    results['std_absolute_error'] = np.std(np.abs(y_test - y_pred))\n",
    "    results['root_mean_squared_error'] = np.sqrt(mean_squared_error(y_test, y_pred))\n",
    "    # get std of sequared errors\n",
    "    results['std_squared_error'] = np.std((y_test - y_pred) ** 2)\n",
    "\n",
    "    # Calculate residuals\n",
    "    residuals = y_test - y_pred\n",
    "    max_error = max(residuals) if abs(max(residuals)) > abs(min(residuals)) else min(residuals)\n",
    "    max_idx = list(residuals).index(max(residuals)) if abs(max(residuals)) > abs(min(residuals)) else list(\n",
    "        residuals).index(min(residuals))\n",
    "    max_true, max_pred = y_test.iloc[max_idx], y_pred[max_idx]\n",
    "    results['max_error'] = abs(max_error)\n",
    "    results['max_true'] = max_true\n",
    "    results['max_predicted'] = max_pred\n",
    "    results['residuals_std'] = np.std(residuals)\n",
    "    results['average_residuals'] = np.mean(abs(residuals))\n",
    "    results['std_deviation'] = np.std(residuals)\n",
    "    rmse_values.append(np.sqrt(mean_squared_error(y_test, y_pred)))\n",
    "    mae_values.append(mean_absolute_error(y_test, y_pred))\n",
    "    all_results.append(results)\n",
    "\n",
    "    # fig, ax = plt.subplots()\n",
    "    # ax.hist(y_pred_test, bins=20, alpha=0.5, label='Predicted Y')\n",
    "    # ax.hist(y_test, bins=20, alpha=0.5, label='True Y')\n",
    "    # ax.legend(loc='upper right')\n",
    "    # plt.show()\n",
    "\n",
    "# Create the results table\n",
    "results_dict = {\n",
    "    'Best Estimator': [str(result['best_estimator']) for result in all_results],\n",
    "    'R2 Score': [round(result['r2_score'], 2) for result in all_results],\n",
    "    #'Mean Absolute Perc Error': [round(result['mean_absolute_perc_error'], 2) for result in all_results],\n",
    "    #'Std Absolute Perc Error': [round(result['std_absolute_perc_error'], 2) for result in all_results],\n",
    "    'Mean Absolute Error': [\"{:,.2f}\".format(result['mean_absolute_error']) for result in all_results],\n",
    "    'Std Absolute Error': [\"{:,.2f}\".format(result['std_absolute_error']) for result in all_results],\n",
    "    'Root Mean Squared Error': [\"{:,.2f}\".format(result['root_mean_squared_error']) for result in all_results],\n",
    "    'Std Squared Error': [\"{:,.2f}\".format(result['std_squared_error']) for result in all_results],\n",
    "    'Max Error': [\"{:,.2f}\".format(result['max_error']) for result in all_results],\n",
    "    'Max True': [\"{:,.2f}\".format(result['max_true']) for result in all_results],\n",
    "    'Max Predicted': [\"{:,.2f}\".format(result['max_predicted']) for result in all_results],\n",
    "    'Average Error': [result['average_residuals'] for result in all_results],\n",
    "    'Residuals Std': [\"{:,.2f}\".format(result['residuals_std']) for result in all_results],\n",
    "}\n",
    "\n",
    "results_table = pd.DataFrame(results_dict, index=results_table.index)\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "outputs": [
    {
     "data": {
      "text/plain": "                                                    Best Estimator  R2 Score  \\\nSeq              LGBMRegressor(colsample_bytree=0.8, learning_r...      0.47   \nSeq_MD           LGBMRegressor(colsample_bytree=0.8, learning_r...      0.50   \nSeq_MD_MDpocket  LGBMRegressor(colsample_bytree=0.8, learning_r...      0.45   \nSeq_MDpocket     LGBMRegressor(colsample_bytree=0.8, learning_r...      0.41   \nMD_MDpocket      LGBMRegressor(colsample_bytree=0.8, learning_r...      0.21   \nMD               LGBMRegressor(colsample_bytree=0.8, learning_r...      0.22   \nMDPocket         LGBMRegressor(colsample_bytree=0.8, learning_r...      0.20   \n\n                Mean Absolute Error Std Absolute Error  \\\nSeq                            0.85               0.86   \nSeq_MD                         0.83               0.84   \nSeq_MD_MDpocket                0.84               0.91   \nSeq_MDpocket                   0.86               0.95   \nMD_MDpocket                    1.08               1.02   \nMD                             1.08               1.00   \nMDPocket                       1.16               0.94   \n\n                Root Mean Squared Error Std Squared Error Max Error Max True  \\\nSeq                                1.21              3.29      4.60    -4.20   \nSeq_MD                             1.18              3.19      4.51    -4.20   \nSeq_MD_MDpocket                    1.24              3.49      4.78    -4.20   \nSeq_MDpocket                       1.28              3.73      4.95    -4.20   \nMD_MDpocket                        1.48              4.35      5.12    -4.20   \nMD                                 1.47              4.28      5.00    -4.20   \nMDPocket                           1.49              3.71      4.51    -4.20   \n\n                Max Predicted  Average Error Residuals Std  \nSeq                      0.40       0.851938          1.21  \nSeq_MD                   0.31       0.829303          1.18  \nSeq_MD_MDpocket          0.58       0.836820          1.24  \nSeq_MDpocket             0.75       0.857435          1.28  \nMD_MDpocket              0.92       1.076091          1.48  \nMD                       0.80       1.076739          1.47  \nMDPocket                 0.31       1.159177          1.49  ",
      "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>Best Estimator</th>\n      <th>R2 Score</th>\n      <th>Mean Absolute Error</th>\n      <th>Std Absolute Error</th>\n      <th>Root Mean Squared Error</th>\n      <th>Std Squared Error</th>\n      <th>Max Error</th>\n      <th>Max True</th>\n      <th>Max Predicted</th>\n      <th>Average Error</th>\n      <th>Residuals Std</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>Seq</th>\n      <td>LGBMRegressor(colsample_bytree=0.8, learning_r...</td>\n      <td>0.47</td>\n      <td>0.85</td>\n      <td>0.86</td>\n      <td>1.21</td>\n      <td>3.29</td>\n      <td>4.60</td>\n      <td>-4.20</td>\n      <td>0.40</td>\n      <td>0.851938</td>\n      <td>1.21</td>\n    </tr>\n    <tr>\n      <th>Seq_MD</th>\n      <td>LGBMRegressor(colsample_bytree=0.8, learning_r...</td>\n      <td>0.50</td>\n      <td>0.83</td>\n      <td>0.84</td>\n      <td>1.18</td>\n      <td>3.19</td>\n      <td>4.51</td>\n      <td>-4.20</td>\n      <td>0.31</td>\n      <td>0.829303</td>\n      <td>1.18</td>\n    </tr>\n    <tr>\n      <th>Seq_MD_MDpocket</th>\n      <td>LGBMRegressor(colsample_bytree=0.8, learning_r...</td>\n      <td>0.45</td>\n      <td>0.84</td>\n      <td>0.91</td>\n      <td>1.24</td>\n      <td>3.49</td>\n      <td>4.78</td>\n      <td>-4.20</td>\n      <td>0.58</td>\n      <td>0.836820</td>\n      <td>1.24</td>\n    </tr>\n    <tr>\n      <th>Seq_MDpocket</th>\n      <td>LGBMRegressor(colsample_bytree=0.8, learning_r...</td>\n      <td>0.41</td>\n      <td>0.86</td>\n      <td>0.95</td>\n      <td>1.28</td>\n      <td>3.73</td>\n      <td>4.95</td>\n      <td>-4.20</td>\n      <td>0.75</td>\n      <td>0.857435</td>\n      <td>1.28</td>\n    </tr>\n    <tr>\n      <th>MD_MDpocket</th>\n      <td>LGBMRegressor(colsample_bytree=0.8, learning_r...</td>\n      <td>0.21</td>\n      <td>1.08</td>\n      <td>1.02</td>\n      <td>1.48</td>\n      <td>4.35</td>\n      <td>5.12</td>\n      <td>-4.20</td>\n      <td>0.92</td>\n      <td>1.076091</td>\n      <td>1.48</td>\n    </tr>\n    <tr>\n      <th>MD</th>\n      <td>LGBMRegressor(colsample_bytree=0.8, learning_r...</td>\n      <td>0.22</td>\n      <td>1.08</td>\n      <td>1.00</td>\n      <td>1.47</td>\n      <td>4.28</td>\n      <td>5.00</td>\n      <td>-4.20</td>\n      <td>0.80</td>\n      <td>1.076739</td>\n      <td>1.47</td>\n    </tr>\n    <tr>\n      <th>MDPocket</th>\n      <td>LGBMRegressor(colsample_bytree=0.8, learning_r...</td>\n      <td>0.20</td>\n      <td>1.16</td>\n      <td>0.94</td>\n      <td>1.49</td>\n      <td>3.71</td>\n      <td>4.51</td>\n      <td>-4.20</td>\n      <td>0.31</td>\n      <td>1.159177</td>\n      <td>1.49</td>\n    </tr>\n  </tbody>\n</table>\n</div>"
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_table"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2023-10-15T16:43:00.029165597Z",
     "start_time": "2023-10-15T16:43:00.027545750Z"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-10-15T16:43:00.029406324Z",
     "start_time": "2023-10-15T16:43:00.027824240Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": "                                                    Best Estimator  R2 Score  \\\nSeq              LGBMRegressor(colsample_bytree=0.8, learning_r...      0.47   \nSeq_MD           LGBMRegressor(colsample_bytree=0.8, learning_r...      0.50   \nSeq_MD_MDpocket  LGBMRegressor(colsample_bytree=0.8, learning_r...      0.45   \nSeq_MDpocket     LGBMRegressor(colsample_bytree=0.8, learning_r...      0.41   \nMD_MDpocket      LGBMRegressor(colsample_bytree=0.8, learning_r...      0.21   \nMD               LGBMRegressor(colsample_bytree=0.8, learning_r...      0.22   \nMDPocket         LGBMRegressor(colsample_bytree=0.8, learning_r...      0.20   \n\n                Mean Absolute Error Std Absolute Error  \\\nSeq                            0.85               0.86   \nSeq_MD                         0.83               0.84   \nSeq_MD_MDpocket                0.84               0.91   \nSeq_MDpocket                   0.86               0.95   \nMD_MDpocket                    1.08               1.02   \nMD                             1.08               1.00   \nMDPocket                       1.16               0.94   \n\n                Root Mean Squared Error Std Squared Error Max Error Max True  \\\nSeq                                1.21              3.29      4.60    -4.20   \nSeq_MD                             1.18              3.19      4.51    -4.20   \nSeq_MD_MDpocket                    1.24              3.49      4.78    -4.20   \nSeq_MDpocket                       1.28              3.73      4.95    -4.20   \nMD_MDpocket                        1.48              4.35      5.12    -4.20   \nMD                                 1.47              4.28      5.00    -4.20   \nMDPocket                           1.49              3.71      4.51    -4.20   \n\n                Max Predicted  Average Error Residuals Std  \nSeq                      0.40       0.851938          1.21  \nSeq_MD                   0.31       0.829303          1.18  \nSeq_MD_MDpocket          0.58       0.836820          1.24  \nSeq_MDpocket             0.75       0.857435          1.28  \nMD_MDpocket              0.92       1.076091          1.48  \nMD                       0.80       1.076739          1.47  \nMDPocket                 0.31       1.159177          1.49  ",
      "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>Best Estimator</th>\n      <th>R2 Score</th>\n      <th>Mean Absolute Error</th>\n      <th>Std Absolute Error</th>\n      <th>Root Mean Squared Error</th>\n      <th>Std Squared Error</th>\n      <th>Max Error</th>\n      <th>Max True</th>\n      <th>Max Predicted</th>\n      <th>Average Error</th>\n      <th>Residuals Std</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>Seq</th>\n      <td>LGBMRegressor(colsample_bytree=0.8, learning_r...</td>\n      <td>0.47</td>\n      <td>0.85</td>\n      <td>0.86</td>\n      <td>1.21</td>\n      <td>3.29</td>\n      <td>4.60</td>\n      <td>-4.20</td>\n      <td>0.40</td>\n      <td>0.851938</td>\n      <td>1.21</td>\n    </tr>\n    <tr>\n      <th>Seq_MD</th>\n      <td>LGBMRegressor(colsample_bytree=0.8, learning_r...</td>\n      <td>0.50</td>\n      <td>0.83</td>\n      <td>0.84</td>\n      <td>1.18</td>\n      <td>3.19</td>\n      <td>4.51</td>\n      <td>-4.20</td>\n      <td>0.31</td>\n      <td>0.829303</td>\n      <td>1.18</td>\n    </tr>\n    <tr>\n      <th>Seq_MD_MDpocket</th>\n      <td>LGBMRegressor(colsample_bytree=0.8, learning_r...</td>\n      <td>0.45</td>\n      <td>0.84</td>\n      <td>0.91</td>\n      <td>1.24</td>\n      <td>3.49</td>\n      <td>4.78</td>\n      <td>-4.20</td>\n      <td>0.58</td>\n      <td>0.836820</td>\n      <td>1.24</td>\n    </tr>\n    <tr>\n      <th>Seq_MDpocket</th>\n      <td>LGBMRegressor(colsample_bytree=0.8, learning_r...</td>\n      <td>0.41</td>\n      <td>0.86</td>\n      <td>0.95</td>\n      <td>1.28</td>\n      <td>3.73</td>\n      <td>4.95</td>\n      <td>-4.20</td>\n      <td>0.75</td>\n      <td>0.857435</td>\n      <td>1.28</td>\n    </tr>\n    <tr>\n      <th>MD_MDpocket</th>\n      <td>LGBMRegressor(colsample_bytree=0.8, learning_r...</td>\n      <td>0.21</td>\n      <td>1.08</td>\n      <td>1.02</td>\n      <td>1.48</td>\n      <td>4.35</td>\n      <td>5.12</td>\n      <td>-4.20</td>\n      <td>0.92</td>\n      <td>1.076091</td>\n      <td>1.48</td>\n    </tr>\n    <tr>\n      <th>MD</th>\n      <td>LGBMRegressor(colsample_bytree=0.8, learning_r...</td>\n      <td>0.22</td>\n      <td>1.08</td>\n      <td>1.00</td>\n      <td>1.47</td>\n      <td>4.28</td>\n      <td>5.00</td>\n      <td>-4.20</td>\n      <td>0.80</td>\n      <td>1.076739</td>\n      <td>1.47</td>\n    </tr>\n    <tr>\n      <th>MDPocket</th>\n      <td>LGBMRegressor(colsample_bytree=0.8, learning_r...</td>\n      <td>0.20</td>\n      <td>1.16</td>\n      <td>0.94</td>\n      <td>1.49</td>\n      <td>3.71</td>\n      <td>4.51</td>\n      <td>-4.20</td>\n      <td>0.31</td>\n      <td>1.159177</td>\n      <td>1.49</td>\n    </tr>\n  </tbody>\n</table>\n</div>"
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_table"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "outputs": [
    {
     "data": {
      "text/plain": "'LGBMRegressor(colsample_bytree=0.8, learning_rate=0.008, max_depth=41,\\n              min_split_gain=1, n_estimators=520, num_leaves=42,\\n              random_state=3031, reg_alpha=0.6, reg_lambda=7)'"
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_table['Best Estimator']['Seq_MD']"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2023-10-16T08:12:23.857399066Z",
     "start_time": "2023-10-16T08:12:23.856965337Z"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Bootstrap for Seq: 100%|██████████| 500/500 [7:22:01<00:00, 53.04s/it]   \n",
      "Bootstrap for Seq_MD: 100%|██████████| 500/500 [9:42:36<00:00, 69.91s/it]    \n",
      "Bootstrap for Seq_MD_MDpocket: 100%|██████████| 500/500 [11:32:47<00:00, 83.13s/it]   \n",
      "Bootstrap for Seq_MDpocket: 100%|██████████| 500/500 [10:12:03<00:00, 73.45s/it]  \n",
      "Bootstrap for MD_MDpocket: 100%|██████████| 500/500 [2:51:59<00:00, 20.64s/it]  \n",
      "Bootstrap for MD: 100%|██████████| 500/500 [2:12:34<00:00, 15.91s/it]  \n",
      "Bootstrap for MDPocket: 100%|██████████| 500/500 [5:58:05<00:00, 42.97s/it]  \n"
     ]
    }
   ],
   "source": [
    "dataset_names = ['Seq', 'Seq_MD', 'Seq_MD_MDpocket', 'Seq_MDpocket', 'MD_MDpocket', 'MD', 'MDPocket']\n",
    "\n",
    "# Lists to store mean and std values for each metric\n",
    "mean_metrics = []\n",
    "std_metrics = []\n",
    "\n",
    "# Initialize lists to store R2, RMSE, and MAE dataframes for each dataset\n",
    "r2_values_bs = []\n",
    "rmse_values_bs = []\n",
    "mae_values_bs = []\n",
    "\n",
    "# Iterate over datasets with a progress bar\n",
    "for dataset, dataset_name in zip(datasets, dataset_names):\n",
    "    X_train, X_test, y_train, y_test = train_test_split(dataset, Y, test_size=0.2, random_state=random_seed)\n",
    "\n",
    "    # Get hyperparameters for this dataset\n",
    "    param_string = results_table['Best Estimator'][dataset_name]\n",
    "    params = dict(re.findall(r\"(\\w+)=(.*?)(?:(?:,|$|\\n\\s+)|\\))\", param_string))\n",
    "    reg_lambda = params.get('reg_lambda', '')\n",
    "    params['reg_lambda'] = reg_lambda.strip(\")\")\n",
    "\n",
    "    # Convert values to their appropriate data types if needed\n",
    "    params = {k: eval(v) if v.isdigit() else v for k, v in params.items()}\n",
    "\n",
    "    # Initialize lists to store R2, RMSE, and MAE for this dataset\n",
    "    bootstrap_r2_scores = []\n",
    "    bootstrap_rmse_scores = []\n",
    "    bootstrap_mae_scores = []\n",
    "\n",
    "    # Number of bootstrap iterations\n",
    "    n_bootstrap_iterations = 500\n",
    "\n",
    "    # Create a tqdm progress bar for the loop\n",
    "    for _ in tqdm(range(n_bootstrap_iterations), desc=f'Bootstrap for {dataset_name}'):\n",
    "        # Create a bootstrap sample by randomly selecting with replacement\n",
    "        bootstrap_indices = np.random.choice(len(X_train), size=len(X_train), replace=True)\n",
    "        X_train_bootstrap = X_train.iloc[bootstrap_indices]\n",
    "        y_train_bootstrap = y_train.iloc[bootstrap_indices]\n",
    "\n",
    "        # Train your model on the bootstrap sample with hyperparameters\n",
    "        xgb_model = LGBMRegressor(**params)\n",
    "        xgb_model.fit(X_train_bootstrap, y_train_bootstrap)\n",
    "\n",
    "        # Make predictions on the test set\n",
    "        y_pred = xgb_model.predict(X_test)\n",
    "\n",
    "        # Calculate metrics for this bootstrap iteration\n",
    "        r2 = r2_score(y_test, y_pred)\n",
    "        rmse = np.sqrt(mean_squared_error(y_test, y_pred))\n",
    "        mae = mean_absolute_error(y_test, y_pred)\n",
    "\n",
    "        # Append the metrics to their respective lists for this dataset\n",
    "        bootstrap_r2_scores.append(r2)\n",
    "        bootstrap_rmse_scores.append(rmse)\n",
    "        bootstrap_mae_scores.append(mae)\n",
    "\n",
    "    # Create dataframes to store R2, RMSE, and MAE for this dataset\n",
    "    r2_df = pd.DataFrame({'R2': bootstrap_r2_scores})\n",
    "    rmse_df = pd.DataFrame({'RMSE': bootstrap_rmse_scores})\n",
    "    mae_df = pd.DataFrame({'MAE': bootstrap_mae_scores})\n",
    "\n",
    "    # Append the dataframes to the lists\n",
    "    r2_values_bs.append(r2_df)\n",
    "    rmse_values_bs.append(rmse_df)\n",
    "    mae_values_bs.append(mae_df)"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2023-10-18T11:42:46.200611902Z",
     "start_time": "2023-10-16T09:50:37.503584508Z"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "outputs": [],
   "source": [
    "# # Import necessary library\n",
    "# import pandas as pd\n",
    "#\n",
    "# # Create an Excel writer object\n",
    "# with pd.ExcelWriter('metrics.xlsx') as writer:\n",
    "#     # Iterate through the lists and dataset names and save each DataFrame to a different sheet\n",
    "#     for dataset_name, r2_df, rmse_df, mae_df in zip(dataset_names, r2_values_bs, rmse_values_bs, mae_values_bs):\n",
    "#         # Use the dataset_name to create unique sheet names\n",
    "#         r2_df.to_excel(writer, sheet_name=f'{dataset_name}_R2')\n",
    "#         rmse_df.to_excel(writer, sheet_name=f'{dataset_name}_RMSE')\n",
    "#         mae_df.to_excel(writer, sheet_name=f'{dataset_name}_MAE')\n"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2023-10-18T11:42:56.869346494Z",
     "start_time": "2023-10-18T11:42:56.089333065Z"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 224,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "7\n",
      "ANOVA table:\n",
      "              sum_sq      df            F  PR(>F)\n",
      "C(group)  70.672951     6.0  6597.346874     0.0\n",
      "Residual   6.236361  3493.0          NaN     NaN\n",
      "Tukey HSD results:\n",
      "             Multiple Comparison of Means - Tukey HSD, FWER=0.05             \n",
      "============================================================================\n",
      "      group1             group2       meandiff p-adj   lower   upper  reject\n",
      "----------------------------------------------------------------------------\n",
      "       MDPocket_R2     MD_MDpocket_R2   0.0827    0.0  0.0748  0.0905   True\n",
      "       MDPocket_R2              MD_R2   0.0702    0.0  0.0623  0.0781   True\n",
      "       MDPocket_R2 Seq_MD_MDpocket_R2   0.3237    0.0  0.3158  0.3316   True\n",
      "       MDPocket_R2          Seq_MD_R2   0.3511    0.0  0.3432   0.359   True\n",
      "       MDPocket_R2    Seq_MDpocket_R2   0.3078    0.0     0.3  0.3157   True\n",
      "       MDPocket_R2             Seq_R2   0.3483    0.0  0.3404  0.3561   True\n",
      "    MD_MDpocket_R2              MD_R2  -0.0124 0.0001 -0.0203 -0.0046   True\n",
      "    MD_MDpocket_R2 Seq_MD_MDpocket_R2    0.241    0.0  0.2331  0.2489   True\n",
      "    MD_MDpocket_R2          Seq_MD_R2   0.2684    0.0  0.2605  0.2763   True\n",
      "    MD_MDpocket_R2    Seq_MDpocket_R2   0.2252    0.0  0.2173  0.2331   True\n",
      "    MD_MDpocket_R2             Seq_R2   0.2656    0.0  0.2577  0.2735   True\n",
      "             MD_R2 Seq_MD_MDpocket_R2   0.2535    0.0  0.2456  0.2614   True\n",
      "             MD_R2          Seq_MD_R2   0.2809    0.0   0.273  0.2888   True\n",
      "             MD_R2    Seq_MDpocket_R2   0.2376    0.0  0.2297  0.2455   True\n",
      "             MD_R2             Seq_R2   0.2781    0.0  0.2702  0.2859   True\n",
      "Seq_MD_MDpocket_R2          Seq_MD_R2   0.0274    0.0  0.0195  0.0353   True\n",
      "Seq_MD_MDpocket_R2    Seq_MDpocket_R2  -0.0158    0.0 -0.0237  -0.008   True\n",
      "Seq_MD_MDpocket_R2             Seq_R2   0.0246    0.0  0.0167  0.0325   True\n",
      "         Seq_MD_R2    Seq_MDpocket_R2  -0.0432    0.0 -0.0511 -0.0354   True\n",
      "         Seq_MD_R2             Seq_R2  -0.0028  0.941 -0.0107  0.0051  False\n",
      "   Seq_MDpocket_R2             Seq_R2   0.0404    0.0  0.0325  0.0483   True\n",
      "----------------------------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import statsmodels.api as sm\n",
    "from statsmodels.formula.api import ols\n",
    "from statsmodels.stats.multicomp import pairwise_tukeyhsd\n",
    "\n",
    "# Load the Excel file\n",
    "with pd.ExcelFile('/home/nick/PycharmProjects/Engineering_EKB/Python_scripts/metrics.xlsx') as xls:\n",
    "    # Get the sheet names for R2\n",
    "    r2_sheet_names = [sheet for sheet in xls.sheet_names if 'R2' in sheet]\n",
    "    print(len(r2_sheet_names))\n",
    "\n",
    "    # Create an empty DataFrame to hold the combined data\n",
    "    all_data = pd.DataFrame()\n",
    "\n",
    "    for sheet_name in r2_sheet_names:\n",
    "        # Read the data from the current sheet\n",
    "        data = pd.read_excel(xls, sheet_name=sheet_name)\n",
    "\n",
    "\n",
    "        # Add a 'group' column to indicate the sheet name\n",
    "        data['group'] = sheet_name\n",
    "\n",
    "        # Append this data to the combined DataFrame\n",
    "        all_data = pd.concat([all_data, data], ignore_index=True)\n",
    "\n",
    "    # Now you have all data combined with a 'group' column indicating the sheet name\n",
    "    formula = 'R2 ~ C(group)'\n",
    "    model = ols(formula, all_data).fit()\n",
    "    anova_table = sm.stats.anova_lm(model, typ=2)\n",
    "    print(f'ANOVA table:\\n', anova_table)\n",
    "\n",
    "    # Perform Tukey HSD test\n",
    "    tukey = pairwise_tukeyhsd(all_data['R2'], all_data['group'])\n",
    "    print(f'Tukey HSD results:\\n', tukey)\n"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2023-10-19T18:19:13.145411498Z",
     "start_time": "2023-10-19T18:19:12.726114896Z"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 227,
   "outputs": [
    {
     "data": {
      "text/plain": "<Figure size 1000x600 with 1 Axes>",
      "image/png": "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"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "labels = all_data['group'].unique()\n",
    "# Set the aesthetic style of the plots\n",
    "sns.set_style(\"whitegrid\")\n",
    "\n",
    "# Create a figure and axis object\n",
    "fig, ax = plt.subplots(figsize=(10,6))\n",
    "\n",
    "# Iterate through each group and plot the KDE\n",
    "for label in labels:\n",
    "    group_data = all_data[all_data['group'] == label]\n",
    "    sns.kdeplot(data=group_data, x='R2', ax=ax, label=label, fill=True)\n",
    "\n",
    "# Optionally, set the title and labels\n",
    "# ax.set_title('Distribution of R2 values across different groups')\n",
    "ax.set_xlabel(r'$R^2$', fontsize=16)\n",
    "ax.set_ylabel('Density', fontsize=16)\n",
    "ax.tick_params(axis='x', labelsize=14)  # Adjust the x-axis tick label size\n",
    "ax.tick_params(axis='y', labelsize=14)\n",
    "\n",
    "\n",
    "handles, labels = ax.get_legend_handles_labels()\n",
    "modified_labels = [label.rsplit('_', 1)[0].replace('_', '+') for label in labels]\n",
    "# Ensure the legend is displayed\n",
    "# plt.legend(handles, modified_labels, title='Group', title_fontsize='14', fontsize='12', loc='upper left')\n",
    "# plt.savefig('/home/nick/PycharmProjects/Engineering_EKB/Figures/R2_boostrap.png')\n",
    "# Display the plot\n",
    "plt.show()\n"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2023-10-19T18:20:01.558709121Z",
     "start_time": "2023-10-19T18:20:01.281890367Z"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 222,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ANOVA table:\n",
      "              sum_sq      df            F  PR(>F)\n",
      "C(group)  70.032013     6.0  6667.870931     0.0\n",
      "Residual   6.114441  3493.0          NaN     NaN\n",
      "Tukey HSD results:\n",
      "               Multiple Comparison of Means - Tukey HSD, FWER=0.05               \n",
      "================================================================================\n",
      "       group1               group2        meandiff p-adj   lower   upper  reject\n",
      "--------------------------------------------------------------------------------\n",
      "       MDPocket_RMSE     MD_MDpocket_RMSE  -0.0739    0.0 -0.0817 -0.0661   True\n",
      "       MDPocket_RMSE              MD_RMSE  -0.0623    0.0 -0.0701 -0.0545   True\n",
      "       MDPocket_RMSE Seq_MD_MDpocket_RMSE  -0.3163    0.0 -0.3241 -0.3085   True\n",
      "       MDPocket_RMSE          Seq_MD_RMSE  -0.3467    0.0 -0.3545 -0.3389   True\n",
      "       MDPocket_RMSE    Seq_MDpocket_RMSE  -0.2991    0.0 -0.3069 -0.2913   True\n",
      "       MDPocket_RMSE             Seq_RMSE  -0.3436    0.0 -0.3514 -0.3358   True\n",
      "    MD_MDpocket_RMSE              MD_RMSE   0.0116 0.0003  0.0038  0.0194   True\n",
      "    MD_MDpocket_RMSE Seq_MD_MDpocket_RMSE  -0.2424    0.0 -0.2502 -0.2346   True\n",
      "    MD_MDpocket_RMSE          Seq_MD_RMSE  -0.2728    0.0 -0.2806  -0.265   True\n",
      "    MD_MDpocket_RMSE    Seq_MDpocket_RMSE  -0.2252    0.0  -0.233 -0.2174   True\n",
      "    MD_MDpocket_RMSE             Seq_RMSE  -0.2697    0.0 -0.2775 -0.2619   True\n",
      "             MD_RMSE Seq_MD_MDpocket_RMSE  -0.2539    0.0 -0.2617 -0.2461   True\n",
      "             MD_RMSE          Seq_MD_RMSE  -0.2843    0.0 -0.2921 -0.2765   True\n",
      "             MD_RMSE    Seq_MDpocket_RMSE  -0.2367    0.0 -0.2445 -0.2289   True\n",
      "             MD_RMSE             Seq_RMSE  -0.2813    0.0 -0.2891 -0.2735   True\n",
      "Seq_MD_MDpocket_RMSE          Seq_MD_RMSE  -0.0304    0.0 -0.0382 -0.0226   True\n",
      "Seq_MD_MDpocket_RMSE    Seq_MDpocket_RMSE   0.0172    0.0  0.0094   0.025   True\n",
      "Seq_MD_MDpocket_RMSE             Seq_RMSE  -0.0273    0.0 -0.0351 -0.0195   True\n",
      "         Seq_MD_RMSE    Seq_MDpocket_RMSE   0.0476    0.0  0.0398  0.0554   True\n",
      "         Seq_MD_RMSE             Seq_RMSE   0.0031 0.9081 -0.0047  0.0109  False\n",
      "   Seq_MDpocket_RMSE             Seq_RMSE  -0.0445    0.0 -0.0523 -0.0367   True\n",
      "--------------------------------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import statsmodels.api as sm\n",
    "from statsmodels.formula.api import ols\n",
    "from statsmodels.stats.multicomp import pairwise_tukeyhsd\n",
    "\n",
    "# Load the Excel file\n",
    "with pd.ExcelFile('/home/nick/PycharmProjects/Engineering_EKB/Python_scripts/metrics.xlsx') as xls:\n",
    "    # Get the sheet names for R2\n",
    "    rmse_sheet_names = [sheet for sheet in xls.sheet_names if 'RMSE' in sheet]\n",
    "\n",
    "    # Create an empty DataFrame to hold the combined data\n",
    "    all_data = pd.DataFrame()\n",
    "\n",
    "    for sheet_name in rmse_sheet_names:\n",
    "        # Read the data from the current sheet\n",
    "        data = pd.read_excel(xls, sheet_name=sheet_name)\n",
    "\n",
    "\n",
    "        # Add a 'group' column to indicate the sheet name\n",
    "        data['group'] = sheet_name\n",
    "\n",
    "        # Append this data to the combined DataFrame\n",
    "        all_data = pd.concat([all_data, data], ignore_index=True)\n",
    "\n",
    "    # Now you have all data combined with a 'group' column indicating the sheet name\n",
    "    formula = 'RMSE ~ C(group)'\n",
    "    model = ols(formula, all_data).fit()\n",
    "    anova_table = sm.stats.anova_lm(model, typ=2)\n",
    "    print(f'ANOVA table:\\n', anova_table)\n",
    "\n",
    "    # Perform Tukey HSD test\n",
    "    tukey = pairwise_tukeyhsd(all_data['RMSE'], all_data['group'])\n",
    "    print(f'Tukey HSD results:\\n', tukey)\n"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2023-10-19T18:18:44.186127112Z",
     "start_time": "2023-10-19T18:18:43.755277875Z"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 223,
   "outputs": [
    {
     "data": {
      "text/plain": "<Figure size 1000x600 with 1 Axes>",
      "image/png": "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"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Assume all_data is your DataFrame\n",
    "labels = all_data['group'].unique()\n",
    "\n",
    "# Set the aesthetic style of the plots\n",
    "sns.set_style(\"whitegrid\")\n",
    "\n",
    "# Create a figure and axis object\n",
    "fig, ax = plt.subplots(figsize=(10,6))\n",
    "\n",
    "# Iterate through each group and plot the KDE\n",
    "for label in labels:\n",
    "    group_data = all_data[all_data['group'] == label]\n",
    "    sns.kdeplot(data=group_data, x='RMSE', ax=ax, label=label, fill=True)\n",
    "\n",
    "# Optionally, set the title and labels\n",
    "\n",
    "\n",
    "ax.tick_params(axis='x', labelsize=14)  # Adjust the x-axis tick label size\n",
    "ax.tick_params(axis='y', labelsize=14)\n",
    "ax.set_xlabel('RMSE', fontsize=16)\n",
    "ax.set_ylabel('Density', fontsize=16)\n",
    "\n",
    "handles, labels = ax.get_legend_handles_labels()\n",
    "modified_labels = [label.rsplit('_', 1)[0].replace('_', '+') for label in labels]\n",
    "# Ensure the legend is displayed\n",
    "# plt.legend(handles, modified_labels, title='Group', title_fontsize='14', fontsize='12')\n",
    "# plt.savefig('/home/nick/PycharmProjects/Engineering_EKB/Figures/RMSE_boostrap.png')\n",
    "# Display the plot\n",
    "plt.show()"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2023-10-19T18:18:46.186709859Z",
     "start_time": "2023-10-19T18:18:45.897075934Z"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 186,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ANOVA table:\n",
      "              sum_sq      df            F  PR(>F)\n",
      "C(group)  59.952970     6.0  6651.608718     0.0\n",
      "Residual   5.247245  3493.0          NaN     NaN\n",
      "Tukey HSD results:\n",
      "              Multiple Comparison of Means - Tukey HSD, FWER=0.05              \n",
      "==============================================================================\n",
      "       group1              group2       meandiff p-adj   lower   upper  reject\n",
      "------------------------------------------------------------------------------\n",
      "       MDPocket_MAE              MD_MAE  -0.1168    0.0  -0.124 -0.1095   True\n",
      "       MDPocket_MAE     MD_MDpocket_MAE  -0.1253    0.0 -0.1325  -0.118   True\n",
      "       MDPocket_MAE             Seq_MAE  -0.3418    0.0  -0.349 -0.3345   True\n",
      "       MDPocket_MAE          Seq_MD_MAE  -0.3398    0.0  -0.347 -0.3325   True\n",
      "       MDPocket_MAE Seq_MD_MDpocket_MAE  -0.3312    0.0 -0.3384 -0.3239   True\n",
      "       MDPocket_MAE    Seq_MDpocket_MAE  -0.3227    0.0 -0.3299 -0.3155   True\n",
      "             MD_MAE     MD_MDpocket_MAE  -0.0085 0.0093 -0.0157 -0.0013   True\n",
      "             MD_MAE             Seq_MAE   -0.225    0.0 -0.2322 -0.2178   True\n",
      "             MD_MAE          Seq_MD_MAE   -0.223    0.0 -0.2302 -0.2158   True\n",
      "             MD_MAE Seq_MD_MDpocket_MAE  -0.2144    0.0 -0.2217 -0.2072   True\n",
      "             MD_MAE    Seq_MDpocket_MAE  -0.2059    0.0 -0.2132 -0.1987   True\n",
      "    MD_MDpocket_MAE             Seq_MAE  -0.2165    0.0 -0.2237 -0.2092   True\n",
      "    MD_MDpocket_MAE          Seq_MD_MAE  -0.2145    0.0 -0.2217 -0.2073   True\n",
      "    MD_MDpocket_MAE Seq_MD_MDpocket_MAE  -0.2059    0.0 -0.2131 -0.1987   True\n",
      "    MD_MDpocket_MAE    Seq_MDpocket_MAE  -0.1974    0.0 -0.2046 -0.1902   True\n",
      "            Seq_MAE          Seq_MD_MAE    0.002 0.9844 -0.0053  0.0092  False\n",
      "            Seq_MAE Seq_MD_MDpocket_MAE   0.0106 0.0003  0.0033  0.0178   True\n",
      "            Seq_MAE    Seq_MDpocket_MAE   0.0191    0.0  0.0118  0.0263   True\n",
      "         Seq_MD_MAE Seq_MD_MDpocket_MAE   0.0086 0.0083  0.0014  0.0158   True\n",
      "         Seq_MD_MAE    Seq_MDpocket_MAE   0.0171    0.0  0.0099  0.0243   True\n",
      "Seq_MD_MDpocket_MAE    Seq_MDpocket_MAE   0.0085 0.0096  0.0013  0.0157   True\n",
      "------------------------------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import statsmodels.api as sm\n",
    "from statsmodels.formula.api import ols\n",
    "from statsmodels.stats.multicomp import pairwise_tukeyhsd\n",
    "\n",
    "# Load the Excel file\n",
    "with pd.ExcelFile('/home/nick/PycharmProjects/Engineering_EKB/Python_scripts/metrics.xlsx') as xls:\n",
    "    # Get the sheet names for R2\n",
    "    mae_sheet_names = [sheet for sheet in xls.sheet_names if 'MAE' in sheet]\n",
    "\n",
    "    # Create an empty DataFrame to hold the combined data\n",
    "    all_data = pd.DataFrame()\n",
    "\n",
    "    for sheet_name in mae_sheet_names:\n",
    "        # Read the data from the current sheet\n",
    "        data = pd.read_excel(xls, sheet_name=sheet_name)\n",
    "\n",
    "\n",
    "        # Add a 'group' column to indicate the sheet name\n",
    "        data['group'] = sheet_name\n",
    "\n",
    "        # Append this data to the combined DataFrame\n",
    "        all_data = pd.concat([all_data, data], ignore_index=True)\n",
    "\n",
    "    # Now you have all data combined with a 'group' column indicating the sheet name\n",
    "    formula = 'MAE ~ C(group)'\n",
    "    model = ols(formula, all_data).fit()\n",
    "    anova_table = sm.stats.anova_lm(model, typ=2)\n",
    "    print(f'ANOVA table:\\n', anova_table)\n",
    "\n",
    "    # Perform Tukey HSD test\n",
    "    tukey = pairwise_tukeyhsd(all_data['MAE'], all_data['group'])\n",
    "    print(f'Tukey HSD results:\\n', tukey)\n"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2023-10-19T09:38:13.244457838Z",
     "start_time": "2023-10-19T09:38:12.739926403Z"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 221,
   "outputs": [
    {
     "data": {
      "text/plain": "<Figure size 1000x600 with 1 Axes>",
      "image/png": "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"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Assume all_data is your DataFrame\n",
    "labels = all_data['group'].unique()\n",
    "\n",
    "# Set the aesthetic style of the plots\n",
    "sns.set_style(\"whitegrid\")\n",
    "\n",
    "# Create a figure and axis object\n",
    "fig, ax = plt.subplots(figsize=(10,6))\n",
    "\n",
    "# ... rest of the plotting code ...\n",
    "\n",
    "# Iterate through each group and plot the KDE\n",
    "for label in labels:\n",
    "    group_data = all_data[all_data['group'] == label]\n",
    "    sns.kdeplot(data=group_data, x='MAE', ax=ax, label=label, fill=True)\n",
    "\n",
    "# Optionally, set the title and labels\n",
    "# ax.set_title('Distribution of MAE values across different groups')\n",
    "ax.tick_params(axis='x', labelsize=14)  # Adjust the x-axis tick label size\n",
    "ax.tick_params(axis='y', labelsize=14)\n",
    "ax.set_xlabel('MAE', fontsize=16)\n",
    "ax.set_ylabel('Density', fontsize=16)\n",
    "\n",
    "handles, labels = ax.get_legend_handles_labels()\n",
    "modified_labels = [label.rsplit('_', 1)[0].replace('_', '+') for label in labels]\n",
    "\n",
    "\n",
    "# Ensure the legend is displayed\n",
    "legend = plt.legend(handles, modified_labels, title='Datasets', title_fontsize='14', fontsize='20', loc='center left', bbox_to_anchor=(1, 0.5))\n",
    "plt.legend(handles, modified_labels, title='Datasets', title_fontsize='14', fontsize='12', loc='center left', bbox_to_anchor=(1, 0.5))\n",
    "plt.savefig('/home/nick/PycharmProjects/Engineering_EKB/Figures/MAE_boostrap.png', bbox_extra_artists=(legend,), bbox_inches='tight')\n",
    "# Display the plot\n",
    "plt.show()"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2023-10-19T18:17:54.974868138Z",
     "start_time": "2023-10-19T18:17:54.554264663Z"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 216,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1000, 600)\n",
      "(1079, 525)\n"
     ]
    }
   ],
   "source": [
    "from PIL import Image\n",
    "\n",
    "# Load your images\n",
    "img1 = Image.open('/home/nick/PycharmProjects/Engineering_EKB/Figures/R2_boostrap.png')\n",
    "img2 = Image.open('/home/nick/PycharmProjects/Engineering_EKB/Figures/RMSE_boostrap.png')\n",
    "img3 = Image.open('/home/nick/PycharmProjects/Engineering_EKB/Figures/MAE_boostrap.png')\n",
    "print(img2.size)\n",
    "print(img3.size)\n",
    "# Define cropping boxes (left, upper, right, lower)\n",
    "crop_box1 = (0, 0, img1.width- 100, img1.height )  # Adjust the values as necessary\n",
    "crop_box2 = (0, 0, img2.width-100, img2.height)  # Adjust the values as necessary\n",
    "crop_box3 = (0, 0, img3.width + 100, img3.height+75)       # Adjust the values as necessary\n",
    "\n",
    "# Crop the images\n",
    "img1 = img1.crop(crop_box1)\n",
    "img2 = img2.crop(crop_box2)\n",
    "img3 = img3.crop(crop_box3)\n",
    "\n",
    "# Get dimensions of the individual images\n",
    "width1, height1 = img1.size\n",
    "width2, height2 = img2.size\n",
    "width3, height3 = img3.size\n",
    "\n",
    "# Create a new image with a width that is the sum of the widths of the individual images\n",
    "# and a height that is the maximum height of the individual images\n",
    "total_width = width1 + width2 + width3\n",
    "max_height = max(height1, height2, height3)\n",
    "\n",
    "composite_img = Image.new('RGBA', (total_width, max_height))\n",
    "\n",
    "# Paste each image into the composite image\n",
    "composite_img.paste(img1, (0, 0))\n",
    "composite_img.paste(img2, (width1, 0))\n",
    "composite_img.paste(img3, (width1 + width2, 0))\n",
    "\n",
    "# Save the composite image\n",
    "composite_img.save('/home/nick/PycharmProjects/Engineering_EKB/Figures/combined_metrics_distributions.png')\n",
    "\n",
    "# Optionally, display the composite image\n",
    "composite_img.show()\n"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2023-10-19T09:55:37.785678723Z",
     "start_time": "2023-10-19T09:55:37.639280931Z"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "outputs": [],
   "source": [
    "performance = []\n",
    "split = []\n",
    "gain = []\n",
    "for i in range(len(datasets)):\n",
    "    param_string = results_table['Best Estimator'][i]\n",
    "    params = dict(re.findall(r\"(\\w+)=(.*?)(?:(?:,|$|\\n\\s+)|\\))\", param_string))\n",
    "    reg_lambda = params.get('reg_lambda', '')\n",
    "    params['reg_lambda'] = reg_lambda.strip(\")\")\n",
    "\n",
    "    # Convert values to their appropriate data types if needed\n",
    "    params = {k: eval(v) if v.isdigit() else v for k, v in params.items()}\n",
    "\n",
    "    # Create the LGBMRegressor object with the extracted parameters\n",
    "    Best_model = LGBMRegressor(**params)\n",
    "    X_train, X_test, y_train, y_test = train_test_split(datasets[i], Y, test_size=0.2, random_state=random_seed)\n",
    "    eval_set = [(X_train, y_train), (X_test, y_test)]\n",
    "\n",
    "    Best_model.fit(X_train, y_train, eval_metric=[\"mae\", \"rmse\"], eval_set=eval_set, verbose=True, early_stopping_rounds=10)\n",
    "    y_pred = Best_model.predict(X_test)\n",
    "    predictions = [value for value in y_pred]\n",
    "    results = Best_model.evals_result_\n",
    "    perm = PermutationImportance(Best_model, random_state=random_seed).fit(X_test, y_test)\n",
    "\n",
    "    weights_perm = eli5.show_weights(perm, feature_names=X_test.columns.tolist())\n",
    "    performance.append(weights_perm)\n",
    "    weights_split = eli5.show_weights(Best_model, importance_type=\"split\")\n",
    "    split.append(weights_split)\n",
    "    weights_gain = eli5.show_weights(Best_model, importance_type=\"gain\")\n",
    "    gain.append(weights_gain)\n"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create an empty DataFrame\n",
    "performances = pd.DataFrame()\n",
    "\n",
    "for i, html in enumerate(performance):\n",
    "    # Parse the HTML\n",
    "    soup = BeautifulSoup(html.data, 'html.parser')\n",
    "\n",
    "    # Find the table element in the HTML\n",
    "    table = soup.find('table')\n",
    "\n",
    "    # Convert the table to a DataFrame\n",
    "    temp_df = pd.read_html(str(table))[0]\n",
    "\n",
    "    # Rename the weight and feature columns\n",
    "    temp_df.rename(columns={'weight': f'Performance {i+1} Weight', 'feature': f'Performance {i+1} Feature'}, inplace=True)\n",
    "\n",
    "    # Append the temporary DataFrame to the main DataFrame\n",
    "    performances = pd.concat([performances, temp_df], axis=1)\n",
    "\n",
    "# Display the combined DataFrame\n",
    "performances\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Convert 'Weight' column to numeric values and extract the first digit\n",
    "Seq_MD_perf = performances.iloc[:15, 2:4]\n",
    "# Convert 'Weight' column to numeric values and extract the first digit\n",
    "\n",
    "Seq_MD_perf['Weight'] = pd.to_numeric(Seq_MD_perf['Weight'].str.extract(r'(\\d+\\.\\d+|\\d+)', expand=False), errors='coerce')\n",
    "Seq_MD_perf_sorted = Seq_MD_perf.sort_values(by='Weight', ascending=True)\n",
    "\n",
    "# Create a horizontal bar plot\n",
    "plt.figure(figsize=(10, 6))\n",
    "plt.barh(Seq_MD_perf_sorted['Feature'], Seq_MD_perf_sorted['Weight'], color='blue')\n",
    "plt.title('Feature Weights (Top 10)')\n",
    "plt.xlabel('Weight')\n",
    "plt.ylabel('Feature')\n",
    "plt.xlim(0, Seq_MD_perf_sorted['Weight'].max()+0.01)  # Set the x-axis range\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Convert 'Weight' column to numeric values and extract the first digit and error value\n",
    "Seq_MD_perf = performances.iloc[:15, 2:4]\n",
    "Seq_MD_perf[['Weight', 'Error']] = Seq_MD_perf['Weight'].str.extract(r'(\\d+\\.\\d+)\\s*±\\s*(\\d+\\.\\d+)', expand=True)\n",
    "\n",
    "# Convert the extracted columns to numeric values\n",
    "Seq_MD_perf['Weight'] = pd.to_numeric(Seq_MD_perf['Weight'], errors='coerce')\n",
    "Seq_MD_perf['Error'] = pd.to_numeric(Seq_MD_perf['Error'], errors='coerce')\n",
    "\n",
    "Seq_MD_perf_sorted = Seq_MD_perf.sort_values(by='Weight', ascending=True)\n",
    "\n",
    "# Create a horizontal bar plot with error bars\n",
    "plt.figure(figsize=(10, 6))\n",
    "plt.barh(Seq_MD_perf_sorted['Feature'], Seq_MD_perf_sorted['Weight'], color='blue')\n",
    "plt.errorbar(Seq_MD_perf_sorted['Weight'], Seq_MD_perf_sorted['Feature'], xerr=Seq_MD_perf_sorted['Error'], fmt='none', ecolor='black', capsize=4)\n",
    "plt.xlabel('Weight by Permutation Feature Importance')\n",
    "plt.ylabel('Feature')\n",
    "plt.xlim(0, Seq_MD_perf_sorted['Weight'].max() + Seq_MD_perf_sorted['Error'].max() + 0.01)  # Set the x-axis range\n",
    "plt.tight_layout()\n",
    "\n",
    "# Save the figure\n",
    "#plt.savefig('PIF_ranking.png', dpi=1200)\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create an empty DataFrame\n",
    "splits = pd.DataFrame()\n",
    "\n",
    "for i, html in enumerate(split):\n",
    "    # Parse the HTML\n",
    "    soup = BeautifulSoup(html.data, 'html.parser')\n",
    "\n",
    "    # Find the table element in the HTML\n",
    "    table = soup.find('table')\n",
    "\n",
    "    # Convert the table to a DataFrame\n",
    "    temp_df = pd.read_html(str(table))[0]\n",
    "\n",
    "    # Rename the weight and feature columns\n",
    "    temp_df.rename(columns={'weight': f'Performance {i+1} Weight', 'feature': f'Performance {i+1} Feature'}, inplace=True)\n",
    "\n",
    "    # Append the temporary DataFrame to the main DataFrame\n",
    "    splits = pd.concat([splits, temp_df], axis=1)\n",
    "\n",
    "# Display the combined DataFrame\n",
    "splits\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Convert 'Weight' column to numeric values and extract the first digit\n",
    "Seq_MD_splits = splits.iloc[:15, 2:4]\n",
    "# Convert 'Weight' column to numeric values and extract the first digit\n",
    "\n",
    "Seq_MD_splits['Weight'] = pd.to_numeric(Seq_MD_splits['Weight'].str.extract(r'(\\d+\\.\\d+|\\d+)', expand=False), errors='coerce')\n",
    "Seq_MD_splits_sorted = Seq_MD_splits.sort_values(by='Weight', ascending=True)\n",
    "\n",
    "# Create a horizontal bar plot\n",
    "plt.figure(figsize=(10, 6))\n",
    "plt.barh(Seq_MD_splits_sorted['Feature'], Seq_MD_splits_sorted['Weight'], color='blue')\n",
    "plt.xlabel('Weight by Split Importance')\n",
    "plt.ylabel('Feature')\n",
    "plt.xlim(0, Seq_MD_splits_sorted['Weight'].max()+0.01)  # Set the x-axis range\n",
    "plt.tight_layout()\n",
    "# plt.savefig('PIF_split.png', dpi=1200)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create an empty DataFrame\n",
    "gains = pd.DataFrame()\n",
    "\n",
    "for i, html in enumerate(gain):\n",
    "    # Parse the HTML\n",
    "    soup = BeautifulSoup(html.data, 'html.parser')\n",
    "\n",
    "    # Find the table element in the HTML\n",
    "    table = soup.find('table')\n",
    "\n",
    "    # Convert the table to a DataFrame\n",
    "    temp_df = pd.read_html(str(table))[0]\n",
    "\n",
    "    # Rename the weight and feature columns\n",
    "    temp_df.rename(columns={'weight': f'Performance {i+1} Weight', 'feature': f'Performance {i+1} Feature'}, inplace=True)\n",
    "\n",
    "    # Append the temporary DataFrame to the main DataFrame\n",
    "    gains = pd.concat([gains, temp_df], axis=1)\n",
    "\n",
    "# Display the combined DataFrame\n",
    "gains"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Convert 'Weight' column to numeric values and extract the first digit\n",
    "Seq_MD_gains = gains.iloc[:15, 2:4]\n",
    "# Convert 'Weight' column to numeric values and extract the first digit\n",
    "\n",
    "Seq_MD_gains['Weight'] = pd.to_numeric(Seq_MD_gains['Weight'].str.extract(r'(\\d+\\.\\d+|\\d+)', expand=False), errors='coerce')\n",
    "Seq_MD_gains_sorted = Seq_MD_gains.sort_values(by='Weight', ascending=True)\n",
    "\n",
    "# Create a horizontal bar plot\n",
    "plt.figure(figsize=(10, 6))\n",
    "plt.barh(Seq_MD_gains_sorted['Feature'], Seq_MD_gains_sorted['Weight'], color='blue')\n",
    "\n",
    "plt.xlabel('Weight by Performance Gain')\n",
    "plt.ylabel('Feature')\n",
    "plt.xlim(0, Seq_MD_gains_sorted['Weight'].max()+0.01)  # Set the x-axis range\n",
    "plt.tight_layout()\n",
    "#  plt.savefig('PIF_gain.png', dpi=1200)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import shap"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "BestSeqMD = LGBMRegressor(colsample_bytree=0.8, learning_rate=0.008, max_depth=41,\n",
    "                          min_split_gain=1, n_estimators=520, num_leaves=42,random_state=3031, reg_alpha=0.6, reg_lambda=7)\n",
    "X_train, X_test, y_train, y_test = train_test_split(Seq_MD, Y, test_size=0.2, random_state=random_seed)\n",
    "eval_set = [(X_train, y_train), (X_test, y_test)]\n",
    "BestSeqMD.fit(X_train, y_train, eval_metric=[\"mae\", \"rmse\"], eval_set=eval_set, verbose=True, early_stopping_rounds=10)\n",
    "y_pred = BestSeqMD.predict(X_test)\n",
    "predictions = [value for value in y_pred]\n",
    "results = BestSeqMD.evals_result_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "shap_values = shap.Explainer(BestSeqMD).shap_values(X_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "shap.summary_plot(shap_values, X_train, plot_type=\"bar\", show=False)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "f = plt.figure()\n",
    "shap.summary_plot(shap_values, X_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "shap.dependence_plot('halpha', shap_values, X_train, dot_size=32, show=False)\n",
    "# Set the DPI (dots per inch) to 1200\n",
    "# plt.savefig('/home/nick/PycharmProjects/Engineering_EKB/Figures/dependence_plot_halpha_SEQMD.png', dpi=1200)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# fig, ax = plt.subplots()\n",
    "# heatmap = shap.plots.heatmap(shap_values, show=False, cmap='viridis')\n",
    "# ax.set_aspect('auto')\n",
    "#\n",
    "# # Add colorbar on the side and move it to the right\n",
    "# cbar = plt.colorbar(heatmap, ax=ax, fraction=0.03, pad=0.08)\n",
    "# cbar.set_label('SHAP Value')\n",
    "#\n",
    "# plt.savefig('/home/nick/PycharmProjects/Engineering_EKB/Figures/heatmap_SEQMD.png', dpi=1200)\n",
    "# plt.tight_layout()\n",
    "# # Display the plot\n",
    "# plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "shap_values = shap.Explainer(BestSeqMD).shap_values(X_train)\n",
    "f = plt.figure()\n",
    "shap.summary_plot(shap_values, X_train, cmap='plasma', show=False)\n",
    "# plt.savefig('/home/nick/PycharmProjects/Engineering_EKB/Figures/summary_plot_SEQMD.png', dpi=1200)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import statsmodels.api as sm\n",
    "import scipy.stats as stats"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "diff = y_test - y_pred\n",
    "min_value = np.min(diff)  # find minimum value in diff\n",
    "\n",
    "if min_value < 0:  # check if minimum value is negative\n",
    "    diff = diff - (min_value +0.000001)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "plt.hist(diff, bins='auto', density=True)\n",
    "plt.xlabel('Differences')\n",
    "plt.ylabel('Frequency')\n",
    "plt.title('Histogram of Differences')\n",
    "plt.show()\n",
    "\n",
    "# Q-Q plot\n",
    "# Q-Q plot using scipy.stats\n",
    "stats.probplot(diff, dist=\"norm\", plot=plt)\n",
    "plt.xlabel('Theoretical Quantiles')\n",
    "plt.ylabel('Sample Quantiles')\n",
    "plt.title('Q-Q Plot of Differences')\n",
    "plt.show()\n",
    "\n",
    "# Shapiro-Wilk test\n",
    "_, p_value = stats.shapiro(diff)\n",
    "alpha = 0.05  # significance level\n",
    "\n",
    "print(f\"Shapiro-Wilk test p-value: {p_value}\")\n",
    "if p_value > alpha:\n",
    "    print(\"Differences likely follow a normal distribution\")\n",
    "else:\n",
    "    print(\"Differences do not follow a normal distribution\")\n",
    "# plot the diff box plot\n",
    "plt.boxplot(diff)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# # Normalise the differences to a standard normal distribution\n",
    "# diff_norm = (diff - np.mean(diff)) / np.std(diff)\n",
    "# plt.hist(diff_norm, bins='auto', density=True)\n",
    "# plt.xlabel('Differences')\n",
    "# plt.ylabel('Frequency')\n",
    "# plt.title('Histogram of Normalised Differences')\n",
    "# plt.show()\n",
    "#\n",
    "# # Q-Q plot using scipy.stats\n",
    "# stats.probplot(diff_norm, dist=\"norm\", plot=plt)\n",
    "# plt.xlabel('Theoretical Quantiles')\n",
    "# plt.ylabel('Sample Quantiles')\n",
    "# plt.title('Q-Q Plot of Differences')\n",
    "# plt.show()\n",
    "# # Shapiro-Wilk test\n",
    "# _, p_value = stats.shapiro(diff_norm)\n",
    "# alpha = 0.05  # significance level\n",
    "#\n",
    "# print(f\"Shapiro-Wilk test p-value: {p_value}\")\n",
    "# if p_value > alpha:\n",
    "#     print(\"Differences likely follow a normal distribution\")\n",
    "# else:\n",
    "#     print(\"Differences do not follow a normal distribution\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# plt.hist(y_test, bins='auto', density=True, alpha=0.5, label='y_test')\n",
    "# plt.hist(diff_norm, bins='auto', density=True, alpha=0.5, label='Normalized Differences')\n",
    "# plt.xlabel('Values')\n",
    "# plt.ylabel('Frequency')\n",
    "# plt.title('Histogram Comparison')\n",
    "# plt.legend()\n",
    "# plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "diff_transformed = np.log(diff)\n",
    "\n",
    "# #Reshape diff_transformed to match the size of diff\n",
    "# diff_transformed = diff_transformed.reshape(-1, 1)\n",
    "plt.hist(diff_transformed, bins='auto', density=True)\n",
    "plt.xlabel('Differences')\n",
    "plt.ylabel('Frequency')\n",
    "plt.title('Histogram of Normalised Differences')\n",
    "plt.show()\n",
    "\n",
    "# Q-Q plot using scipy.stats\n",
    "stats.probplot(diff_transformed, dist=\"norm\", plot=plt)\n",
    "plt.xlabel('Theoretical Quantiles')\n",
    "plt.ylabel('Sample Quantiles')\n",
    "plt.title('Q-Q Plot of Differences')\n",
    "plt.show()\n",
    "# Shapiro-Wilk test\n",
    "_, p_value = stats.shapiro(diff_transformed)\n",
    "alpha = 0.05  # significance level\n",
    "\n",
    "print(f\"Shapiro-Wilk test p-value: {p_value}\")\n",
    "if p_value > alpha:\n",
    "    print(\"Differences likely follow a normal distribution\")\n",
    "else:\n",
    "    print(\"Differences do not follow a normal distribution\")\n",
    "# plot its cdf\n",
    "# Sort the normalized differences in ascending order\n",
    "sorted_diff = np.sort(diff_transformed)\n",
    "cdf = np.cumsum(sorted_diff)\n",
    "plt.plot(sorted_diff, cdf)\n",
    "plt.xlabel('Normalized Differences')\n",
    "plt.ylabel('Cumulative Probability')\n",
    "plt.title('CDF of Normalized Differences')\n",
    "plt.grid(True)\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# now try do log 2 transform"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# # Sort the normalized differences in ascending order\n",
    "# sorted_diff = np.sort(diff_transformed)\n",
    "#\n",
    "# # Calculate the cumulative probabilities\n",
    "# cumulative_prob = np.arange(1, len(sorted_diff) + 1) / len(sorted_diff)\n",
    "#\n",
    "# # Plot the CDF\n",
    "# plt.plot(sorted_diff, cumulative_prob, marker='o')\n",
    "# plt.xlabel('Normalized Differences')\n",
    "# plt.ylabel('Cumulative Probability')\n",
    "# plt.title('CDF of Normalized Differences')\n",
    "# plt.grid(True)\n",
    "# plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# from statsmodels.distributions.empirical_distribution import ECDF\n",
    "# ecdf_y_test = ECDF(y_test)\n",
    "# ecdf_y_pred = ECDF(y_pred)\n",
    "#\n",
    "# # Plot the CDFs\n",
    "# plt.step(ecdf_y_test.x, ecdf_y_test.y, label='y_test')\n",
    "# plt.step(ecdf_y_pred.x, ecdf_y_pred.y, label='y_pred')\n",
    "# plt.xlabel('Value')\n",
    "# plt.ylabel('Cumulative Probability')\n",
    "# plt.title('CDFs of y_test and y_pred')\n",
    "# plt.legend()\n",
    "# plt.grid(True)\n",
    "# plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# # Apply log transformation with an addition of 0.001\n",
    "# y_test_log = np.log(y_test + 0.001)\n",
    "# y_pred_log = np.log(y_pred + 0.001)\n",
    "#\n",
    "# # Calculate the empirical cumulative distribution functions (ECDFs)\n",
    "# ecdf_y_test = ECDF(y_test_log)\n",
    "# ecdf_y_pred = ECDF(y_pred_log)\n",
    "#\n",
    "# # Plot the CDFs\n",
    "# plt.step(ecdf_y_test.x, ecdf_y_test.y, label='y_test')\n",
    "# plt.step(ecdf_y_pred.x, ecdf_y_pred.y, label='y_pred')\n",
    "# plt.xlabel('Value')\n",
    "# plt.ylabel('Cumulative Probability')\n",
    "# plt.title('CDFs of Log-transformed y_test and y_pred')\n",
    "# plt.legend()\n",
    "# plt.grid(True)\n",
    "# plt.show()\n",
    "# # IT HAS NEGATIVE NUMBERS!!!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "diff_res = pd.DataFrame({'y_test': y_test, 'y_pred': y_pred, 'diff': y_test - y_pred})\n",
    "diff_res"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# plot an histogram of all the colums in diff_res on top of each other using seaborn\n",
    "sns.histplot(data=diff_res, x='y_test', bins='auto', color='lightblue', label='y_test')\n",
    "sns.histplot(data=diff_res, x='y_pred', bins='auto', color='red', label='y_pred')\n",
    "sns.histplot(data=diff_res, x='diff', bins='auto', color='lightgreen', label='diff')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# save diff_res to excel\n",
    "#  diff_res.to_excel('diff_res.xlsx')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Min-max normalization\n",
    "def min_max_norm(array):\n",
    "    return (array - np.min(array)) / (np.max(array) - np.min(array))\n",
    "diff_res['y_pred_norm'] = min_max_norm(diff_res['y_pred'])\n",
    "diff_res['y_test_norm'] = min_max_norm(diff_res['y_test'])\n",
    "diff_res['diff_norm'] = min_max_norm(diff_res['diff'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# plot an histogram of all the colums in diff_res on top of each other using seaborn\n",
    "sns.histplot(data=diff_res, x='y_test_norm', bins='auto', color='lightblue', label='y_test_norm')\n",
    "sns.histplot(data=diff_res, x='y_pred_norm', bins='auto', color='red', label='y_pred_norm')\n",
    "sns.histplot(data=diff_res, x='diff_norm', bins='auto', color='lightgreen', label='diff_norm')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# do normality test for the normalised columns\n",
    "# Shapiro-Wilk test\n",
    "_, p_value = stats.shapiro(diff_res['diff_norm'])\n",
    "alpha = 0.05  # significance level\n",
    "print(f\"Shapiro-Wilk test p-value: {p_value}\")\n",
    "if p_value > alpha:\n",
    "    print(\"y_test_norm likely follow a normal distribution\")\n",
    "else:\n",
    "    print(\"y_test_norm do not follow a normal distribution\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# print cdfs of all the normalised columns\n",
    "# Calculate the empirical cumulative distribution functions (ECDFs)\n",
    "from statsmodels.distributions.empirical_distribution import ECDF\n",
    "ecdf_y_test = ECDF(diff_res['y_test_norm'])\n",
    "ecdf_y_pred = ECDF(diff_res['y_pred_norm'])\n",
    "ecdf_diff = ECDF(diff_res['diff_norm'])\n",
    "ecdf_yp = ECDF(diff_res['y_pred'])\n",
    "ecdf_yt = ECDF(diff_res['y_test'])\n",
    "ecdf_d = ECDF(diff_res['diff'])\n",
    "\n",
    "# Plot the CDFs\n",
    "\n",
    "plt.step(ecdf_y_test.x, ecdf_y_test.y, label='y_test_norm')\n",
    "plt.step(ecdf_y_pred.x, ecdf_y_pred.y, label='y_pred_norm')\n",
    "plt.step(ecdf_diff.x, ecdf_diff.y, label='diff_norm')\n",
    "plt.xlabel('Value')\n",
    "plt.ylabel('Cumulative Probability')\n",
    "plt.title('CDFs of Normalised Min/Max y_test, y_pred and diff')\n",
    "plt.legend()\n",
    "plt.grid(True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.step(ecdf_yt.x, ecdf_yt.y, label='y_test')\n",
    "plt.step(ecdf_yp.x, ecdf_yp.y, label='y_pred')\n",
    "plt.step(ecdf_d.x, ecdf_d.y, label='diff')\n",
    "plt.xlabel('Value')\n",
    "plt.ylabel('Cumulative Probability')\n",
    "plt.title('CDFs of y_test, y_pred and diff')\n",
    "plt.legend()\n",
    "plt.grid(True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# plot hist of diff\n",
    "sns.histplot(data=diff_res, x='diff', bins='auto', color='lightgreen', label='diff')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "diff = y_test - y_pred\n",
    "\n",
    "sorted_differences = np.sort(diff)\n",
    "\n",
    "# Step 3: Determine the threshold for the upper tail\n",
    "desired_percentile_upper = 95  # Define the desired percentile for the upper tail\n",
    "index_upper = int(np.ceil(desired_percentile_upper / 100 * len(sorted_differences)))\n",
    "threshold_upper = sorted_differences[index_upper - 1]  # Subtract 1 because Python uses zero-based indexing\n",
    "\n",
    "# Step 4: Determine the threshold for the lower tail\n",
    "desired_percentile_lower = 5  # Define the desired percentile for the lower tail\n",
    "index_lower = int(np.ceil(desired_percentile_lower / 100 * len(sorted_differences)))\n",
    "threshold_lower = sorted_differences[index_lower - 1]  # Subtract 1 because Python uses zero-based indexing\n",
    "\n",
    "# Step 5: Determine the range around the median (close to zero)\n",
    "\n",
    "# Determine the range around the median (close to zero)\n",
    "desired_range_around_median = 45  # Define the desired range around the median in percentage\n",
    "median = np.median(diff)\n",
    "threshold_lower_range = median - (median * desired_range_around_median / 100)\n",
    "threshold_upper_range = median + (median * desired_range_around_median / 100)\n",
    "\n",
    "\n",
    "# Print the thresholds\n",
    "print(\"Threshold for upper tail:\", threshold_upper)\n",
    "print(\"Threshold for lower tail:\", threshold_lower)\n",
    "print(\"Thresholds for range around median:\", threshold_lower_range, \"-\", threshold_upper_range)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Retrieve the indices of variants below the lower threshold\n",
    "indices_below_threshold = np.where(diff < threshold_lower)[0]\n",
    "\n",
    "# Retrieve the indices of variants above the upper threshold\n",
    "indices_above_threshold = np.where(diff > threshold_upper)[0]\n",
    "\n",
    "# Retrieve the indices of variants within the range around the median\n",
    "indices_within_range = np.where((diff >= threshold_lower_range) & (diff <= threshold_upper_range))[0]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# print the index names of diff\n",
    "print(diff_res.index[indices_below_threshold])\n",
    "print(diff_res.index[indices_above_threshold])\n",
    "print(diff_res.index[indices_within_range])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 171 = WLEK0650 - V15Q/V43M/R82P/C112S - 4 (V43M only twice in the dataset)\n",
    "# 251 = WLEK0778 - V15Q/Q24L/G27S/R82P/E99A/C112S/I128V/A129T/Q137R - 9\n",
    "# 280 = WLEK0819 - V15Q/I70V/R82P/C112S/Q160L/L191I - 6\n",
    "# 92 = WLEK0500 - V15Q/S38T/L74F/R82P/M100K/C112S/S127T/Y175F - 8\n",
    "# 90 = WLEK0498 - /V15Q/S38T/L74F/R82P/M100K/C112S/S127T/V180A - 8\n",
    "# 194 = WLEK0702 - V15Q/S38T/L74F/R82P/M100K/C112S/S127T/P162S - 8\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import glob\n",
    "from Bio import PDB\n",
    "def compare_sequences(pdb_file, wild_type_pdb_file):\n",
    "    parser = PDB.PDBParser()\n",
    "\n",
    "    # Parse the current structure\n",
    "    structure = parser.get_structure(\"structure\", pdb_file)\n",
    "\n",
    "    # Parse the wild-type structure\n",
    "    wild_type_structure = parser.get_structure(\"wild_type_structure\", wild_type_pdb_file)\n",
    "\n",
    "    # Get the chains from the structures\n",
    "    chains = structure[0]\n",
    "    wild_type_chains = wild_type_structure[0]\n",
    "\n",
    "    # Initialize the mutation count\n",
    "    mutation_count = 0\n",
    "\n",
    "    # Compare the residues in each chain with the wild-type structure\n",
    "    for chain, wild_type_chain in zip(chains, wild_type_chains):\n",
    "        for residue, wild_type_residue in zip(chain, wild_type_chain):\n",
    "            residue_id = residue.get_resname() + str(residue.get_id()[1])\n",
    "            wild_type_residue_id = wild_type_residue.get_resname() + str(wild_type_residue.get_id()[1])\n",
    "            if residue_id != wild_type_residue_id:\n",
    "                mutation_count += 1\n",
    "\n",
    "    return mutation_count"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Directory containing the PDB files\n",
    "pdb_directory = \"/home/nick/Documents/PDBs\"\n",
    "\n",
    "# Wild-type PDB file\n",
    "wild_type_pdb_file = \"/home/nick/PycharmProjects/Engineering_EKB/Python_scripts/wt.pdb\"\n",
    "\n",
    "mutations = {}  # Dictionary to store folder name as key and mutation count as value\n",
    "\n",
    "for folder_name in glob.glob(os.path.join(pdb_directory, \"*\")):\n",
    "    folder_base_name = os.path.basename(folder_name)\n",
    "    if os.path.isdir(folder_name):\n",
    "        for pdb_file in glob.glob(os.path.join(folder_name, \"*.pdb\")):\n",
    "            mutation_count = compare_sequences(pdb_file, wild_type_pdb_file)\n",
    "            mutations[folder_base_name] = mutations.get(folder_base_name, 0) + mutation_count\n",
    "\n",
    "# Create a DataFrame from the mutations dictionary\n",
    "mut_seq = pd.DataFrame(list(mutations.items()), columns=['folder_name', 'mutation_count'])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "mut_seq"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# make of index of diff a column in diff_res\n",
    "diff_res['folder_name'] = diff_res.index\n",
    "diff_res['folder_name'] = diff_res['folder_name'].astype(str)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# merge diff_res and mut_seq at folder_name\n",
    "merged_df = pd.merge(diff_res, mut_seq, on='folder_name', how='left')\n",
    "merged_df[]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# print the count of times mutations is bigger or equal to 5\n",
    "print(merged_df[merged_df['mutation_count'] >= 5].count())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 171 = WLEK0650 - V43M - 1\n",
    "# 251 = WLEK0778 - Q24L/G27S/E99A/I128V/A129T/Q137R - 6\n",
    "# 280 = WLEK0819 - I70V/Q160L/L191I - 3\n",
    "# 92 = WLEK0500 - S38T/L74F/M100K/S127T/Y175F - 5\n",
    "# 90 = WLEK0498 - S38T/L74F/M100K/S127T/V180A - 5\n",
    "# 194 = WLEK0702 - S38T/L74F/M100K/S127T/P162S - 5\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# print max number of mutations\n",
    "# print how many entries have more than 5 mutations\n",
    "print(merged_df[merged_df['mutation_count'] > 5].count())"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.17"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}