LBGMClassifier.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [],
   "source": [
    "from lightgbm import LGBMClassifier, plot_importance\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import accuracy_score, precision_recall_fscore_support\n",
    "from matplotlib import pyplot as plt\n",
    "import seaborn as sns\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from sklearn.model_selection import GridSearchCV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train = train_df.pop('label')\n",
    "x_train = train_df.drop(columns = ['Timestamp', 'Arbitration_ID'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 170,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, Y_train, Y_test = train_test_split(x_train, y_train, test_size=0.01, random_state=123)\n",
    "X_test, X_val, Y_test, Y_val = train_test_split(X_test, Y_test, test_size=0.5, random_state=123)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1]\ttraining's binary_logloss: 0.145543\tvalid_1's binary_logloss: 0.145914\n",
      "Training until validation scores don't improve for 100 rounds\n",
      "[2]\ttraining's binary_logloss: 0.115047\tvalid_1's binary_logloss: 0.116085\n",
      "[3]\ttraining's binary_logloss: 0.0946565\tvalid_1's binary_logloss: 0.095451\n",
      "[4]\ttraining's binary_logloss: 0.080016\tvalid_1's binary_logloss: 0.0805839\n",
      "[5]\ttraining's binary_logloss: 0.0689236\tvalid_1's binary_logloss: 0.0694276\n",
      "[6]\ttraining's binary_logloss: 0.0597133\tvalid_1's binary_logloss: 0.060105\n",
      "[7]\ttraining's binary_logloss: 0.0525229\tvalid_1's binary_logloss: 0.0526963\n",
      "[8]\ttraining's binary_logloss: 0.0462211\tvalid_1's binary_logloss: 0.0465476\n",
      "[9]\ttraining's binary_logloss: 0.0412714\tvalid_1's binary_logloss: 0.0416943\n",
      "[10]\ttraining's binary_logloss: 0.0372544\tvalid_1's binary_logloss: 0.0375388\n",
      "[11]\ttraining's binary_logloss: 0.034003\tvalid_1's binary_logloss: 0.0342123\n",
      "[12]\ttraining's binary_logloss: 0.03066\tvalid_1's binary_logloss: 0.0309853\n",
      "[13]\ttraining's binary_logloss: 0.0281666\tvalid_1's binary_logloss: 0.0285512\n",
      "[14]\ttraining's binary_logloss: 0.0258326\tvalid_1's binary_logloss: 0.0267936\n",
      "[15]\ttraining's binary_logloss: 0.0239394\tvalid_1's binary_logloss: 0.0249261\n",
      "[16]\ttraining's binary_logloss: 0.0219855\tvalid_1's binary_logloss: 0.023072\n",
      "[17]\ttraining's binary_logloss: 0.0204005\tvalid_1's binary_logloss: 0.0214118\n",
      "[18]\ttraining's binary_logloss: 0.0192762\tvalid_1's binary_logloss: 0.0203738\n",
      "[19]\ttraining's binary_logloss: 0.018202\tvalid_1's binary_logloss: 0.0192423\n",
      "[20]\ttraining's binary_logloss: 0.017167\tvalid_1's binary_logloss: 0.0180811\n",
      "[21]\ttraining's binary_logloss: 0.016248\tvalid_1's binary_logloss: 0.0171781\n",
      "[22]\ttraining's binary_logloss: 0.0154215\tvalid_1's binary_logloss: 0.0164509\n",
      "[23]\ttraining's binary_logloss: 0.0147454\tvalid_1's binary_logloss: 0.0158797\n",
      "[24]\ttraining's binary_logloss: 0.0141118\tvalid_1's binary_logloss: 0.0152786\n",
      "[25]\ttraining's binary_logloss: 0.0134688\tvalid_1's binary_logloss: 0.0145761\n",
      "[26]\ttraining's binary_logloss: 0.012997\tvalid_1's binary_logloss: 0.0141931\n",
      "[27]\ttraining's binary_logloss: 0.0125624\tvalid_1's binary_logloss: 0.0137994\n",
      "[28]\ttraining's binary_logloss: 0.0121138\tvalid_1's binary_logloss: 0.0134189\n",
      "[29]\ttraining's binary_logloss: 0.0117503\tvalid_1's binary_logloss: 0.013046\n",
      "[30]\ttraining's binary_logloss: 0.0114013\tvalid_1's binary_logloss: 0.0127782\n",
      "[31]\ttraining's binary_logloss: 0.0110581\tvalid_1's binary_logloss: 0.0124586\n",
      "[32]\ttraining's binary_logloss: 0.010722\tvalid_1's binary_logloss: 0.0120969\n",
      "[33]\ttraining's binary_logloss: 0.0103994\tvalid_1's binary_logloss: 0.0119302\n",
      "[34]\ttraining's binary_logloss: 0.0101288\tvalid_1's binary_logloss: 0.0116558\n",
      "[35]\ttraining's binary_logloss: 0.00988233\tvalid_1's binary_logloss: 0.0114444\n",
      "[36]\ttraining's binary_logloss: 0.0096281\tvalid_1's binary_logloss: 0.0110894\n",
      "[37]\ttraining's binary_logloss: 0.00934849\tvalid_1's binary_logloss: 0.0107742\n",
      "[38]\ttraining's binary_logloss: 0.00910523\tvalid_1's binary_logloss: 0.0105471\n",
      "[39]\ttraining's binary_logloss: 0.00889915\tvalid_1's binary_logloss: 0.010293\n",
      "[40]\ttraining's binary_logloss: 0.00870216\tvalid_1's binary_logloss: 0.0101177\n",
      "[41]\ttraining's binary_logloss: 0.00850136\tvalid_1's binary_logloss: 0.00990547\n",
      "[42]\ttraining's binary_logloss: 0.00830586\tvalid_1's binary_logloss: 0.00970429\n",
      "[43]\ttraining's binary_logloss: 0.0081523\tvalid_1's binary_logloss: 0.00949276\n",
      "[44]\ttraining's binary_logloss: 0.00800879\tvalid_1's binary_logloss: 0.00933893\n",
      "[45]\ttraining's binary_logloss: 0.0078339\tvalid_1's binary_logloss: 0.00926304\n",
      "[46]\ttraining's binary_logloss: 0.007675\tvalid_1's binary_logloss: 0.00910762\n",
      "[47]\ttraining's binary_logloss: 0.00756767\tvalid_1's binary_logloss: 0.00902666\n",
      "[48]\ttraining's binary_logloss: 0.00744541\tvalid_1's binary_logloss: 0.00892536\n",
      "[49]\ttraining's binary_logloss: 0.00731687\tvalid_1's binary_logloss: 0.00877672\n",
      "[50]\ttraining's binary_logloss: 0.00718776\tvalid_1's binary_logloss: 0.00863746\n",
      "[51]\ttraining's binary_logloss: 0.00710147\tvalid_1's binary_logloss: 0.00855264\n",
      "[52]\ttraining's binary_logloss: 0.00697748\tvalid_1's binary_logloss: 0.00848596\n",
      "[53]\ttraining's binary_logloss: 0.00684643\tvalid_1's binary_logloss: 0.0083976\n",
      "[54]\ttraining's binary_logloss: 0.0067517\tvalid_1's binary_logloss: 0.0083022\n",
      "[55]\ttraining's binary_logloss: 0.00663804\tvalid_1's binary_logloss: 0.00822931\n",
      "[56]\ttraining's binary_logloss: 0.00655694\tvalid_1's binary_logloss: 0.00813281\n",
      "[57]\ttraining's binary_logloss: 0.00645362\tvalid_1's binary_logloss: 0.00801037\n",
      "[58]\ttraining's binary_logloss: 0.00638066\tvalid_1's binary_logloss: 0.00792631\n",
      "[59]\ttraining's binary_logloss: 0.00631385\tvalid_1's binary_logloss: 0.00785882\n",
      "[60]\ttraining's binary_logloss: 0.00622287\tvalid_1's binary_logloss: 0.00775328\n",
      "[61]\ttraining's binary_logloss: 0.00615417\tvalid_1's binary_logloss: 0.00769017\n",
      "[62]\ttraining's binary_logloss: 0.00604587\tvalid_1's binary_logloss: 0.00766128\n",
      "[63]\ttraining's binary_logloss: 0.00597989\tvalid_1's binary_logloss: 0.00762087\n",
      "[64]\ttraining's binary_logloss: 0.00590078\tvalid_1's binary_logloss: 0.00753723\n",
      "[65]\ttraining's binary_logloss: 0.0058224\tvalid_1's binary_logloss: 0.00745444\n",
      "[66]\ttraining's binary_logloss: 0.00574488\tvalid_1's binary_logloss: 0.00733497\n",
      "[67]\ttraining's binary_logloss: 0.00565714\tvalid_1's binary_logloss: 0.00723545\n",
      "[68]\ttraining's binary_logloss: 0.00559778\tvalid_1's binary_logloss: 0.00716224\n",
      "[69]\ttraining's binary_logloss: 0.00554927\tvalid_1's binary_logloss: 0.00711616\n",
      "[70]\ttraining's binary_logloss: 0.00548917\tvalid_1's binary_logloss: 0.00702391\n",
      "[71]\ttraining's binary_logloss: 0.00543653\tvalid_1's binary_logloss: 0.00697075\n",
      "[72]\ttraining's binary_logloss: 0.00538012\tvalid_1's binary_logloss: 0.00689158\n",
      "[73]\ttraining's binary_logloss: 0.00534858\tvalid_1's binary_logloss: 0.00686872\n",
      "[74]\ttraining's binary_logloss: 0.00529629\tvalid_1's binary_logloss: 0.00682809\n",
      "[75]\ttraining's binary_logloss: 0.00522673\tvalid_1's binary_logloss: 0.0067471\n",
      "[76]\ttraining's binary_logloss: 0.00516157\tvalid_1's binary_logloss: 0.00669177\n",
      "[77]\ttraining's binary_logloss: 0.00511836\tvalid_1's binary_logloss: 0.00663115\n",
      "[78]\ttraining's binary_logloss: 0.00506129\tvalid_1's binary_logloss: 0.00659092\n",
      "[79]\ttraining's binary_logloss: 0.00501838\tvalid_1's binary_logloss: 0.00656255\n",
      "[80]\ttraining's binary_logloss: 0.00496874\tvalid_1's binary_logloss: 0.00651731\n",
      "[81]\ttraining's binary_logloss: 0.00493463\tvalid_1's binary_logloss: 0.00649004\n",
      "[82]\ttraining's binary_logloss: 0.00490294\tvalid_1's binary_logloss: 0.00643836\n",
      "[83]\ttraining's binary_logloss: 0.00485625\tvalid_1's binary_logloss: 0.0064156\n",
      "[84]\ttraining's binary_logloss: 0.00482373\tvalid_1's binary_logloss: 0.00636232\n",
      "[85]\ttraining's binary_logloss: 0.00478413\tvalid_1's binary_logloss: 0.00634381\n",
      "[86]\ttraining's binary_logloss: 0.00474839\tvalid_1's binary_logloss: 0.00628555\n",
      "[87]\ttraining's binary_logloss: 0.00470484\tvalid_1's binary_logloss: 0.00624521\n",
      "[88]\ttraining's binary_logloss: 0.0046654\tvalid_1's binary_logloss: 0.00614276\n",
      "[89]\ttraining's binary_logloss: 0.00462864\tvalid_1's binary_logloss: 0.00610791\n",
      "[90]\ttraining's binary_logloss: 0.00459762\tvalid_1's binary_logloss: 0.00607021\n",
      "[91]\ttraining's binary_logloss: 0.00455749\tvalid_1's binary_logloss: 0.00604917\n",
      "[92]\ttraining's binary_logloss: 0.00451573\tvalid_1's binary_logloss: 0.0060175\n",
      "[93]\ttraining's binary_logloss: 0.00447783\tvalid_1's binary_logloss: 0.005984\n",
      "[94]\ttraining's binary_logloss: 0.00444414\tvalid_1's binary_logloss: 0.00594913\n",
      "[95]\ttraining's binary_logloss: 0.00441238\tvalid_1's binary_logloss: 0.00593584\n",
      "[96]\ttraining's binary_logloss: 0.00435917\tvalid_1's binary_logloss: 0.00589137\n",
      "[97]\ttraining's binary_logloss: 0.00432869\tvalid_1's binary_logloss: 0.00587926\n",
      "[98]\ttraining's binary_logloss: 0.00428847\tvalid_1's binary_logloss: 0.00580936\n",
      "[99]\ttraining's binary_logloss: 0.00426852\tvalid_1's binary_logloss: 0.00578246\n",
      "[100]\ttraining's binary_logloss: 0.00424675\tvalid_1's binary_logloss: 0.00576128\n",
      "[101]\ttraining's binary_logloss: 0.00421643\tvalid_1's binary_logloss: 0.00572352\n",
      "[102]\ttraining's binary_logloss: 0.0041912\tvalid_1's binary_logloss: 0.00568947\n",
      "[103]\ttraining's binary_logloss: 0.00416147\tvalid_1's binary_logloss: 0.00564381\n",
      "[104]\ttraining's binary_logloss: 0.00413969\tvalid_1's binary_logloss: 0.00561223\n",
      "[105]\ttraining's binary_logloss: 0.00410136\tvalid_1's binary_logloss: 0.00560318\n",
      "[106]\ttraining's binary_logloss: 0.00407842\tvalid_1's binary_logloss: 0.0055934\n",
      "[107]\ttraining's binary_logloss: 0.00405454\tvalid_1's binary_logloss: 0.00556516\n",
      "[108]\ttraining's binary_logloss: 0.00402896\tvalid_1's binary_logloss: 0.00553461\n",
      "[109]\ttraining's binary_logloss: 0.00400308\tvalid_1's binary_logloss: 0.00552258\n",
      "[110]\ttraining's binary_logloss: 0.00396606\tvalid_1's binary_logloss: 0.00548031\n",
      "[111]\ttraining's binary_logloss: 0.00392609\tvalid_1's binary_logloss: 0.00541177\n",
      "[112]\ttraining's binary_logloss: 0.00389133\tvalid_1's binary_logloss: 0.00540197\n",
      "[113]\ttraining's binary_logloss: 0.00385799\tvalid_1's binary_logloss: 0.00533432\n",
      "[114]\ttraining's binary_logloss: 0.00383891\tvalid_1's binary_logloss: 0.00531039\n",
      "[115]\ttraining's binary_logloss: 0.0038204\tvalid_1's binary_logloss: 0.00531451\n",
      "[116]\ttraining's binary_logloss: 0.00380086\tvalid_1's binary_logloss: 0.00531256\n",
      "[117]\ttraining's binary_logloss: 0.00377589\tvalid_1's binary_logloss: 0.00525781\n",
      "[118]\ttraining's binary_logloss: 0.00374752\tvalid_1's binary_logloss: 0.00523378\n",
      "[119]\ttraining's binary_logloss: 0.00373124\tvalid_1's binary_logloss: 0.00522823\n",
      "[120]\ttraining's binary_logloss: 0.00370674\tvalid_1's binary_logloss: 0.00521192\n",
      "[121]\ttraining's binary_logloss: 0.00368008\tvalid_1's binary_logloss: 0.00518143\n",
      "[122]\ttraining's binary_logloss: 0.00366424\tvalid_1's binary_logloss: 0.00517447\n",
      "[123]\ttraining's binary_logloss: 0.00364358\tvalid_1's binary_logloss: 0.00515159\n",
      "[124]\ttraining's binary_logloss: 0.00361559\tvalid_1's binary_logloss: 0.00512878\n",
      "[125]\ttraining's binary_logloss: 0.00359403\tvalid_1's binary_logloss: 0.00510471\n",
      "[126]\ttraining's binary_logloss: 0.00357291\tvalid_1's binary_logloss: 0.00505892\n",
      "[127]\ttraining's binary_logloss: 0.00354917\tvalid_1's binary_logloss: 0.00502781\n",
      "[128]\ttraining's binary_logloss: 0.00351174\tvalid_1's binary_logloss: 0.00500831\n",
      "[129]\ttraining's binary_logloss: 0.00349182\tvalid_1's binary_logloss: 0.00498779\n",
      "[130]\ttraining's binary_logloss: 0.00347485\tvalid_1's binary_logloss: 0.00497649\n",
      "[131]\ttraining's binary_logloss: 0.00345842\tvalid_1's binary_logloss: 0.00495962\n",
      "[132]\ttraining's binary_logloss: 0.0034227\tvalid_1's binary_logloss: 0.00491804\n",
      "[133]\ttraining's binary_logloss: 0.00339358\tvalid_1's binary_logloss: 0.00491762\n",
      "[134]\ttraining's binary_logloss: 0.00337322\tvalid_1's binary_logloss: 0.00490223\n",
      "[135]\ttraining's binary_logloss: 0.00335971\tvalid_1's binary_logloss: 0.00489634\n",
      "[136]\ttraining's binary_logloss: 0.00334003\tvalid_1's binary_logloss: 0.00488073\n",
      "[137]\ttraining's binary_logloss: 0.00332328\tvalid_1's binary_logloss: 0.00487761\n",
      "[138]\ttraining's binary_logloss: 0.0033094\tvalid_1's binary_logloss: 0.00484515\n",
      "[139]\ttraining's binary_logloss: 0.00329157\tvalid_1's binary_logloss: 0.00482704\n",
      "[140]\ttraining's binary_logloss: 0.00327818\tvalid_1's binary_logloss: 0.00482082\n",
      "[141]\ttraining's binary_logloss: 0.00326896\tvalid_1's binary_logloss: 0.00482917\n",
      "[142]\ttraining's binary_logloss: 0.00324788\tvalid_1's binary_logloss: 0.00483508\n",
      "[143]\ttraining's binary_logloss: 0.0032316\tvalid_1's binary_logloss: 0.00481523\n",
      "[144]\ttraining's binary_logloss: 0.00321252\tvalid_1's binary_logloss: 0.00479012\n",
      "[145]\ttraining's binary_logloss: 0.00320232\tvalid_1's binary_logloss: 0.00480646\n",
      "[146]\ttraining's binary_logloss: 0.00318831\tvalid_1's binary_logloss: 0.00479667\n",
      "[147]\ttraining's binary_logloss: 0.0031763\tvalid_1's binary_logloss: 0.004801\n",
      "[148]\ttraining's binary_logloss: 0.0031605\tvalid_1's binary_logloss: 0.00478669\n",
      "[149]\ttraining's binary_logloss: 0.00314074\tvalid_1's binary_logloss: 0.00478036\n",
      "[150]\ttraining's binary_logloss: 0.00312864\tvalid_1's binary_logloss: 0.00475903\n",
      "[151]\ttraining's binary_logloss: 0.00311113\tvalid_1's binary_logloss: 0.00471369\n",
      "[152]\ttraining's binary_logloss: 0.00309435\tvalid_1's binary_logloss: 0.00470452\n",
      "[153]\ttraining's binary_logloss: 0.00307812\tvalid_1's binary_logloss: 0.00472721\n",
      "[154]\ttraining's binary_logloss: 0.00306463\tvalid_1's binary_logloss: 0.00472415\n",
      "[155]\ttraining's binary_logloss: 0.00304847\tvalid_1's binary_logloss: 0.00468331\n",
      "[156]\ttraining's binary_logloss: 0.00303651\tvalid_1's binary_logloss: 0.00466383\n",
      "[157]\ttraining's binary_logloss: 0.0030174\tvalid_1's binary_logloss: 0.00466159\n",
      "[158]\ttraining's binary_logloss: 0.0030061\tvalid_1's binary_logloss: 0.00466124\n",
      "[159]\ttraining's binary_logloss: 0.00299544\tvalid_1's binary_logloss: 0.00465829\n",
      "[160]\ttraining's binary_logloss: 0.00298384\tvalid_1's binary_logloss: 0.00465833\n",
      "[161]\ttraining's binary_logloss: 0.00297217\tvalid_1's binary_logloss: 0.0046558\n",
      "[162]\ttraining's binary_logloss: 0.00294945\tvalid_1's binary_logloss: 0.0046065\n",
      "[163]\ttraining's binary_logloss: 0.00293758\tvalid_1's binary_logloss: 0.00459128\n",
      "[164]\ttraining's binary_logloss: 0.00292803\tvalid_1's binary_logloss: 0.00458174\n",
      "[165]\ttraining's binary_logloss: 0.0029193\tvalid_1's binary_logloss: 0.00457686\n",
      "[166]\ttraining's binary_logloss: 0.00290648\tvalid_1's binary_logloss: 0.0045694\n",
      "[167]\ttraining's binary_logloss: 0.00289095\tvalid_1's binary_logloss: 0.00455208\n",
      "[168]\ttraining's binary_logloss: 0.00287579\tvalid_1's binary_logloss: 0.00453814\n",
      "[169]\ttraining's binary_logloss: 0.00286062\tvalid_1's binary_logloss: 0.00451871\n",
      "[170]\ttraining's binary_logloss: 0.00284839\tvalid_1's binary_logloss: 0.00450701\n",
      "[171]\ttraining's binary_logloss: 0.00283739\tvalid_1's binary_logloss: 0.00448679\n",
      "[172]\ttraining's binary_logloss: 0.00282566\tvalid_1's binary_logloss: 0.00446247\n",
      "[173]\ttraining's binary_logloss: 0.00281041\tvalid_1's binary_logloss: 0.00444786\n",
      "[174]\ttraining's binary_logloss: 0.00279898\tvalid_1's binary_logloss: 0.00442452\n",
      "[175]\ttraining's binary_logloss: 0.00279004\tvalid_1's binary_logloss: 0.00442072\n",
      "[176]\ttraining's binary_logloss: 0.00278023\tvalid_1's binary_logloss: 0.00441196\n",
      "[177]\ttraining's binary_logloss: 0.00276844\tvalid_1's binary_logloss: 0.00440043\n",
      "[178]\ttraining's binary_logloss: 0.00275931\tvalid_1's binary_logloss: 0.00440036\n",
      "[179]\ttraining's binary_logloss: 0.00274782\tvalid_1's binary_logloss: 0.00440006\n",
      "[180]\ttraining's binary_logloss: 0.00273988\tvalid_1's binary_logloss: 0.00438263\n",
      "[181]\ttraining's binary_logloss: 0.00273203\tvalid_1's binary_logloss: 0.00439733\n",
      "[182]\ttraining's binary_logloss: 0.00272364\tvalid_1's binary_logloss: 0.00439858\n",
      "[183]\ttraining's binary_logloss: 0.00270805\tvalid_1's binary_logloss: 0.00438322\n",
      "[184]\ttraining's binary_logloss: 0.00269869\tvalid_1's binary_logloss: 0.00437689\n",
      "[185]\ttraining's binary_logloss: 0.00268673\tvalid_1's binary_logloss: 0.00437508\n",
      "[186]\ttraining's binary_logloss: 0.00267378\tvalid_1's binary_logloss: 0.00436302\n",
      "[187]\ttraining's binary_logloss: 0.00266459\tvalid_1's binary_logloss: 0.00434688\n",
      "[188]\ttraining's binary_logloss: 0.00265776\tvalid_1's binary_logloss: 0.00434565\n",
      "[189]\ttraining's binary_logloss: 0.00264951\tvalid_1's binary_logloss: 0.00433443\n",
      "[190]\ttraining's binary_logloss: 0.00263854\tvalid_1's binary_logloss: 0.00433256\n",
      "[191]\ttraining's binary_logloss: 0.00262893\tvalid_1's binary_logloss: 0.00432959\n",
      "[192]\ttraining's binary_logloss: 0.00261741\tvalid_1's binary_logloss: 0.00431388\n",
      "[193]\ttraining's binary_logloss: 0.00261071\tvalid_1's binary_logloss: 0.00430993\n",
      "[194]\ttraining's binary_logloss: 0.00260352\tvalid_1's binary_logloss: 0.00430231\n",
      "[195]\ttraining's binary_logloss: 0.00259506\tvalid_1's binary_logloss: 0.00431696\n",
      "[196]\ttraining's binary_logloss: 0.00258633\tvalid_1's binary_logloss: 0.00430536\n",
      "[197]\ttraining's binary_logloss: 0.00257982\tvalid_1's binary_logloss: 0.00430222\n",
      "[198]\ttraining's binary_logloss: 0.00257181\tvalid_1's binary_logloss: 0.0042841\n",
      "[199]\ttraining's binary_logloss: 0.00256763\tvalid_1's binary_logloss: 0.00429037\n",
      "[200]\ttraining's binary_logloss: 0.002562\tvalid_1's binary_logloss: 0.00427947\n",
      "[201]\ttraining's binary_logloss: 0.00255249\tvalid_1's binary_logloss: 0.00426422\n",
      "[202]\ttraining's binary_logloss: 0.00254076\tvalid_1's binary_logloss: 0.00426215\n",
      "[203]\ttraining's binary_logloss: 0.00253218\tvalid_1's binary_logloss: 0.00424056\n",
      "[204]\ttraining's binary_logloss: 0.0025243\tvalid_1's binary_logloss: 0.00424148\n",
      "[205]\ttraining's binary_logloss: 0.00251746\tvalid_1's binary_logloss: 0.00422707\n",
      "[206]\ttraining's binary_logloss: 0.00250772\tvalid_1's binary_logloss: 0.00422933\n",
      "[207]\ttraining's binary_logloss: 0.00250039\tvalid_1's binary_logloss: 0.00423107\n",
      "[208]\ttraining's binary_logloss: 0.00249151\tvalid_1's binary_logloss: 0.0042151\n",
      "[209]\ttraining's binary_logloss: 0.00247861\tvalid_1's binary_logloss: 0.00419286\n",
      "[210]\ttraining's binary_logloss: 0.00247073\tvalid_1's binary_logloss: 0.00418027\n",
      "[211]\ttraining's binary_logloss: 0.00246269\tvalid_1's binary_logloss: 0.00417883\n",
      "[212]\ttraining's binary_logloss: 0.00245383\tvalid_1's binary_logloss: 0.00417499\n",
      "[213]\ttraining's binary_logloss: 0.00244561\tvalid_1's binary_logloss: 0.00417349\n",
      "[214]\ttraining's binary_logloss: 0.00243723\tvalid_1's binary_logloss: 0.00415218\n",
      "[215]\ttraining's binary_logloss: 0.00242931\tvalid_1's binary_logloss: 0.00414284\n",
      "[216]\ttraining's binary_logloss: 0.00242319\tvalid_1's binary_logloss: 0.00414345\n",
      "[217]\ttraining's binary_logloss: 0.00241738\tvalid_1's binary_logloss: 0.0041523\n",
      "[218]\ttraining's binary_logloss: 0.00241011\tvalid_1's binary_logloss: 0.00415062\n",
      "[219]\ttraining's binary_logloss: 0.00239887\tvalid_1's binary_logloss: 0.0041388\n",
      "[220]\ttraining's binary_logloss: 0.0023889\tvalid_1's binary_logloss: 0.00413108\n",
      "[221]\ttraining's binary_logloss: 0.00237811\tvalid_1's binary_logloss: 0.00408943\n",
      "[222]\ttraining's binary_logloss: 0.00236956\tvalid_1's binary_logloss: 0.00408049\n",
      "[223]\ttraining's binary_logloss: 0.00236074\tvalid_1's binary_logloss: 0.00407437\n",
      "[224]\ttraining's binary_logloss: 0.00235541\tvalid_1's binary_logloss: 0.00406722\n",
      "[225]\ttraining's binary_logloss: 0.00234508\tvalid_1's binary_logloss: 0.00405485\n",
      "[226]\ttraining's binary_logloss: 0.00233735\tvalid_1's binary_logloss: 0.00405054\n",
      "[227]\ttraining's binary_logloss: 0.00232742\tvalid_1's binary_logloss: 0.00402155\n",
      "[228]\ttraining's binary_logloss: 0.00232112\tvalid_1's binary_logloss: 0.00402411\n",
      "[229]\ttraining's binary_logloss: 0.00231511\tvalid_1's binary_logloss: 0.00401081\n",
      "[230]\ttraining's binary_logloss: 0.00230791\tvalid_1's binary_logloss: 0.00400413\n",
      "[231]\ttraining's binary_logloss: 0.00230189\tvalid_1's binary_logloss: 0.00399912\n",
      "[232]\ttraining's binary_logloss: 0.00229514\tvalid_1's binary_logloss: 0.00399882\n",
      "[233]\ttraining's binary_logloss: 0.00228579\tvalid_1's binary_logloss: 0.00398492\n",
      "[234]\ttraining's binary_logloss: 0.00227897\tvalid_1's binary_logloss: 0.0039754\n",
      "[235]\ttraining's binary_logloss: 0.00226894\tvalid_1's binary_logloss: 0.00397343\n",
      "[236]\ttraining's binary_logloss: 0.00226317\tvalid_1's binary_logloss: 0.00395742\n",
      "[237]\ttraining's binary_logloss: 0.00225737\tvalid_1's binary_logloss: 0.00394862\n",
      "[238]\ttraining's binary_logloss: 0.00225074\tvalid_1's binary_logloss: 0.00394691\n",
      "[239]\ttraining's binary_logloss: 0.00224454\tvalid_1's binary_logloss: 0.00393424\n",
      "[240]\ttraining's binary_logloss: 0.00223806\tvalid_1's binary_logloss: 0.00393239\n",
      "[241]\ttraining's binary_logloss: 0.00223057\tvalid_1's binary_logloss: 0.00391667\n",
      "[242]\ttraining's binary_logloss: 0.00222546\tvalid_1's binary_logloss: 0.00392186\n",
      "[243]\ttraining's binary_logloss: 0.00222121\tvalid_1's binary_logloss: 0.00391851\n",
      "[244]\ttraining's binary_logloss: 0.00221361\tvalid_1's binary_logloss: 0.0039173\n",
      "[245]\ttraining's binary_logloss: 0.00220731\tvalid_1's binary_logloss: 0.00390196\n",
      "[246]\ttraining's binary_logloss: 0.00220071\tvalid_1's binary_logloss: 0.0038853\n",
      "[247]\ttraining's binary_logloss: 0.00219309\tvalid_1's binary_logloss: 0.00388483\n",
      "[248]\ttraining's binary_logloss: 0.00218609\tvalid_1's binary_logloss: 0.00387434\n",
      "[249]\ttraining's binary_logloss: 0.00217793\tvalid_1's binary_logloss: 0.00385212\n",
      "[250]\ttraining's binary_logloss: 0.00217047\tvalid_1's binary_logloss: 0.00384767\n",
      "[251]\ttraining's binary_logloss: 0.00216564\tvalid_1's binary_logloss: 0.00385433\n",
      "[252]\ttraining's binary_logloss: 0.00216051\tvalid_1's binary_logloss: 0.00384368\n",
      "[253]\ttraining's binary_logloss: 0.0021547\tvalid_1's binary_logloss: 0.00383574\n",
      "[254]\ttraining's binary_logloss: 0.00214958\tvalid_1's binary_logloss: 0.00383856\n",
      "[255]\ttraining's binary_logloss: 0.00214331\tvalid_1's binary_logloss: 0.00383449\n",
      "[256]\ttraining's binary_logloss: 0.00213457\tvalid_1's binary_logloss: 0.00382673\n",
      "[257]\ttraining's binary_logloss: 0.00212802\tvalid_1's binary_logloss: 0.00381255\n",
      "[258]\ttraining's binary_logloss: 0.00211698\tvalid_1's binary_logloss: 0.00381595\n",
      "[259]\ttraining's binary_logloss: 0.00211189\tvalid_1's binary_logloss: 0.00381127\n",
      "[260]\ttraining's binary_logloss: 0.002105\tvalid_1's binary_logloss: 0.00380338\n",
      "[261]\ttraining's binary_logloss: 0.0020981\tvalid_1's binary_logloss: 0.0037989\n",
      "[262]\ttraining's binary_logloss: 0.00209326\tvalid_1's binary_logloss: 0.00379228\n",
      "[263]\ttraining's binary_logloss: 0.00208789\tvalid_1's binary_logloss: 0.00379012\n",
      "[264]\ttraining's binary_logloss: 0.00208199\tvalid_1's binary_logloss: 0.00377519\n",
      "[265]\ttraining's binary_logloss: 0.00207784\tvalid_1's binary_logloss: 0.00377433\n",
      "[266]\ttraining's binary_logloss: 0.0020716\tvalid_1's binary_logloss: 0.00376945\n",
      "[267]\ttraining's binary_logloss: 0.00206588\tvalid_1's binary_logloss: 0.00376918\n",
      "[268]\ttraining's binary_logloss: 0.00205965\tvalid_1's binary_logloss: 0.0037665\n",
      "[269]\ttraining's binary_logloss: 0.00205298\tvalid_1's binary_logloss: 0.00375304\n",
      "[270]\ttraining's binary_logloss: 0.00204632\tvalid_1's binary_logloss: 0.00375351\n",
      "[271]\ttraining's binary_logloss: 0.00204143\tvalid_1's binary_logloss: 0.00374938\n",
      "[272]\ttraining's binary_logloss: 0.00203618\tvalid_1's binary_logloss: 0.00375242\n",
      "[273]\ttraining's binary_logloss: 0.0020318\tvalid_1's binary_logloss: 0.00373305\n",
      "[274]\ttraining's binary_logloss: 0.00202593\tvalid_1's binary_logloss: 0.00372535\n",
      "[275]\ttraining's binary_logloss: 0.00202028\tvalid_1's binary_logloss: 0.00372774\n",
      "[276]\ttraining's binary_logloss: 0.00201445\tvalid_1's binary_logloss: 0.00372153\n",
      "[277]\ttraining's binary_logloss: 0.00201071\tvalid_1's binary_logloss: 0.00371153\n",
      "[278]\ttraining's binary_logloss: 0.00200418\tvalid_1's binary_logloss: 0.00371746\n",
      "[279]\ttraining's binary_logloss: 0.00199733\tvalid_1's binary_logloss: 0.00371466\n",
      "[280]\ttraining's binary_logloss: 0.00199351\tvalid_1's binary_logloss: 0.00370212\n",
      "[281]\ttraining's binary_logloss: 0.00198761\tvalid_1's binary_logloss: 0.00368981\n",
      "[282]\ttraining's binary_logloss: 0.00198166\tvalid_1's binary_logloss: 0.00365223\n",
      "[283]\ttraining's binary_logloss: 0.0019769\tvalid_1's binary_logloss: 0.00365799\n",
      "[284]\ttraining's binary_logloss: 0.0019723\tvalid_1's binary_logloss: 0.00364406\n",
      "[285]\ttraining's binary_logloss: 0.00196628\tvalid_1's binary_logloss: 0.00362973\n",
      "[286]\ttraining's binary_logloss: 0.00196236\tvalid_1's binary_logloss: 0.00363239\n",
      "[287]\ttraining's binary_logloss: 0.00195446\tvalid_1's binary_logloss: 0.0036282\n",
      "[288]\ttraining's binary_logloss: 0.00194971\tvalid_1's binary_logloss: 0.00362544\n",
      "[289]\ttraining's binary_logloss: 0.00194421\tvalid_1's binary_logloss: 0.00361756\n",
      "[290]\ttraining's binary_logloss: 0.00193917\tvalid_1's binary_logloss: 0.00361885\n",
      "[291]\ttraining's binary_logloss: 0.00193495\tvalid_1's binary_logloss: 0.00361826\n",
      "[292]\ttraining's binary_logloss: 0.00193018\tvalid_1's binary_logloss: 0.00359819\n",
      "[293]\ttraining's binary_logloss: 0.00192545\tvalid_1's binary_logloss: 0.00359958\n",
      "[294]\ttraining's binary_logloss: 0.00192086\tvalid_1's binary_logloss: 0.00359593\n",
      "[295]\ttraining's binary_logloss: 0.00191611\tvalid_1's binary_logloss: 0.00360146\n",
      "[296]\ttraining's binary_logloss: 0.00191089\tvalid_1's binary_logloss: 0.00360363\n",
      "[297]\ttraining's binary_logloss: 0.00190593\tvalid_1's binary_logloss: 0.00360892\n",
      "[298]\ttraining's binary_logloss: 0.0019021\tvalid_1's binary_logloss: 0.00360465\n",
      "[299]\ttraining's binary_logloss: 0.00189715\tvalid_1's binary_logloss: 0.00359318\n",
      "[300]\ttraining's binary_logloss: 0.00189263\tvalid_1's binary_logloss: 0.00358199\n",
      "[301]\ttraining's binary_logloss: 0.001886\tvalid_1's binary_logloss: 0.00355933\n",
      "[302]\ttraining's binary_logloss: 0.00188272\tvalid_1's binary_logloss: 0.00356504\n",
      "[303]\ttraining's binary_logloss: 0.00187891\tvalid_1's binary_logloss: 0.00357742\n",
      "[304]\ttraining's binary_logloss: 0.00187475\tvalid_1's binary_logloss: 0.00357809\n",
      "[305]\ttraining's binary_logloss: 0.00186786\tvalid_1's binary_logloss: 0.00357522\n",
      "[306]\ttraining's binary_logloss: 0.00186312\tvalid_1's binary_logloss: 0.00356941\n",
      "[307]\ttraining's binary_logloss: 0.00186009\tvalid_1's binary_logloss: 0.00355879\n",
      "[308]\ttraining's binary_logloss: 0.00185648\tvalid_1's binary_logloss: 0.00355935\n",
      "[309]\ttraining's binary_logloss: 0.00185088\tvalid_1's binary_logloss: 0.00355996\n",
      "[310]\ttraining's binary_logloss: 0.00184648\tvalid_1's binary_logloss: 0.00357821\n",
      "[311]\ttraining's binary_logloss: 0.00184176\tvalid_1's binary_logloss: 0.00357241\n",
      "[312]\ttraining's binary_logloss: 0.00183555\tvalid_1's binary_logloss: 0.00356132\n",
      "[313]\ttraining's binary_logloss: 0.00183112\tvalid_1's binary_logloss: 0.00356439\n",
      "[314]\ttraining's binary_logloss: 0.00182594\tvalid_1's binary_logloss: 0.00355722\n",
      "[315]\ttraining's binary_logloss: 0.00182303\tvalid_1's binary_logloss: 0.00355046\n",
      "[316]\ttraining's binary_logloss: 0.00181878\tvalid_1's binary_logloss: 0.00354964\n",
      "[317]\ttraining's binary_logloss: 0.00181568\tvalid_1's binary_logloss: 0.00355506\n",
      "[318]\ttraining's binary_logloss: 0.00181104\tvalid_1's binary_logloss: 0.0035504\n",
      "[319]\ttraining's binary_logloss: 0.00180663\tvalid_1's binary_logloss: 0.00356004\n",
      "[320]\ttraining's binary_logloss: 0.00180309\tvalid_1's binary_logloss: 0.00355673\n",
      "[321]\ttraining's binary_logloss: 0.00179862\tvalid_1's binary_logloss: 0.00356231\n",
      "[322]\ttraining's binary_logloss: 0.0017948\tvalid_1's binary_logloss: 0.00355945\n",
      "[323]\ttraining's binary_logloss: 0.0017911\tvalid_1's binary_logloss: 0.00356727\n",
      "[324]\ttraining's binary_logloss: 0.00178835\tvalid_1's binary_logloss: 0.00357015\n",
      "[325]\ttraining's binary_logloss: 0.00178485\tvalid_1's binary_logloss: 0.0035653\n",
      "[326]\ttraining's binary_logloss: 0.00177972\tvalid_1's binary_logloss: 0.00356431\n",
      "[327]\ttraining's binary_logloss: 0.00177623\tvalid_1's binary_logloss: 0.00356136\n",
      "[328]\ttraining's binary_logloss: 0.00177247\tvalid_1's binary_logloss: 0.00357655\n",
      "[329]\ttraining's binary_logloss: 0.00176768\tvalid_1's binary_logloss: 0.00356356\n",
      "[330]\ttraining's binary_logloss: 0.0017647\tvalid_1's binary_logloss: 0.00356116\n",
      "[331]\ttraining's binary_logloss: 0.00176092\tvalid_1's binary_logloss: 0.00355898\n",
      "[332]\ttraining's binary_logloss: 0.00175726\tvalid_1's binary_logloss: 0.00355128\n",
      "[333]\ttraining's binary_logloss: 0.00175332\tvalid_1's binary_logloss: 0.00355268\n",
      "[334]\ttraining's binary_logloss: 0.00175024\tvalid_1's binary_logloss: 0.0035545\n",
      "[335]\ttraining's binary_logloss: 0.0017465\tvalid_1's binary_logloss: 0.00354149\n",
      "[336]\ttraining's binary_logloss: 0.00174344\tvalid_1's binary_logloss: 0.00354435\n",
      "[337]\ttraining's binary_logloss: 0.00173944\tvalid_1's binary_logloss: 0.00355572\n",
      "[338]\ttraining's binary_logloss: 0.00173453\tvalid_1's binary_logloss: 0.00355245\n",
      "[339]\ttraining's binary_logloss: 0.00173054\tvalid_1's binary_logloss: 0.00355075\n",
      "[340]\ttraining's binary_logloss: 0.00172631\tvalid_1's binary_logloss: 0.0035425\n",
      "[341]\ttraining's binary_logloss: 0.00172236\tvalid_1's binary_logloss: 0.00354038\n",
      "[342]\ttraining's binary_logloss: 0.00171791\tvalid_1's binary_logloss: 0.00353122\n",
      "[343]\ttraining's binary_logloss: 0.00171421\tvalid_1's binary_logloss: 0.00353128\n",
      "[344]\ttraining's binary_logloss: 0.00171119\tvalid_1's binary_logloss: 0.00352895\n",
      "[345]\ttraining's binary_logloss: 0.00170786\tvalid_1's binary_logloss: 0.00352092\n",
      "[346]\ttraining's binary_logloss: 0.00170462\tvalid_1's binary_logloss: 0.00352669\n",
      "[347]\ttraining's binary_logloss: 0.00170125\tvalid_1's binary_logloss: 0.00351065\n",
      "[348]\ttraining's binary_logloss: 0.0016977\tvalid_1's binary_logloss: 0.00351805\n",
      "[349]\ttraining's binary_logloss: 0.00169399\tvalid_1's binary_logloss: 0.00351661\n",
      "[350]\ttraining's binary_logloss: 0.00169164\tvalid_1's binary_logloss: 0.00352106\n",
      "[351]\ttraining's binary_logloss: 0.00168785\tvalid_1's binary_logloss: 0.00351916\n",
      "[352]\ttraining's binary_logloss: 0.00168396\tvalid_1's binary_logloss: 0.00352795\n",
      "[353]\ttraining's binary_logloss: 0.00168082\tvalid_1's binary_logloss: 0.00352132\n",
      "[354]\ttraining's binary_logloss: 0.00167725\tvalid_1's binary_logloss: 0.00351596\n",
      "[355]\ttraining's binary_logloss: 0.00167459\tvalid_1's binary_logloss: 0.00351346\n",
      "[356]\ttraining's binary_logloss: 0.00167092\tvalid_1's binary_logloss: 0.00352365\n",
      "[357]\ttraining's binary_logloss: 0.00166759\tvalid_1's binary_logloss: 0.00352387\n",
      "[358]\ttraining's binary_logloss: 0.0016651\tvalid_1's binary_logloss: 0.00352234\n",
      "[359]\ttraining's binary_logloss: 0.0016595\tvalid_1's binary_logloss: 0.00352428\n",
      "[360]\ttraining's binary_logloss: 0.00165728\tvalid_1's binary_logloss: 0.00351162\n",
      "[361]\ttraining's binary_logloss: 0.00165208\tvalid_1's binary_logloss: 0.00351698\n",
      "[362]\ttraining's binary_logloss: 0.00164978\tvalid_1's binary_logloss: 0.00349961\n",
      "[363]\ttraining's binary_logloss: 0.00164691\tvalid_1's binary_logloss: 0.00350026\n",
      "[364]\ttraining's binary_logloss: 0.00164314\tvalid_1's binary_logloss: 0.00349342\n",
      "[365]\ttraining's binary_logloss: 0.00163943\tvalid_1's binary_logloss: 0.00349324\n",
      "[366]\ttraining's binary_logloss: 0.0016364\tvalid_1's binary_logloss: 0.00350752\n",
      "[367]\ttraining's binary_logloss: 0.00163369\tvalid_1's binary_logloss: 0.00352506\n",
      "[368]\ttraining's binary_logloss: 0.00163082\tvalid_1's binary_logloss: 0.00351994\n",
      "[369]\ttraining's binary_logloss: 0.00162736\tvalid_1's binary_logloss: 0.00352919\n",
      "[370]\ttraining's binary_logloss: 0.00162435\tvalid_1's binary_logloss: 0.00352769\n",
      "[371]\ttraining's binary_logloss: 0.00162157\tvalid_1's binary_logloss: 0.00353504\n",
      "[372]\ttraining's binary_logloss: 0.00161835\tvalid_1's binary_logloss: 0.00353312\n",
      "[373]\ttraining's binary_logloss: 0.001615\tvalid_1's binary_logloss: 0.00352765\n",
      "[374]\ttraining's binary_logloss: 0.00161211\tvalid_1's binary_logloss: 0.00353257\n",
      "[375]\ttraining's binary_logloss: 0.00160876\tvalid_1's binary_logloss: 0.0035261\n",
      "[376]\ttraining's binary_logloss: 0.00160582\tvalid_1's binary_logloss: 0.00352073\n",
      "[377]\ttraining's binary_logloss: 0.00160264\tvalid_1's binary_logloss: 0.00352685\n",
      "[378]\ttraining's binary_logloss: 0.00159898\tvalid_1's binary_logloss: 0.00352552\n",
      "[379]\ttraining's binary_logloss: 0.00159586\tvalid_1's binary_logloss: 0.00352132\n",
      "[380]\ttraining's binary_logloss: 0.00159331\tvalid_1's binary_logloss: 0.00352055\n",
      "[381]\ttraining's binary_logloss: 0.00159072\tvalid_1's binary_logloss: 0.0035051\n",
      "[382]\ttraining's binary_logloss: 0.00158831\tvalid_1's binary_logloss: 0.00350714\n",
      "[383]\ttraining's binary_logloss: 0.0015864\tvalid_1's binary_logloss: 0.00349301\n",
      "[384]\ttraining's binary_logloss: 0.00158346\tvalid_1's binary_logloss: 0.00349266\n",
      "[385]\ttraining's binary_logloss: 0.00158077\tvalid_1's binary_logloss: 0.00349252\n",
      "[386]\ttraining's binary_logloss: 0.00157765\tvalid_1's binary_logloss: 0.00349449\n",
      "[387]\ttraining's binary_logloss: 0.00157433\tvalid_1's binary_logloss: 0.00350174\n",
      "[388]\ttraining's binary_logloss: 0.00157172\tvalid_1's binary_logloss: 0.00350127\n",
      "[389]\ttraining's binary_logloss: 0.00156836\tvalid_1's binary_logloss: 0.00349961\n",
      "[390]\ttraining's binary_logloss: 0.00156593\tvalid_1's binary_logloss: 0.00350908\n",
      "[391]\ttraining's binary_logloss: 0.00156314\tvalid_1's binary_logloss: 0.00351432\n",
      "[392]\ttraining's binary_logloss: 0.00155997\tvalid_1's binary_logloss: 0.00351929\n",
      "[393]\ttraining's binary_logloss: 0.00155777\tvalid_1's binary_logloss: 0.0035097\n",
      "[394]\ttraining's binary_logloss: 0.00155388\tvalid_1's binary_logloss: 0.00350986\n",
      "[395]\ttraining's binary_logloss: 0.00155147\tvalid_1's binary_logloss: 0.00351332\n",
      "[396]\ttraining's binary_logloss: 0.00154908\tvalid_1's binary_logloss: 0.00351505\n",
      "[397]\ttraining's binary_logloss: 0.00154618\tvalid_1's binary_logloss: 0.00350795\n",
      "[398]\ttraining's binary_logloss: 0.00154249\tvalid_1's binary_logloss: 0.00350762\n",
      "[399]\ttraining's binary_logloss: 0.00153975\tvalid_1's binary_logloss: 0.00350855\n",
      "[400]\ttraining's binary_logloss: 0.00153724\tvalid_1's binary_logloss: 0.00350046\n",
      "[401]\ttraining's binary_logloss: 0.00153453\tvalid_1's binary_logloss: 0.00346886\n",
      "[402]\ttraining's binary_logloss: 0.00153206\tvalid_1's binary_logloss: 0.00347132\n",
      "[403]\ttraining's binary_logloss: 0.00152903\tvalid_1's binary_logloss: 0.00347193\n",
      "[404]\ttraining's binary_logloss: 0.0015264\tvalid_1's binary_logloss: 0.00348071\n",
      "[405]\ttraining's binary_logloss: 0.00152332\tvalid_1's binary_logloss: 0.00347674\n",
      "[406]\ttraining's binary_logloss: 0.00152084\tvalid_1's binary_logloss: 0.00347989\n",
      "[407]\ttraining's binary_logloss: 0.0015177\tvalid_1's binary_logloss: 0.00347882\n",
      "[408]\ttraining's binary_logloss: 0.00151469\tvalid_1's binary_logloss: 0.00347853\n",
      "[409]\ttraining's binary_logloss: 0.00151232\tvalid_1's binary_logloss: 0.00347485\n",
      "[410]\ttraining's binary_logloss: 0.00150981\tvalid_1's binary_logloss: 0.00347611\n",
      "[411]\ttraining's binary_logloss: 0.00150695\tvalid_1's binary_logloss: 0.003466\n",
      "[412]\ttraining's binary_logloss: 0.0015043\tvalid_1's binary_logloss: 0.00346661\n",
      "[413]\ttraining's binary_logloss: 0.0015019\tvalid_1's binary_logloss: 0.00345876\n",
      "[414]\ttraining's binary_logloss: 0.0014987\tvalid_1's binary_logloss: 0.00345789\n",
      "[415]\ttraining's binary_logloss: 0.00149569\tvalid_1's binary_logloss: 0.00345338\n",
      "[416]\ttraining's binary_logloss: 0.00149274\tvalid_1's binary_logloss: 0.00344928\n",
      "[417]\ttraining's binary_logloss: 0.0014902\tvalid_1's binary_logloss: 0.00344843\n",
      "[418]\ttraining's binary_logloss: 0.00148789\tvalid_1's binary_logloss: 0.00345463\n",
      "[419]\ttraining's binary_logloss: 0.00148327\tvalid_1's binary_logloss: 0.00345065\n",
      "[420]\ttraining's binary_logloss: 0.00147985\tvalid_1's binary_logloss: 0.00343807\n",
      "[421]\ttraining's binary_logloss: 0.0014777\tvalid_1's binary_logloss: 0.00344041\n",
      "[422]\ttraining's binary_logloss: 0.0014751\tvalid_1's binary_logloss: 0.00343196\n",
      "[423]\ttraining's binary_logloss: 0.0014728\tvalid_1's binary_logloss: 0.00343118\n",
      "[424]\ttraining's binary_logloss: 0.00147006\tvalid_1's binary_logloss: 0.00343477\n",
      "[425]\ttraining's binary_logloss: 0.00146765\tvalid_1's binary_logloss: 0.00341983\n",
      "[426]\ttraining's binary_logloss: 0.00146523\tvalid_1's binary_logloss: 0.00341129\n",
      "[427]\ttraining's binary_logloss: 0.00146258\tvalid_1's binary_logloss: 0.00340018\n",
      "[428]\ttraining's binary_logloss: 0.00145987\tvalid_1's binary_logloss: 0.00339794\n",
      "[429]\ttraining's binary_logloss: 0.00145678\tvalid_1's binary_logloss: 0.00339794\n",
      "[430]\ttraining's binary_logloss: 0.0014546\tvalid_1's binary_logloss: 0.0034012\n",
      "[431]\ttraining's binary_logloss: 0.00145239\tvalid_1's binary_logloss: 0.00340288\n",
      "[432]\ttraining's binary_logloss: 0.0014497\tvalid_1's binary_logloss: 0.00339234\n",
      "[433]\ttraining's binary_logloss: 0.00144731\tvalid_1's binary_logloss: 0.0033875\n",
      "[434]\ttraining's binary_logloss: 0.00144519\tvalid_1's binary_logloss: 0.00338405\n",
      "[435]\ttraining's binary_logloss: 0.00144304\tvalid_1's binary_logloss: 0.0033868\n",
      "[436]\ttraining's binary_logloss: 0.00144088\tvalid_1's binary_logloss: 0.00338512\n",
      "[437]\ttraining's binary_logloss: 0.00143873\tvalid_1's binary_logloss: 0.00339343\n",
      "[438]\ttraining's binary_logloss: 0.0014365\tvalid_1's binary_logloss: 0.0033936\n",
      "[439]\ttraining's binary_logloss: 0.00143452\tvalid_1's binary_logloss: 0.00339384\n",
      "[440]\ttraining's binary_logloss: 0.00143213\tvalid_1's binary_logloss: 0.00340068\n",
      "[441]\ttraining's binary_logloss: 0.00143013\tvalid_1's binary_logloss: 0.00340089\n",
      "[442]\ttraining's binary_logloss: 0.00142729\tvalid_1's binary_logloss: 0.00339491\n",
      "[443]\ttraining's binary_logloss: 0.00142526\tvalid_1's binary_logloss: 0.00340299\n",
      "[444]\ttraining's binary_logloss: 0.0014229\tvalid_1's binary_logloss: 0.00340268\n",
      "[445]\ttraining's binary_logloss: 0.00142076\tvalid_1's binary_logloss: 0.00339795\n",
      "[446]\ttraining's binary_logloss: 0.00141824\tvalid_1's binary_logloss: 0.00339598\n",
      "[447]\ttraining's binary_logloss: 0.00141633\tvalid_1's binary_logloss: 0.00339552\n",
      "[448]\ttraining's binary_logloss: 0.00141423\tvalid_1's binary_logloss: 0.00340077\n",
      "[449]\ttraining's binary_logloss: 0.00141187\tvalid_1's binary_logloss: 0.00339416\n",
      "[450]\ttraining's binary_logloss: 0.00141041\tvalid_1's binary_logloss: 0.00339682\n",
      "[451]\ttraining's binary_logloss: 0.00140756\tvalid_1's binary_logloss: 0.00337195\n",
      "[452]\ttraining's binary_logloss: 0.00140533\tvalid_1's binary_logloss: 0.00337047\n",
      "[453]\ttraining's binary_logloss: 0.00140318\tvalid_1's binary_logloss: 0.00337882\n",
      "[454]\ttraining's binary_logloss: 0.0014013\tvalid_1's binary_logloss: 0.00337943\n",
      "[455]\ttraining's binary_logloss: 0.00139947\tvalid_1's binary_logloss: 0.00338623\n",
      "[456]\ttraining's binary_logloss: 0.00139774\tvalid_1's binary_logloss: 0.00337835\n",
      "[457]\ttraining's binary_logloss: 0.00139573\tvalid_1's binary_logloss: 0.0033705\n",
      "[458]\ttraining's binary_logloss: 0.00139413\tvalid_1's binary_logloss: 0.00337402\n",
      "[459]\ttraining's binary_logloss: 0.00139163\tvalid_1's binary_logloss: 0.00337304\n",
      "[460]\ttraining's binary_logloss: 0.00138949\tvalid_1's binary_logloss: 0.00337525\n",
      "[461]\ttraining's binary_logloss: 0.00138799\tvalid_1's binary_logloss: 0.00337707\n",
      "[462]\ttraining's binary_logloss: 0.00138529\tvalid_1's binary_logloss: 0.00337367\n",
      "[463]\ttraining's binary_logloss: 0.00138331\tvalid_1's binary_logloss: 0.00337173\n",
      "[464]\ttraining's binary_logloss: 0.0013814\tvalid_1's binary_logloss: 0.00337019\n",
      "[465]\ttraining's binary_logloss: 0.00137881\tvalid_1's binary_logloss: 0.00336771\n",
      "[466]\ttraining's binary_logloss: 0.00137627\tvalid_1's binary_logloss: 0.00337589\n",
      "[467]\ttraining's binary_logloss: 0.00137362\tvalid_1's binary_logloss: 0.00337633\n",
      "[468]\ttraining's binary_logloss: 0.001372\tvalid_1's binary_logloss: 0.00337906\n",
      "[469]\ttraining's binary_logloss: 0.00137002\tvalid_1's binary_logloss: 0.00337177\n",
      "[470]\ttraining's binary_logloss: 0.00136868\tvalid_1's binary_logloss: 0.00337267\n",
      "[471]\ttraining's binary_logloss: 0.00136713\tvalid_1's binary_logloss: 0.00337573\n",
      "[472]\ttraining's binary_logloss: 0.00136565\tvalid_1's binary_logloss: 0.0033852\n",
      "[473]\ttraining's binary_logloss: 0.00136384\tvalid_1's binary_logloss: 0.00337653\n",
      "[474]\ttraining's binary_logloss: 0.00136119\tvalid_1's binary_logloss: 0.00337806\n",
      "[475]\ttraining's binary_logloss: 0.00135944\tvalid_1's binary_logloss: 0.00337877\n",
      "[476]\ttraining's binary_logloss: 0.00135758\tvalid_1's binary_logloss: 0.00337427\n",
      "[477]\ttraining's binary_logloss: 0.00135573\tvalid_1's binary_logloss: 0.00337026\n",
      "[478]\ttraining's binary_logloss: 0.00135381\tvalid_1's binary_logloss: 0.00336704\n",
      "[479]\ttraining's binary_logloss: 0.001352\tvalid_1's binary_logloss: 0.00336028\n",
      "[480]\ttraining's binary_logloss: 0.00135053\tvalid_1's binary_logloss: 0.00335832\n",
      "[481]\ttraining's binary_logloss: 0.00134907\tvalid_1's binary_logloss: 0.00336084\n",
      "[482]\ttraining's binary_logloss: 0.00134698\tvalid_1's binary_logloss: 0.00336064\n",
      "[483]\ttraining's binary_logloss: 0.00134522\tvalid_1's binary_logloss: 0.00334693\n",
      "[484]\ttraining's binary_logloss: 0.00134304\tvalid_1's binary_logloss: 0.00334739\n",
      "[485]\ttraining's binary_logloss: 0.00134103\tvalid_1's binary_logloss: 0.00334763\n",
      "[486]\ttraining's binary_logloss: 0.00133948\tvalid_1's binary_logloss: 0.00335079\n",
      "[487]\ttraining's binary_logloss: 0.0013377\tvalid_1's binary_logloss: 0.0033329\n",
      "[488]\ttraining's binary_logloss: 0.00133458\tvalid_1's binary_logloss: 0.00333556\n",
      "[489]\ttraining's binary_logloss: 0.00133264\tvalid_1's binary_logloss: 0.00334051\n",
      "[490]\ttraining's binary_logloss: 0.00133038\tvalid_1's binary_logloss: 0.00333544\n",
      "[491]\ttraining's binary_logloss: 0.00132881\tvalid_1's binary_logloss: 0.00335309\n",
      "[492]\ttraining's binary_logloss: 0.00132606\tvalid_1's binary_logloss: 0.00334228\n",
      "[493]\ttraining's binary_logloss: 0.00132444\tvalid_1's binary_logloss: 0.00335565\n",
      "[494]\ttraining's binary_logloss: 0.00132257\tvalid_1's binary_logloss: 0.00335637\n",
      "[495]\ttraining's binary_logloss: 0.00132026\tvalid_1's binary_logloss: 0.00336494\n",
      "[496]\ttraining's binary_logloss: 0.00131847\tvalid_1's binary_logloss: 0.00336493\n",
      "[497]\ttraining's binary_logloss: 0.0013171\tvalid_1's binary_logloss: 0.00336402\n",
      "[498]\ttraining's binary_logloss: 0.00131528\tvalid_1's binary_logloss: 0.00334708\n",
      "[499]\ttraining's binary_logloss: 0.00131358\tvalid_1's binary_logloss: 0.00335758\n",
      "[500]\ttraining's binary_logloss: 0.00131204\tvalid_1's binary_logloss: 0.0033504\n",
      "[501]\ttraining's binary_logloss: 0.00131061\tvalid_1's binary_logloss: 0.00334733\n",
      "[502]\ttraining's binary_logloss: 0.00130889\tvalid_1's binary_logloss: 0.00333841\n",
      "[503]\ttraining's binary_logloss: 0.0013067\tvalid_1's binary_logloss: 0.00333205\n",
      "[504]\ttraining's binary_logloss: 0.00130508\tvalid_1's binary_logloss: 0.0033241\n",
      "[505]\ttraining's binary_logloss: 0.00130344\tvalid_1's binary_logloss: 0.00332506\n",
      "[506]\ttraining's binary_logloss: 0.00130166\tvalid_1's binary_logloss: 0.00333574\n",
      "[507]\ttraining's binary_logloss: 0.00129996\tvalid_1's binary_logloss: 0.00334212\n",
      "[508]\ttraining's binary_logloss: 0.0012984\tvalid_1's binary_logloss: 0.00334002\n",
      "[509]\ttraining's binary_logloss: 0.00129712\tvalid_1's binary_logloss: 0.00334404\n",
      "[510]\ttraining's binary_logloss: 0.00129577\tvalid_1's binary_logloss: 0.00333864\n",
      "[511]\ttraining's binary_logloss: 0.00129382\tvalid_1's binary_logloss: 0.00334097\n",
      "[512]\ttraining's binary_logloss: 0.00129177\tvalid_1's binary_logloss: 0.00333773\n",
      "[513]\ttraining's binary_logloss: 0.00129006\tvalid_1's binary_logloss: 0.0033368\n",
      "[514]\ttraining's binary_logloss: 0.00128867\tvalid_1's binary_logloss: 0.00334083\n",
      "[515]\ttraining's binary_logloss: 0.00128675\tvalid_1's binary_logloss: 0.00334027\n",
      "[516]\ttraining's binary_logloss: 0.00128499\tvalid_1's binary_logloss: 0.00333871\n",
      "[517]\ttraining's binary_logloss: 0.00128328\tvalid_1's binary_logloss: 0.00334208\n",
      "[518]\ttraining's binary_logloss: 0.00128213\tvalid_1's binary_logloss: 0.00334431\n",
      "[519]\ttraining's binary_logloss: 0.00128067\tvalid_1's binary_logloss: 0.00334206\n",
      "[520]\ttraining's binary_logloss: 0.00127885\tvalid_1's binary_logloss: 0.00335064\n",
      "[521]\ttraining's binary_logloss: 0.0012772\tvalid_1's binary_logloss: 0.00334909\n",
      "[522]\ttraining's binary_logloss: 0.00127572\tvalid_1's binary_logloss: 0.00335183\n",
      "[523]\ttraining's binary_logloss: 0.00127416\tvalid_1's binary_logloss: 0.00334919\n",
      "[524]\ttraining's binary_logloss: 0.00127279\tvalid_1's binary_logloss: 0.00332994\n",
      "[525]\ttraining's binary_logloss: 0.00127043\tvalid_1's binary_logloss: 0.00333235\n",
      "[526]\ttraining's binary_logloss: 0.00126888\tvalid_1's binary_logloss: 0.00333076\n",
      "[527]\ttraining's binary_logloss: 0.00126748\tvalid_1's binary_logloss: 0.00333249\n",
      "[528]\ttraining's binary_logloss: 0.00126552\tvalid_1's binary_logloss: 0.00333835\n",
      "[529]\ttraining's binary_logloss: 0.00126444\tvalid_1's binary_logloss: 0.00333437\n",
      "[530]\ttraining's binary_logloss: 0.00126314\tvalid_1's binary_logloss: 0.00334755\n",
      "[531]\ttraining's binary_logloss: 0.0012622\tvalid_1's binary_logloss: 0.00334642\n",
      "[532]\ttraining's binary_logloss: 0.00126084\tvalid_1's binary_logloss: 0.00334065\n",
      "[533]\ttraining's binary_logloss: 0.00125919\tvalid_1's binary_logloss: 0.00335175\n",
      "[534]\ttraining's binary_logloss: 0.00125769\tvalid_1's binary_logloss: 0.00335007\n",
      "[535]\ttraining's binary_logloss: 0.00125593\tvalid_1's binary_logloss: 0.00335332\n",
      "[536]\ttraining's binary_logloss: 0.00125406\tvalid_1's binary_logloss: 0.00335992\n",
      "[537]\ttraining's binary_logloss: 0.00125252\tvalid_1's binary_logloss: 0.00336692\n",
      "[538]\ttraining's binary_logloss: 0.0012511\tvalid_1's binary_logloss: 0.00336463\n",
      "[539]\ttraining's binary_logloss: 0.00124933\tvalid_1's binary_logloss: 0.00336365\n",
      "[540]\ttraining's binary_logloss: 0.00124831\tvalid_1's binary_logloss: 0.00336532\n",
      "[541]\ttraining's binary_logloss: 0.00124718\tvalid_1's binary_logloss: 0.00337186\n",
      "[542]\ttraining's binary_logloss: 0.00124599\tvalid_1's binary_logloss: 0.0033644\n",
      "[543]\ttraining's binary_logloss: 0.00124433\tvalid_1's binary_logloss: 0.00336624\n",
      "[544]\ttraining's binary_logloss: 0.00124213\tvalid_1's binary_logloss: 0.00336614\n",
      "[545]\ttraining's binary_logloss: 0.00124048\tvalid_1's binary_logloss: 0.00335798\n",
      "[546]\ttraining's binary_logloss: 0.00123906\tvalid_1's binary_logloss: 0.00335291\n",
      "[547]\ttraining's binary_logloss: 0.00123754\tvalid_1's binary_logloss: 0.00335654\n",
      "[548]\ttraining's binary_logloss: 0.00123619\tvalid_1's binary_logloss: 0.00334683\n",
      "[549]\ttraining's binary_logloss: 0.00123515\tvalid_1's binary_logloss: 0.00335003\n",
      "[550]\ttraining's binary_logloss: 0.0012338\tvalid_1's binary_logloss: 0.00335057\n",
      "[551]\ttraining's binary_logloss: 0.00123222\tvalid_1's binary_logloss: 0.00334852\n",
      "[552]\ttraining's binary_logloss: 0.00123053\tvalid_1's binary_logloss: 0.00334918\n",
      "[553]\ttraining's binary_logloss: 0.00122923\tvalid_1's binary_logloss: 0.00334784\n",
      "[554]\ttraining's binary_logloss: 0.00122694\tvalid_1's binary_logloss: 0.00335384\n",
      "[555]\ttraining's binary_logloss: 0.00122554\tvalid_1's binary_logloss: 0.00334832\n",
      "[556]\ttraining's binary_logloss: 0.00122405\tvalid_1's binary_logloss: 0.00334107\n",
      "[557]\ttraining's binary_logloss: 0.00122269\tvalid_1's binary_logloss: 0.00333858\n",
      "[558]\ttraining's binary_logloss: 0.00122126\tvalid_1's binary_logloss: 0.0033394\n",
      "[559]\ttraining's binary_logloss: 0.00122036\tvalid_1's binary_logloss: 0.0033345\n",
      "[560]\ttraining's binary_logloss: 0.00121866\tvalid_1's binary_logloss: 0.00333733\n",
      "[561]\ttraining's binary_logloss: 0.00121768\tvalid_1's binary_logloss: 0.00333914\n",
      "[562]\ttraining's binary_logloss: 0.00121653\tvalid_1's binary_logloss: 0.00334001\n",
      "[563]\ttraining's binary_logloss: 0.00121507\tvalid_1's binary_logloss: 0.0033353\n",
      "[564]\ttraining's binary_logloss: 0.00121339\tvalid_1's binary_logloss: 0.00334369\n",
      "[565]\ttraining's binary_logloss: 0.00121252\tvalid_1's binary_logloss: 0.00333952\n",
      "[566]\ttraining's binary_logloss: 0.00121123\tvalid_1's binary_logloss: 0.00334449\n",
      "[567]\ttraining's binary_logloss: 0.00120987\tvalid_1's binary_logloss: 0.00333502\n",
      "[568]\ttraining's binary_logloss: 0.00120853\tvalid_1's binary_logloss: 0.00331851\n",
      "[569]\ttraining's binary_logloss: 0.00120726\tvalid_1's binary_logloss: 0.00332605\n",
      "[570]\ttraining's binary_logloss: 0.00120619\tvalid_1's binary_logloss: 0.00332131\n",
      "[571]\ttraining's binary_logloss: 0.0012045\tvalid_1's binary_logloss: 0.00332337\n",
      "[572]\ttraining's binary_logloss: 0.00120319\tvalid_1's binary_logloss: 0.00331999\n",
      "[573]\ttraining's binary_logloss: 0.00120173\tvalid_1's binary_logloss: 0.00331565\n",
      "[574]\ttraining's binary_logloss: 0.00119994\tvalid_1's binary_logloss: 0.00331455\n",
      "[575]\ttraining's binary_logloss: 0.00119892\tvalid_1's binary_logloss: 0.00331154\n",
      "[576]\ttraining's binary_logloss: 0.0011976\tvalid_1's binary_logloss: 0.00331124\n",
      "[577]\ttraining's binary_logloss: 0.00119618\tvalid_1's binary_logloss: 0.00328711\n",
      "[578]\ttraining's binary_logloss: 0.0011944\tvalid_1's binary_logloss: 0.00328332\n",
      "[579]\ttraining's binary_logloss: 0.00119309\tvalid_1's binary_logloss: 0.00328234\n",
      "[580]\ttraining's binary_logloss: 0.00119172\tvalid_1's binary_logloss: 0.00328218\n",
      "[581]\ttraining's binary_logloss: 0.00119054\tvalid_1's binary_logloss: 0.00328328\n",
      "[582]\ttraining's binary_logloss: 0.00118928\tvalid_1's binary_logloss: 0.00328715\n",
      "[583]\ttraining's binary_logloss: 0.00118788\tvalid_1's binary_logloss: 0.00328753\n",
      "[584]\ttraining's binary_logloss: 0.00118568\tvalid_1's binary_logloss: 0.00328567\n",
      "[585]\ttraining's binary_logloss: 0.00118458\tvalid_1's binary_logloss: 0.00328982\n",
      "[586]\ttraining's binary_logloss: 0.0011829\tvalid_1's binary_logloss: 0.00329692\n",
      "[587]\ttraining's binary_logloss: 0.00118174\tvalid_1's binary_logloss: 0.00329558\n",
      "[588]\ttraining's binary_logloss: 0.00118063\tvalid_1's binary_logloss: 0.00329737\n",
      "[589]\ttraining's binary_logloss: 0.00117933\tvalid_1's binary_logloss: 0.00328797\n",
      "[590]\ttraining's binary_logloss: 0.00117745\tvalid_1's binary_logloss: 0.00328818\n",
      "[591]\ttraining's binary_logloss: 0.00117622\tvalid_1's binary_logloss: 0.00328756\n",
      "[592]\ttraining's binary_logloss: 0.00117498\tvalid_1's binary_logloss: 0.00328536\n",
      "[593]\ttraining's binary_logloss: 0.00117356\tvalid_1's binary_logloss: 0.0032872\n",
      "[594]\ttraining's binary_logloss: 0.00117237\tvalid_1's binary_logloss: 0.00328413\n",
      "[595]\ttraining's binary_logloss: 0.00117118\tvalid_1's binary_logloss: 0.00327901\n",
      "[596]\ttraining's binary_logloss: 0.00116996\tvalid_1's binary_logloss: 0.00328064\n",
      "[597]\ttraining's binary_logloss: 0.00116864\tvalid_1's binary_logloss: 0.0032683\n",
      "[598]\ttraining's binary_logloss: 0.00116647\tvalid_1's binary_logloss: 0.00327212\n",
      "[599]\ttraining's binary_logloss: 0.00116517\tvalid_1's binary_logloss: 0.00327768\n",
      "[600]\ttraining's binary_logloss: 0.00116393\tvalid_1's binary_logloss: 0.00327762\n",
      "[601]\ttraining's binary_logloss: 0.00116257\tvalid_1's binary_logloss: 0.00327366\n",
      "[602]\ttraining's binary_logloss: 0.00116154\tvalid_1's binary_logloss: 0.00327014\n",
      "[603]\ttraining's binary_logloss: 0.00116049\tvalid_1's binary_logloss: 0.00328294\n",
      "[604]\ttraining's binary_logloss: 0.00115924\tvalid_1's binary_logloss: 0.00328814\n",
      "[605]\ttraining's binary_logloss: 0.00115753\tvalid_1's binary_logloss: 0.00329396\n",
      "[606]\ttraining's binary_logloss: 0.0011566\tvalid_1's binary_logloss: 0.00329542\n",
      "[607]\ttraining's binary_logloss: 0.00115569\tvalid_1's binary_logloss: 0.00328565\n",
      "[608]\ttraining's binary_logloss: 0.00115458\tvalid_1's binary_logloss: 0.00328567\n",
      "[609]\ttraining's binary_logloss: 0.00115347\tvalid_1's binary_logloss: 0.00328307\n",
      "[610]\ttraining's binary_logloss: 0.00115266\tvalid_1's binary_logloss: 0.00328352\n",
      "[611]\ttraining's binary_logloss: 0.00115136\tvalid_1's binary_logloss: 0.00327726\n",
      "[612]\ttraining's binary_logloss: 0.0011505\tvalid_1's binary_logloss: 0.00327881\n",
      "[613]\ttraining's binary_logloss: 0.00114939\tvalid_1's binary_logloss: 0.00327218\n",
      "[614]\ttraining's binary_logloss: 0.00114813\tvalid_1's binary_logloss: 0.00327058\n",
      "[615]\ttraining's binary_logloss: 0.00114684\tvalid_1's binary_logloss: 0.0032618\n",
      "[616]\ttraining's binary_logloss: 0.00114594\tvalid_1's binary_logloss: 0.00327156\n",
      "[617]\ttraining's binary_logloss: 0.00114486\tvalid_1's binary_logloss: 0.00327363\n",
      "[618]\ttraining's binary_logloss: 0.00114324\tvalid_1's binary_logloss: 0.00327382\n",
      "[619]\ttraining's binary_logloss: 0.00114213\tvalid_1's binary_logloss: 0.00327842\n",
      "[620]\ttraining's binary_logloss: 0.00114098\tvalid_1's binary_logloss: 0.00328563\n",
      "[621]\ttraining's binary_logloss: 0.00113988\tvalid_1's binary_logloss: 0.00328968\n",
      "[622]\ttraining's binary_logloss: 0.00113904\tvalid_1's binary_logloss: 0.0032912\n",
      "[623]\ttraining's binary_logloss: 0.00113774\tvalid_1's binary_logloss: 0.00329064\n",
      "[624]\ttraining's binary_logloss: 0.00113641\tvalid_1's binary_logloss: 0.00328907\n",
      "[625]\ttraining's binary_logloss: 0.00113536\tvalid_1's binary_logloss: 0.00329529\n",
      "[626]\ttraining's binary_logloss: 0.00113429\tvalid_1's binary_logloss: 0.00328793\n",
      "[627]\ttraining's binary_logloss: 0.00113299\tvalid_1's binary_logloss: 0.0032839\n",
      "[628]\ttraining's binary_logloss: 0.00113144\tvalid_1's binary_logloss: 0.00328575\n",
      "[629]\ttraining's binary_logloss: 0.00113042\tvalid_1's binary_logloss: 0.00328022\n",
      "[630]\ttraining's binary_logloss: 0.00112889\tvalid_1's binary_logloss: 0.00326253\n",
      "[631]\ttraining's binary_logloss: 0.00112773\tvalid_1's binary_logloss: 0.00325756\n",
      "[632]\ttraining's binary_logloss: 0.00112686\tvalid_1's binary_logloss: 0.00325259\n",
      "[633]\ttraining's binary_logloss: 0.00112574\tvalid_1's binary_logloss: 0.00325595\n",
      "[634]\ttraining's binary_logloss: 0.00112489\tvalid_1's binary_logloss: 0.00325638\n",
      "[635]\ttraining's binary_logloss: 0.00112417\tvalid_1's binary_logloss: 0.00325556\n",
      "[636]\ttraining's binary_logloss: 0.00112308\tvalid_1's binary_logloss: 0.00325442\n",
      "[637]\ttraining's binary_logloss: 0.00112187\tvalid_1's binary_logloss: 0.00325289\n",
      "[638]\ttraining's binary_logloss: 0.00112102\tvalid_1's binary_logloss: 0.00325075\n",
      "[639]\ttraining's binary_logloss: 0.00111997\tvalid_1's binary_logloss: 0.00325349\n",
      "[640]\ttraining's binary_logloss: 0.00111906\tvalid_1's binary_logloss: 0.00325722\n",
      "[641]\ttraining's binary_logloss: 0.0011179\tvalid_1's binary_logloss: 0.00325535\n",
      "[642]\ttraining's binary_logloss: 0.00111678\tvalid_1's binary_logloss: 0.00325554\n",
      "[643]\ttraining's binary_logloss: 0.00111612\tvalid_1's binary_logloss: 0.00325119\n",
      "[644]\ttraining's binary_logloss: 0.00111555\tvalid_1's binary_logloss: 0.00325295\n",
      "[645]\ttraining's binary_logloss: 0.00111473\tvalid_1's binary_logloss: 0.00325006\n",
      "[646]\ttraining's binary_logloss: 0.00111354\tvalid_1's binary_logloss: 0.00324722\n",
      "[647]\ttraining's binary_logloss: 0.00111251\tvalid_1's binary_logloss: 0.00324056\n",
      "[648]\ttraining's binary_logloss: 0.00111099\tvalid_1's binary_logloss: 0.00323816\n",
      "[649]\ttraining's binary_logloss: 0.00110963\tvalid_1's binary_logloss: 0.00323621\n",
      "[650]\ttraining's binary_logloss: 0.0011085\tvalid_1's binary_logloss: 0.00322724\n",
      "[651]\ttraining's binary_logloss: 0.00110728\tvalid_1's binary_logloss: 0.00323616\n",
      "[652]\ttraining's binary_logloss: 0.00110634\tvalid_1's binary_logloss: 0.00324237\n",
      "[653]\ttraining's binary_logloss: 0.00110518\tvalid_1's binary_logloss: 0.00323538\n",
      "[654]\ttraining's binary_logloss: 0.0011043\tvalid_1's binary_logloss: 0.00323573\n",
      "[655]\ttraining's binary_logloss: 0.00110338\tvalid_1's binary_logloss: 0.00324287\n",
      "[656]\ttraining's binary_logloss: 0.00110266\tvalid_1's binary_logloss: 0.00323609\n",
      "[657]\ttraining's binary_logloss: 0.00110157\tvalid_1's binary_logloss: 0.00322951\n",
      "[658]\ttraining's binary_logloss: 0.00110089\tvalid_1's binary_logloss: 0.00323616\n",
      "[659]\ttraining's binary_logloss: 0.00109975\tvalid_1's binary_logloss: 0.00323588\n",
      "[660]\ttraining's binary_logloss: 0.00109871\tvalid_1's binary_logloss: 0.00323061\n",
      "[661]\ttraining's binary_logloss: 0.00109789\tvalid_1's binary_logloss: 0.00323707\n",
      "[662]\ttraining's binary_logloss: 0.0010971\tvalid_1's binary_logloss: 0.00323863\n",
      "[663]\ttraining's binary_logloss: 0.00109598\tvalid_1's binary_logloss: 0.00323675\n",
      "[664]\ttraining's binary_logloss: 0.0010947\tvalid_1's binary_logloss: 0.00324648\n",
      "[665]\ttraining's binary_logloss: 0.00109364\tvalid_1's binary_logloss: 0.00324538\n",
      "[666]\ttraining's binary_logloss: 0.00109276\tvalid_1's binary_logloss: 0.00325159\n",
      "[667]\ttraining's binary_logloss: 0.00109175\tvalid_1's binary_logloss: 0.00325491\n",
      "[668]\ttraining's binary_logloss: 0.00109087\tvalid_1's binary_logloss: 0.00325023\n",
      "[669]\ttraining's binary_logloss: 0.00109\tvalid_1's binary_logloss: 0.00324407\n",
      "[670]\ttraining's binary_logloss: 0.0010891\tvalid_1's binary_logloss: 0.00324918\n",
      "[671]\ttraining's binary_logloss: 0.00108805\tvalid_1's binary_logloss: 0.00324655\n",
      "[672]\ttraining's binary_logloss: 0.00108717\tvalid_1's binary_logloss: 0.00324755\n",
      "[673]\ttraining's binary_logloss: 0.00108603\tvalid_1's binary_logloss: 0.00324482\n",
      "[674]\ttraining's binary_logloss: 0.00108505\tvalid_1's binary_logloss: 0.00324828\n",
      "[675]\ttraining's binary_logloss: 0.00108373\tvalid_1's binary_logloss: 0.00325086\n",
      "[676]\ttraining's binary_logloss: 0.00108279\tvalid_1's binary_logloss: 0.00325259\n",
      "[677]\ttraining's binary_logloss: 0.00108199\tvalid_1's binary_logloss: 0.00324417\n",
      "[678]\ttraining's binary_logloss: 0.00108083\tvalid_1's binary_logloss: 0.00324169\n",
      "[679]\ttraining's binary_logloss: 0.00107994\tvalid_1's binary_logloss: 0.00324586\n",
      "[680]\ttraining's binary_logloss: 0.00107898\tvalid_1's binary_logloss: 0.00324906\n",
      "[681]\ttraining's binary_logloss: 0.00107826\tvalid_1's binary_logloss: 0.00325259\n",
      "[682]\ttraining's binary_logloss: 0.00107642\tvalid_1's binary_logloss: 0.00325942\n",
      "[683]\ttraining's binary_logloss: 0.00107524\tvalid_1's binary_logloss: 0.00326164\n",
      "[684]\ttraining's binary_logloss: 0.00107435\tvalid_1's binary_logloss: 0.00326467\n",
      "[685]\ttraining's binary_logloss: 0.00107351\tvalid_1's binary_logloss: 0.00325955\n",
      "[686]\ttraining's binary_logloss: 0.00107233\tvalid_1's binary_logloss: 0.00325553\n",
      "[687]\ttraining's binary_logloss: 0.00107143\tvalid_1's binary_logloss: 0.00325818\n",
      "[688]\ttraining's binary_logloss: 0.0010704\tvalid_1's binary_logloss: 0.00324935\n",
      "[689]\ttraining's binary_logloss: 0.0010694\tvalid_1's binary_logloss: 0.003253\n",
      "[690]\ttraining's binary_logloss: 0.00106885\tvalid_1's binary_logloss: 0.00324723\n",
      "[691]\ttraining's binary_logloss: 0.00106784\tvalid_1's binary_logloss: 0.00324789\n",
      "[692]\ttraining's binary_logloss: 0.00106681\tvalid_1's binary_logloss: 0.00324913\n",
      "[693]\ttraining's binary_logloss: 0.00106567\tvalid_1's binary_logloss: 0.00324586\n",
      "[694]\ttraining's binary_logloss: 0.00106495\tvalid_1's binary_logloss: 0.00324155\n",
      "[695]\ttraining's binary_logloss: 0.00106422\tvalid_1's binary_logloss: 0.00323805\n",
      "[696]\ttraining's binary_logloss: 0.00106317\tvalid_1's binary_logloss: 0.00323255\n",
      "[697]\ttraining's binary_logloss: 0.00106232\tvalid_1's binary_logloss: 0.00323304\n",
      "[698]\ttraining's binary_logloss: 0.00106113\tvalid_1's binary_logloss: 0.00323947\n",
      "[699]\ttraining's binary_logloss: 0.00106035\tvalid_1's binary_logloss: 0.00324045\n",
      "[700]\ttraining's binary_logloss: 0.00105934\tvalid_1's binary_logloss: 0.00323981\n",
      "[701]\ttraining's binary_logloss: 0.00105842\tvalid_1's binary_logloss: 0.00324022\n",
      "[702]\ttraining's binary_logloss: 0.00105764\tvalid_1's binary_logloss: 0.0032459\n",
      "[703]\ttraining's binary_logloss: 0.00105666\tvalid_1's binary_logloss: 0.0032524\n",
      "[704]\ttraining's binary_logloss: 0.00105584\tvalid_1's binary_logloss: 0.00324137\n",
      "[705]\ttraining's binary_logloss: 0.0010551\tvalid_1's binary_logloss: 0.00324258\n",
      "[706]\ttraining's binary_logloss: 0.00105422\tvalid_1's binary_logloss: 0.00322652\n",
      "[707]\ttraining's binary_logloss: 0.00105321\tvalid_1's binary_logloss: 0.00322881\n",
      "[708]\ttraining's binary_logloss: 0.00105256\tvalid_1's binary_logloss: 0.00322691\n",
      "[709]\ttraining's binary_logloss: 0.00105188\tvalid_1's binary_logloss: 0.00322085\n",
      "[710]\ttraining's binary_logloss: 0.00105095\tvalid_1's binary_logloss: 0.0032233\n",
      "[711]\ttraining's binary_logloss: 0.00104972\tvalid_1's binary_logloss: 0.00321901\n",
      "[712]\ttraining's binary_logloss: 0.00104909\tvalid_1's binary_logloss: 0.00322347\n",
      "[713]\ttraining's binary_logloss: 0.00104812\tvalid_1's binary_logloss: 0.00322559\n",
      "[714]\ttraining's binary_logloss: 0.00104731\tvalid_1's binary_logloss: 0.00323969\n",
      "[715]\ttraining's binary_logloss: 0.00104639\tvalid_1's binary_logloss: 0.00324043\n",
      "[716]\ttraining's binary_logloss: 0.00104543\tvalid_1's binary_logloss: 0.00323942\n",
      "[717]\ttraining's binary_logloss: 0.00104475\tvalid_1's binary_logloss: 0.00323765\n",
      "[718]\ttraining's binary_logloss: 0.00104375\tvalid_1's binary_logloss: 0.00323586\n",
      "[719]\ttraining's binary_logloss: 0.00104287\tvalid_1's binary_logloss: 0.00323888\n",
      "[720]\ttraining's binary_logloss: 0.00104179\tvalid_1's binary_logloss: 0.00324251\n",
      "[721]\ttraining's binary_logloss: 0.00104097\tvalid_1's binary_logloss: 0.00325983\n",
      "[722]\ttraining's binary_logloss: 0.00104009\tvalid_1's binary_logloss: 0.00325803\n",
      "[723]\ttraining's binary_logloss: 0.00103932\tvalid_1's binary_logloss: 0.00326551\n",
      "[724]\ttraining's binary_logloss: 0.00103829\tvalid_1's binary_logloss: 0.00326496\n",
      "[725]\ttraining's binary_logloss: 0.00103723\tvalid_1's binary_logloss: 0.00326312\n",
      "[726]\ttraining's binary_logloss: 0.00103643\tvalid_1's binary_logloss: 0.00326103\n",
      "[727]\ttraining's binary_logloss: 0.00103528\tvalid_1's binary_logloss: 0.00326207\n",
      "[728]\ttraining's binary_logloss: 0.00103452\tvalid_1's binary_logloss: 0.00326609\n",
      "[729]\ttraining's binary_logloss: 0.00103382\tvalid_1's binary_logloss: 0.00326565\n",
      "[730]\ttraining's binary_logloss: 0.00103295\tvalid_1's binary_logloss: 0.00326452\n",
      "[731]\ttraining's binary_logloss: 0.00103151\tvalid_1's binary_logloss: 0.00326058\n",
      "[732]\ttraining's binary_logloss: 0.00103062\tvalid_1's binary_logloss: 0.00325257\n",
      "[733]\ttraining's binary_logloss: 0.00102998\tvalid_1's binary_logloss: 0.00325434\n",
      "[734]\ttraining's binary_logloss: 0.00102907\tvalid_1's binary_logloss: 0.00325067\n",
      "[735]\ttraining's binary_logloss: 0.00102822\tvalid_1's binary_logloss: 0.00325443\n",
      "[736]\ttraining's binary_logloss: 0.00102759\tvalid_1's binary_logloss: 0.00325904\n",
      "[737]\ttraining's binary_logloss: 0.00102683\tvalid_1's binary_logloss: 0.00326219\n",
      "[738]\ttraining's binary_logloss: 0.00102605\tvalid_1's binary_logloss: 0.00326739\n",
      "[739]\ttraining's binary_logloss: 0.00102532\tvalid_1's binary_logloss: 0.00327135\n",
      "[740]\ttraining's binary_logloss: 0.00102471\tvalid_1's binary_logloss: 0.00327286\n",
      "[741]\ttraining's binary_logloss: 0.00102393\tvalid_1's binary_logloss: 0.00327452\n",
      "[742]\ttraining's binary_logloss: 0.00102306\tvalid_1's binary_logloss: 0.00326818\n",
      "[743]\ttraining's binary_logloss: 0.00102206\tvalid_1's binary_logloss: 0.003276\n",
      "[744]\ttraining's binary_logloss: 0.00102126\tvalid_1's binary_logloss: 0.00328725\n",
      "[745]\ttraining's binary_logloss: 0.00102069\tvalid_1's binary_logloss: 0.00329515\n",
      "[746]\ttraining's binary_logloss: 0.00101984\tvalid_1's binary_logloss: 0.00328856\n",
      "[747]\ttraining's binary_logloss: 0.00101891\tvalid_1's binary_logloss: 0.00329182\n",
      "[748]\ttraining's binary_logloss: 0.00101833\tvalid_1's binary_logloss: 0.00329188\n",
      "[749]\ttraining's binary_logloss: 0.00101773\tvalid_1's binary_logloss: 0.00328981\n",
      "[750]\ttraining's binary_logloss: 0.00101682\tvalid_1's binary_logloss: 0.00329033\n",
      "[751]\ttraining's binary_logloss: 0.00101587\tvalid_1's binary_logloss: 0.00328993\n",
      "[752]\ttraining's binary_logloss: 0.00101499\tvalid_1's binary_logloss: 0.00328042\n",
      "[753]\ttraining's binary_logloss: 0.00101452\tvalid_1's binary_logloss: 0.00327611\n",
      "[754]\ttraining's binary_logloss: 0.0010136\tvalid_1's binary_logloss: 0.00328202\n",
      "[755]\ttraining's binary_logloss: 0.00101282\tvalid_1's binary_logloss: 0.00328134\n",
      "[756]\ttraining's binary_logloss: 0.00101185\tvalid_1's binary_logloss: 0.00328603\n",
      "[757]\ttraining's binary_logloss: 0.00101108\tvalid_1's binary_logloss: 0.00328735\n",
      "[758]\ttraining's binary_logloss: 0.00101004\tvalid_1's binary_logloss: 0.00329068\n",
      "[759]\ttraining's binary_logloss: 0.00100944\tvalid_1's binary_logloss: 0.0032875\n",
      "[760]\ttraining's binary_logloss: 0.00100863\tvalid_1's binary_logloss: 0.00328806\n",
      "[761]\ttraining's binary_logloss: 0.0010078\tvalid_1's binary_logloss: 0.0032786\n",
      "[762]\ttraining's binary_logloss: 0.00100722\tvalid_1's binary_logloss: 0.00327479\n",
      "[763]\ttraining's binary_logloss: 0.00100672\tvalid_1's binary_logloss: 0.00327902\n",
      "[764]\ttraining's binary_logloss: 0.00100594\tvalid_1's binary_logloss: 0.00327549\n",
      "[765]\ttraining's binary_logloss: 0.00100523\tvalid_1's binary_logloss: 0.00327768\n",
      "[766]\ttraining's binary_logloss: 0.0010047\tvalid_1's binary_logloss: 0.00327602\n",
      "[767]\ttraining's binary_logloss: 0.00100401\tvalid_1's binary_logloss: 0.00327568\n",
      "[768]\ttraining's binary_logloss: 0.00100327\tvalid_1's binary_logloss: 0.00326684\n",
      "[769]\ttraining's binary_logloss: 0.00100252\tvalid_1's binary_logloss: 0.00327015\n",
      "[770]\ttraining's binary_logloss: 0.00100177\tvalid_1's binary_logloss: 0.00326958\n",
      "[771]\ttraining's binary_logloss: 0.00100085\tvalid_1's binary_logloss: 0.00326627\n",
      "[772]\ttraining's binary_logloss: 0.0010002\tvalid_1's binary_logloss: 0.00326462\n",
      "[773]\ttraining's binary_logloss: 0.000999583\tvalid_1's binary_logloss: 0.00326543\n",
      "[774]\ttraining's binary_logloss: 0.000998797\tvalid_1's binary_logloss: 0.00326496\n",
      "[775]\ttraining's binary_logloss: 0.000998128\tvalid_1's binary_logloss: 0.00326683\n",
      "[776]\ttraining's binary_logloss: 0.000997243\tvalid_1's binary_logloss: 0.0032657\n",
      "[777]\ttraining's binary_logloss: 0.000996371\tvalid_1's binary_logloss: 0.0032616\n",
      "[778]\ttraining's binary_logloss: 0.000995848\tvalid_1's binary_logloss: 0.00326454\n",
      "[779]\ttraining's binary_logloss: 0.000995029\tvalid_1's binary_logloss: 0.00327396\n",
      "[780]\ttraining's binary_logloss: 0.000994284\tvalid_1's binary_logloss: 0.00326533\n",
      "[781]\ttraining's binary_logloss: 0.000993578\tvalid_1's binary_logloss: 0.0032603\n",
      "[782]\ttraining's binary_logloss: 0.000992971\tvalid_1's binary_logloss: 0.00326375\n",
      "[783]\ttraining's binary_logloss: 0.000992153\tvalid_1's binary_logloss: 0.00326982\n",
      "[784]\ttraining's binary_logloss: 0.000991389\tvalid_1's binary_logloss: 0.00327171\n",
      "[785]\ttraining's binary_logloss: 0.000990594\tvalid_1's binary_logloss: 0.00326731\n",
      "[786]\ttraining's binary_logloss: 0.000989914\tvalid_1's binary_logloss: 0.00326127\n",
      "[787]\ttraining's binary_logloss: 0.000989274\tvalid_1's binary_logloss: 0.00326503\n",
      "[788]\ttraining's binary_logloss: 0.000988338\tvalid_1's binary_logloss: 0.0032627\n",
      "[789]\ttraining's binary_logloss: 0.000987238\tvalid_1's binary_logloss: 0.00326471\n",
      "[790]\ttraining's binary_logloss: 0.000986455\tvalid_1's binary_logloss: 0.00326602\n",
      "[791]\ttraining's binary_logloss: 0.000985688\tvalid_1's binary_logloss: 0.00326428\n",
      "[792]\ttraining's binary_logloss: 0.000984931\tvalid_1's binary_logloss: 0.00326862\n",
      "[793]\ttraining's binary_logloss: 0.000984343\tvalid_1's binary_logloss: 0.00327492\n",
      "[794]\ttraining's binary_logloss: 0.000983481\tvalid_1's binary_logloss: 0.00327453\n",
      "[795]\ttraining's binary_logloss: 0.000982845\tvalid_1's binary_logloss: 0.00327307\n",
      "[796]\ttraining's binary_logloss: 0.0009821\tvalid_1's binary_logloss: 0.00327298\n",
      "[797]\ttraining's binary_logloss: 0.000981504\tvalid_1's binary_logloss: 0.00328486\n",
      "[798]\ttraining's binary_logloss: 0.000980689\tvalid_1's binary_logloss: 0.00328148\n",
      "[799]\ttraining's binary_logloss: 0.000980107\tvalid_1's binary_logloss: 0.00327811\n",
      "[800]\ttraining's binary_logloss: 0.000979419\tvalid_1's binary_logloss: 0.00327891\n",
      "[801]\ttraining's binary_logloss: 0.000978885\tvalid_1's binary_logloss: 0.00328079\n",
      "[802]\ttraining's binary_logloss: 0.000977917\tvalid_1's binary_logloss: 0.00328261\n",
      "[803]\ttraining's binary_logloss: 0.000977344\tvalid_1's binary_logloss: 0.00328708\n",
      "[804]\ttraining's binary_logloss: 0.000976803\tvalid_1's binary_logloss: 0.003285\n",
      "[805]\ttraining's binary_logloss: 0.000976214\tvalid_1's binary_logloss: 0.00328558\n",
      "[806]\ttraining's binary_logloss: 0.000975616\tvalid_1's binary_logloss: 0.00328896\n",
      "[807]\ttraining's binary_logloss: 0.00097487\tvalid_1's binary_logloss: 0.00329245\n",
      "[808]\ttraining's binary_logloss: 0.000974317\tvalid_1's binary_logloss: 0.00329734\n",
      "[809]\ttraining's binary_logloss: 0.000973534\tvalid_1's binary_logloss: 0.00329456\n",
      "[810]\ttraining's binary_logloss: 0.000972795\tvalid_1's binary_logloss: 0.00329109\n",
      "[811]\ttraining's binary_logloss: 0.000972099\tvalid_1's binary_logloss: 0.00329341\n",
      "Early stopping, best iteration is:\n",
      "[711]\ttraining's binary_logloss: 0.00104972\tvalid_1's binary_logloss: 0.00321901\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "LGBMClassifier(learning_rate=0.2, n_estimators=2500, objective='binary',\n",
       "               reg_alpha=1.2, reg_lambda=1.4, subsample=0.75, subsample_freq=1)"
      ]
     },
     "execution_count": 171,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lgb = LGBMClassifier(learning_rate = 0.2,\n",
    "                    n_estimators = 2500,\n",
    "                    objective = 'binary',\n",
    "                    subsample=0.75,\n",
    "                    subsample_freq = 1,\n",
    "                    reg_alpha = 1.2,\n",
    "                    reg_lambda = 1.4)\n",
    "lgb.fit(X_train, Y_train,\n",
    "        early_stopping_rounds = 100,\n",
    "        eval_set = [(X_train, Y_train), (X_val, Y_val)])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Feature Importance Check"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 172,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7efe186f55f8>"
      ]
     },
     "execution_count": 172,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x1440 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(10,20))\n",
    "plot_importance(lgb,ax=ax)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9996187571483035\n",
      "0.998638529611981 0.9966032608695652 0.9976198571914315\n"
     ]
    }
   ],
   "source": [
    "y_pred = lgb.predict(X_test)\n",
    "print(accuracy_score(Y_test, y_pred))\n",
    "precision, recall, fbeta_score, _ = precision_recall_fscore_support(Y_test, y_pred, average = 'binary')\n",
    "print(precision, recall, fbeta_score)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}