{ "cells": [ { "cell_type": "code", "execution_count": 1, "metadata": { "collapsed": true }, "outputs": [], "source": [ "import numpy as np\n", "import pandas as pd\n", "import xgboost as xgb\n", "import lightgbm as lgb\n", "import matplotlib.pyplot as plt\n", "import seaborn\n", "from sklearn import preprocessing\n", "import time\n", "%matplotlib inline" ] }, { "cell_type": "code", "execution_count": 2, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "(61878, 95)\n" ] } ], "source": [ "train = pd.read_csv('input/otto_train.csv')\n", "print(train.shape)" ] }, { "cell_type": "code", "execution_count": 3, "metadata": { "collapsed": true }, "outputs": [], "source": [ "def encode_features(dat):\n", " df = pd.DataFrame(index=dat.index.values)\n", " for c in dat.columns.values:\n", " unq = np.unique(dat[c])\n", " arr = np.zeros(len(df))\n", " for ii, u in enumerate(unq):\n", " flg = (dat[c] == u).values\n", " arr[flg] = ii\n", " df[c] = arr.astype(int)\n", " return df" ] }, { "cell_type": "code", "execution_count": 4, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "(61878, 93) (61878,)\n" ] } ], "source": [ "x = encode_features(train.drop(['id', 'target'], axis=1))\n", "y = np.array([int(v.split('_')[1])-1 for v in train.target])\n", "print(x.shape, y.shape)" ] }, { "cell_type": "code", "execution_count": 5, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "9\n" ] } ], "source": [ "num_cls = len(np.unique(y))\n", "print(num_cls)" ] }, { "cell_type": "code", "execution_count": 6, "metadata": { "collapsed": true }, "outputs": [], "source": [ "prm_xgb = {\n", " 'booster': 'gbtree',\n", " 'objective': 'multi:softprob',\n", " 'num_class': num_cls,\n", " 'max_depth': 5,\n", " 'learning_rate': 0.1,\n", " 'colsample_bytree': 0.9,\n", " 'subsample': 0.9,\n", " 'eval_metric': 'mlogloss',\n", "}\n", "prm_lgb = {\n", " 'boosting_type': 'gbdt',\n", " 'objective': 'multiclass',\n", " 'num_class': num_cls,\n", " 'num_leaves' : 2**5-1,\n", " 'learning_rate': 0.1,\n", " 'feature_fraction': 0.9,\n", " 'bagging_fraction': 0.9,\n", " 'bagging_freq' : 1,\n", " 'metric': 'multi_logloss',\n", "}\n", "num_round = 100" ] }, { "cell_type": "code", "execution_count": 7, "metadata": { "collapsed": true }, "outputs": [], "source": [ "np.random.seed(20161218)\n", "flg_train = np.random.choice([False, True], len(y), p=[0.3, 0.7])\n", "flg_valid = np.logical_not(flg_train)" ] }, { "cell_type": "code", "execution_count": 8, "metadata": { "collapsed": true }, "outputs": [], "source": [ "dt_xgb = xgb.DMatrix(x[flg_train], y[flg_train])\n", "dv_xgb = xgb.DMatrix(x[flg_valid], y[flg_valid])\n", "dt_lgb = lgb.Dataset(x[flg_train], y[flg_train])\n", "dv_lgb = lgb.Dataset(x[flg_valid], y[flg_valid], reference=dt_lgb)\n", "dt_lgb_c = lgb.Dataset(x[flg_train], y[flg_train], free_raw_data=False)\n", "dv_lgb_c = lgb.Dataset(x[flg_valid], y[flg_valid], free_raw_data=False,\n", " reference=dt_lgb)" ] }, { "cell_type": "code", "execution_count": 9, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "[0]\ttrain-mlogloss:1.97769\tvalid-mlogloss:1.98167\n", "[1]\ttrain-mlogloss:1.81571\tvalid-mlogloss:1.82305\n", "[2]\ttrain-mlogloss:1.68693\tvalid-mlogloss:1.69735\n", "[3]\ttrain-mlogloss:1.58784\tvalid-mlogloss:1.60077\n", "[4]\ttrain-mlogloss:1.49629\tvalid-mlogloss:1.51075\n", "[5]\ttrain-mlogloss:1.41958\tvalid-mlogloss:1.43592\n", "[6]\ttrain-mlogloss:1.35004\tvalid-mlogloss:1.368\n", "[7]\ttrain-mlogloss:1.28977\tvalid-mlogloss:1.30914\n", "[8]\ttrain-mlogloss:1.23794\tvalid-mlogloss:1.25887\n", "[9]\ttrain-mlogloss:1.18878\tvalid-mlogloss:1.21078\n", "[10]\ttrain-mlogloss:1.14482\tvalid-mlogloss:1.16806\n", "[11]\ttrain-mlogloss:1.10467\tvalid-mlogloss:1.12878\n", "[12]\ttrain-mlogloss:1.06779\tvalid-mlogloss:1.09323\n", "[13]\ttrain-mlogloss:1.03423\tvalid-mlogloss:1.06093\n", "[14]\ttrain-mlogloss:1.00338\tvalid-mlogloss:1.03145\n", "[15]\ttrain-mlogloss:0.975446\tvalid-mlogloss:1.00447\n", "[16]\ttrain-mlogloss:0.950178\tvalid-mlogloss:0.980001\n", "[17]\ttrain-mlogloss:0.927154\tvalid-mlogloss:0.957823\n", "[18]\ttrain-mlogloss:0.904816\tvalid-mlogloss:0.936438\n", "[19]\ttrain-mlogloss:0.88535\tvalid-mlogloss:0.917764\n", "[20]\ttrain-mlogloss:0.866943\tvalid-mlogloss:0.900236\n", "[21]\ttrain-mlogloss:0.848689\tvalid-mlogloss:0.882548\n", "[22]\ttrain-mlogloss:0.831932\tvalid-mlogloss:0.866694\n", "[23]\ttrain-mlogloss:0.817331\tvalid-mlogloss:0.852802\n", "[24]\ttrain-mlogloss:0.802325\tvalid-mlogloss:0.838711\n", "[25]\ttrain-mlogloss:0.788303\tvalid-mlogloss:0.82546\n", "[26]\ttrain-mlogloss:0.774821\tvalid-mlogloss:0.812747\n", "[27]\ttrain-mlogloss:0.762089\tvalid-mlogloss:0.800748\n", "[28]\ttrain-mlogloss:0.750976\tvalid-mlogloss:0.790482\n", "[29]\ttrain-mlogloss:0.739906\tvalid-mlogloss:0.780105\n", "[30]\ttrain-mlogloss:0.72947\tvalid-mlogloss:0.770532\n", "[31]\ttrain-mlogloss:0.719581\tvalid-mlogloss:0.761236\n", "[32]\ttrain-mlogloss:0.710496\tvalid-mlogloss:0.752896\n", "[33]\ttrain-mlogloss:0.701718\tvalid-mlogloss:0.744823\n", "[34]\ttrain-mlogloss:0.693449\tvalid-mlogloss:0.737215\n", "[35]\ttrain-mlogloss:0.686012\tvalid-mlogloss:0.730414\n", "[36]\ttrain-mlogloss:0.6786\tvalid-mlogloss:0.723613\n", "[37]\ttrain-mlogloss:0.671537\tvalid-mlogloss:0.717258\n", "[38]\ttrain-mlogloss:0.664872\tvalid-mlogloss:0.711308\n", "[39]\ttrain-mlogloss:0.65827\tvalid-mlogloss:0.70538\n", "[40]\ttrain-mlogloss:0.652491\tvalid-mlogloss:0.700172\n", "[41]\ttrain-mlogloss:0.646292\tvalid-mlogloss:0.694649\n", "[42]\ttrain-mlogloss:0.640946\tvalid-mlogloss:0.68997\n", "[43]\ttrain-mlogloss:0.635377\tvalid-mlogloss:0.685187\n", "[44]\ttrain-mlogloss:0.629765\tvalid-mlogloss:0.680131\n", "[45]\ttrain-mlogloss:0.624747\tvalid-mlogloss:0.675822\n", "[46]\ttrain-mlogloss:0.620059\tvalid-mlogloss:0.671835\n", "[47]\ttrain-mlogloss:0.615438\tvalid-mlogloss:0.667818\n", "[48]\ttrain-mlogloss:0.61111\tvalid-mlogloss:0.664124\n", "[49]\ttrain-mlogloss:0.606787\tvalid-mlogloss:0.660404\n", "[50]\ttrain-mlogloss:0.602591\tvalid-mlogloss:0.65673\n", "[51]\ttrain-mlogloss:0.598151\tvalid-mlogloss:0.65303\n", "[52]\ttrain-mlogloss:0.594195\tvalid-mlogloss:0.64961\n", "[53]\ttrain-mlogloss:0.590399\tvalid-mlogloss:0.646296\n", "[54]\ttrain-mlogloss:0.586551\tvalid-mlogloss:0.64311\n", "[55]\ttrain-mlogloss:0.583053\tvalid-mlogloss:0.640342\n", "[56]\ttrain-mlogloss:0.579582\tvalid-mlogloss:0.637335\n", "[57]\ttrain-mlogloss:0.576264\tvalid-mlogloss:0.634617\n", "[58]\ttrain-mlogloss:0.573013\tvalid-mlogloss:0.631998\n", "[59]\ttrain-mlogloss:0.569984\tvalid-mlogloss:0.629461\n", "[60]\ttrain-mlogloss:0.567038\tvalid-mlogloss:0.627098\n", "[61]\ttrain-mlogloss:0.56438\tvalid-mlogloss:0.624942\n", "[62]\ttrain-mlogloss:0.561487\tvalid-mlogloss:0.622638\n", "[63]\ttrain-mlogloss:0.558694\tvalid-mlogloss:0.620354\n", "[64]\ttrain-mlogloss:0.55589\tvalid-mlogloss:0.618263\n", "[65]\ttrain-mlogloss:0.553144\tvalid-mlogloss:0.616094\n", "[66]\ttrain-mlogloss:0.550565\tvalid-mlogloss:0.614126\n", "[67]\ttrain-mlogloss:0.54817\tvalid-mlogloss:0.612366\n", "[68]\ttrain-mlogloss:0.545968\tvalid-mlogloss:0.610506\n", "[69]\ttrain-mlogloss:0.543221\tvalid-mlogloss:0.608496\n", "[70]\ttrain-mlogloss:0.541326\tvalid-mlogloss:0.606942\n", "[71]\ttrain-mlogloss:0.539303\tvalid-mlogloss:0.605424\n", "[72]\ttrain-mlogloss:0.536813\tvalid-mlogloss:0.603583\n", "[73]\ttrain-mlogloss:0.534301\tvalid-mlogloss:0.601599\n", "[74]\ttrain-mlogloss:0.532282\tvalid-mlogloss:0.600154\n", "[75]\ttrain-mlogloss:0.530209\tvalid-mlogloss:0.598514\n", "[76]\ttrain-mlogloss:0.528042\tvalid-mlogloss:0.596927\n", "[77]\ttrain-mlogloss:0.526417\tvalid-mlogloss:0.595635\n", "[78]\ttrain-mlogloss:0.524277\tvalid-mlogloss:0.594087\n", "[79]\ttrain-mlogloss:0.522304\tvalid-mlogloss:0.592697\n", "[80]\ttrain-mlogloss:0.520434\tvalid-mlogloss:0.591274\n", "[81]\ttrain-mlogloss:0.518533\tvalid-mlogloss:0.589946\n", "[82]\ttrain-mlogloss:0.516775\tvalid-mlogloss:0.588603\n", "[83]\ttrain-mlogloss:0.514891\tvalid-mlogloss:0.587333\n", "[84]\ttrain-mlogloss:0.513171\tvalid-mlogloss:0.586085\n", "[85]\ttrain-mlogloss:0.511336\tvalid-mlogloss:0.584773\n", "[86]\ttrain-mlogloss:0.509759\tvalid-mlogloss:0.583657\n", "[87]\ttrain-mlogloss:0.508198\tvalid-mlogloss:0.582465\n", "[88]\ttrain-mlogloss:0.506574\tvalid-mlogloss:0.581387\n", "[89]\ttrain-mlogloss:0.50486\tvalid-mlogloss:0.580269\n", "[90]\ttrain-mlogloss:0.503117\tvalid-mlogloss:0.579108\n", "[91]\ttrain-mlogloss:0.501639\tvalid-mlogloss:0.578292\n", "[92]\ttrain-mlogloss:0.500083\tvalid-mlogloss:0.577266\n", "[93]\ttrain-mlogloss:0.498588\tvalid-mlogloss:0.576308\n", "[94]\ttrain-mlogloss:0.496938\tvalid-mlogloss:0.575252\n", "[95]\ttrain-mlogloss:0.495597\tvalid-mlogloss:0.574365\n", "[96]\ttrain-mlogloss:0.493962\tvalid-mlogloss:0.573261\n", "[97]\ttrain-mlogloss:0.492771\tvalid-mlogloss:0.572487\n", "[98]\ttrain-mlogloss:0.490992\tvalid-mlogloss:0.571281\n", "[99]\ttrain-mlogloss:0.489677\tvalid-mlogloss:0.570366\n", "65.19904208183289\n" ] } ], "source": [ "time_s = time.time()\n", "obj_xgb = xgb.train(\n", " prm_xgb, dt_xgb, num_round,\n", " [(dt_xgb, 'train'), (dv_xgb, 'valid')])\n", "time_t = time.time()\n", "print(time_t - time_s)" ] }, { "cell_type": "code", "execution_count": 10, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "[1]\tvalid_0's multi_logloss:1.96589\n", "[2]\tvalid_0's multi_logloss:1.80069\n", "[3]\tvalid_0's multi_logloss:1.67048\n", "[4]\tvalid_0's multi_logloss:1.56283\n", "[5]\tvalid_0's multi_logloss:1.47489\n", "[6]\tvalid_0's multi_logloss:1.39621\n", "[7]\tvalid_0's multi_logloss:1.32853\n", "[8]\tvalid_0's multi_logloss:1.26781\n", "[9]\tvalid_0's multi_logloss:1.21342\n", "[10]\tvalid_0's multi_logloss:1.16519\n", "[11]\tvalid_0's multi_logloss:1.12163\n", "[12]\tvalid_0's multi_logloss:1.08238\n", "[13]\tvalid_0's multi_logloss:1.0463\n", "[14]\tvalid_0's multi_logloss:1.01403\n", "[15]\tvalid_0's multi_logloss:0.984063\n", "[16]\tvalid_0's multi_logloss:0.956702\n", "[17]\tvalid_0's multi_logloss:0.931502\n", "[18]\tvalid_0's multi_logloss:0.909192\n", "[19]\tvalid_0's multi_logloss:0.887876\n", "[20]\tvalid_0's multi_logloss:0.868728\n", "[21]\tvalid_0's multi_logloss:0.850413\n", "[22]\tvalid_0's multi_logloss:0.833369\n", "[23]\tvalid_0's multi_logloss:0.817609\n", "[24]\tvalid_0's multi_logloss:0.802913\n", "[25]\tvalid_0's multi_logloss:0.789397\n", "[26]\tvalid_0's multi_logloss:0.776372\n", "[27]\tvalid_0's multi_logloss:0.76423\n", "[28]\tvalid_0's multi_logloss:0.752673\n", "[29]\tvalid_0's multi_logloss:0.741827\n", "[30]\tvalid_0's multi_logloss:0.731737\n", "[31]\tvalid_0's multi_logloss:0.722051\n", "[32]\tvalid_0's multi_logloss:0.713151\n", "[33]\tvalid_0's multi_logloss:0.704391\n", "[34]\tvalid_0's multi_logloss:0.696632\n", "[35]\tvalid_0's multi_logloss:0.688742\n", "[36]\tvalid_0's multi_logloss:0.681658\n", "[37]\tvalid_0's multi_logloss:0.675085\n", "[38]\tvalid_0's multi_logloss:0.668548\n", "[39]\tvalid_0's multi_logloss:0.66241\n", "[40]\tvalid_0's multi_logloss:0.656508\n", "[41]\tvalid_0's multi_logloss:0.650796\n", "[42]\tvalid_0's multi_logloss:0.64544\n", "[43]\tvalid_0's multi_logloss:0.640326\n", "[44]\tvalid_0's multi_logloss:0.635305\n", "[45]\tvalid_0's multi_logloss:0.630602\n", "[46]\tvalid_0's multi_logloss:0.626089\n", "[47]\tvalid_0's multi_logloss:0.621855\n", "[48]\tvalid_0's multi_logloss:0.617875\n", "[49]\tvalid_0's multi_logloss:0.613931\n", "[50]\tvalid_0's multi_logloss:0.610281\n", "[51]\tvalid_0's multi_logloss:0.606787\n", "[52]\tvalid_0's multi_logloss:0.603244\n", "[53]\tvalid_0's multi_logloss:0.60016\n", "[54]\tvalid_0's multi_logloss:0.596772\n", "[55]\tvalid_0's multi_logloss:0.593876\n", "[56]\tvalid_0's multi_logloss:0.590949\n", "[57]\tvalid_0's multi_logloss:0.587967\n", "[58]\tvalid_0's multi_logloss:0.585057\n", "[59]\tvalid_0's multi_logloss:0.582379\n", "[60]\tvalid_0's multi_logloss:0.579821\n", "[61]\tvalid_0's multi_logloss:0.577179\n", "[62]\tvalid_0's multi_logloss:0.574902\n", "[63]\tvalid_0's multi_logloss:0.572569\n", "[64]\tvalid_0's multi_logloss:0.570239\n", "[65]\tvalid_0's multi_logloss:0.568136\n", "[66]\tvalid_0's multi_logloss:0.566206\n", "[67]\tvalid_0's multi_logloss:0.564028\n", "[68]\tvalid_0's multi_logloss:0.562047\n", "[69]\tvalid_0's multi_logloss:0.560237\n", "[70]\tvalid_0's multi_logloss:0.558436\n", "[71]\tvalid_0's multi_logloss:0.556645\n", "[72]\tvalid_0's multi_logloss:0.554881\n", "[73]\tvalid_0's multi_logloss:0.553293\n", "[74]\tvalid_0's multi_logloss:0.551645\n", "[75]\tvalid_0's multi_logloss:0.550112\n", "[76]\tvalid_0's multi_logloss:0.54849\n", "[77]\tvalid_0's multi_logloss:0.546681\n", "[78]\tvalid_0's multi_logloss:0.54531\n", "[79]\tvalid_0's multi_logloss:0.543805\n", "[80]\tvalid_0's multi_logloss:0.542504\n", "[81]\tvalid_0's multi_logloss:0.541192\n", "[82]\tvalid_0's multi_logloss:0.539905\n", "[83]\tvalid_0's multi_logloss:0.538626\n", "[84]\tvalid_0's multi_logloss:0.537395\n", "[85]\tvalid_0's multi_logloss:0.536271\n", "[86]\tvalid_0's multi_logloss:0.5353\n", "[87]\tvalid_0's multi_logloss:0.534155\n", "[88]\tvalid_0's multi_logloss:0.53317\n", "[89]\tvalid_0's multi_logloss:0.532131\n", "[90]\tvalid_0's multi_logloss:0.531052\n", "[91]\tvalid_0's multi_logloss:0.530085\n", "[92]\tvalid_0's multi_logloss:0.529274\n", "[93]\tvalid_0's multi_logloss:0.528285\n", "[94]\tvalid_0's multi_logloss:0.527422\n", "[95]\tvalid_0's multi_logloss:0.526541\n", "[96]\tvalid_0's multi_logloss:0.525695\n", "[97]\tvalid_0's multi_logloss:0.524851\n", "[98]\tvalid_0's multi_logloss:0.523952\n", "[99]\tvalid_0's multi_logloss:0.523016\n", "[100]\tvalid_0's multi_logloss:0.522378\n", "16.02284002304077\n" ] } ], "source": [ "time_s = time.time()\n", "obj_lgb = lgb.train(\n", " prm_lgb, dt_lgb, num_boost_round=num_round,\n", " valid_sets=dv_lgb)\n", "time_t = time.time()\n", "print(time_t - time_s)\n", "obj_lgb.save_model('output/lgb.txt')" ] }, { "cell_type": "code", "execution_count": 11, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "[1]\tvalid_0's multi_logloss:2.01067\n", "[2]\tvalid_0's multi_logloss:1.87489\n", "[3]\tvalid_0's multi_logloss:1.76438\n", "[4]\tvalid_0's multi_logloss:1.67353\n", "[5]\tvalid_0's multi_logloss:1.59879\n", "[6]\tvalid_0's multi_logloss:1.53027\n", "[7]\tvalid_0's multi_logloss:1.47235\n", "[8]\tvalid_0's multi_logloss:1.41975\n", "[9]\tvalid_0's multi_logloss:1.37189\n", "[10]\tvalid_0's multi_logloss:1.32934\n", "[11]\tvalid_0's multi_logloss:1.29121\n", "[12]\tvalid_0's multi_logloss:1.25617\n", "[13]\tvalid_0's multi_logloss:1.22375\n", "[14]\tvalid_0's multi_logloss:1.19445\n", "[15]\tvalid_0's multi_logloss:1.1664\n", "[16]\tvalid_0's multi_logloss:1.14143\n", "[17]\tvalid_0's multi_logloss:1.11888\n", "[18]\tvalid_0's multi_logloss:1.09807\n", "[19]\tvalid_0's multi_logloss:1.07792\n", "[20]\tvalid_0's multi_logloss:1.05924\n", "[21]\tvalid_0's multi_logloss:1.04119\n", "[22]\tvalid_0's multi_logloss:1.02435\n", "[23]\tvalid_0's multi_logloss:1.00902\n", "[24]\tvalid_0's multi_logloss:0.994402\n", "[25]\tvalid_0's multi_logloss:0.981219\n", "[26]\tvalid_0's multi_logloss:0.967938\n", "[27]\tvalid_0's multi_logloss:0.955806\n", "[28]\tvalid_0's multi_logloss:0.943997\n", "[29]\tvalid_0's multi_logloss:0.93329\n", "[30]\tvalid_0's multi_logloss:0.92275\n", "[31]\tvalid_0's multi_logloss:0.912987\n", "[32]\tvalid_0's multi_logloss:0.903436\n", "[33]\tvalid_0's multi_logloss:0.893949\n", "[34]\tvalid_0's multi_logloss:0.886012\n", "[35]\tvalid_0's multi_logloss:0.878008\n", "[36]\tvalid_0's multi_logloss:0.869912\n", "[37]\tvalid_0's multi_logloss:0.862095\n", "[38]\tvalid_0's multi_logloss:0.855012\n", "[39]\tvalid_0's multi_logloss:0.847866\n", "[40]\tvalid_0's multi_logloss:0.840964\n", "[41]\tvalid_0's multi_logloss:0.834674\n", "[42]\tvalid_0's multi_logloss:0.828241\n", "[43]\tvalid_0's multi_logloss:0.822593\n", "[44]\tvalid_0's multi_logloss:0.816565\n", "[45]\tvalid_0's multi_logloss:0.811401\n", "[46]\tvalid_0's multi_logloss:0.805926\n", "[47]\tvalid_0's multi_logloss:0.801008\n", "[48]\tvalid_0's multi_logloss:0.796294\n", "[49]\tvalid_0's multi_logloss:0.791922\n", "[50]\tvalid_0's multi_logloss:0.786993\n", "[51]\tvalid_0's multi_logloss:0.782985\n", "[52]\tvalid_0's multi_logloss:0.778408\n", "[53]\tvalid_0's multi_logloss:0.774352\n", "[54]\tvalid_0's multi_logloss:0.770591\n", "[55]\tvalid_0's multi_logloss:0.766681\n", "[56]\tvalid_0's multi_logloss:0.762936\n", "[57]\tvalid_0's multi_logloss:0.759445\n", "[58]\tvalid_0's multi_logloss:0.755821\n", "[59]\tvalid_0's multi_logloss:0.752613\n", "[60]\tvalid_0's multi_logloss:0.749157\n", "[61]\tvalid_0's multi_logloss:0.745849\n", "[62]\tvalid_0's multi_logloss:0.742746\n", "[63]\tvalid_0's multi_logloss:0.739965\n", "[64]\tvalid_0's multi_logloss:0.737107\n", "[65]\tvalid_0's multi_logloss:0.734446\n", "[66]\tvalid_0's multi_logloss:0.731742\n", "[67]\tvalid_0's multi_logloss:0.729159\n", "[68]\tvalid_0's multi_logloss:0.726514\n", "[69]\tvalid_0's multi_logloss:0.723947\n", "[70]\tvalid_0's multi_logloss:0.721495\n", "[71]\tvalid_0's multi_logloss:0.719146\n", "[72]\tvalid_0's multi_logloss:0.716951\n", "[73]\tvalid_0's multi_logloss:0.714741\n", "[74]\tvalid_0's multi_logloss:0.712736\n", "[75]\tvalid_0's multi_logloss:0.710544\n", "[76]\tvalid_0's multi_logloss:0.708489\n", "[77]\tvalid_0's multi_logloss:0.706578\n", "[78]\tvalid_0's multi_logloss:0.704697\n", "[79]\tvalid_0's multi_logloss:0.702595\n", "[80]\tvalid_0's multi_logloss:0.700678\n", "[81]\tvalid_0's multi_logloss:0.698942\n", "[82]\tvalid_0's multi_logloss:0.697189\n", "[83]\tvalid_0's multi_logloss:0.695499\n", "[84]\tvalid_0's multi_logloss:0.693559\n", "[85]\tvalid_0's multi_logloss:0.692008\n", "[86]\tvalid_0's multi_logloss:0.690307\n", "[87]\tvalid_0's multi_logloss:0.688656\n", "[88]\tvalid_0's multi_logloss:0.68721\n", "[89]\tvalid_0's multi_logloss:0.685759\n", "[90]\tvalid_0's multi_logloss:0.684212\n", "[91]\tvalid_0's multi_logloss:0.682738\n", "[92]\tvalid_0's multi_logloss:0.681347\n", "[93]\tvalid_0's multi_logloss:0.680014\n", "[94]\tvalid_0's multi_logloss:0.678764\n", "[95]\tvalid_0's multi_logloss:0.677424\n", "[96]\tvalid_0's multi_logloss:0.676253\n", "[97]\tvalid_0's multi_logloss:0.67492\n", "[98]\tvalid_0's multi_logloss:0.673824\n", "[99]\tvalid_0's multi_logloss:0.672577\n", "[100]\tvalid_0's multi_logloss:0.671286\n", "15.776311874389648\n" ] } ], "source": [ "time_s = time.time()\n", "obj_lgb = lgb.train(\n", " prm_lgb, dt_lgb_c, num_boost_round=num_round,\n", " valid_sets=dv_lgb_c,\n", " categorical_feature=list(range(len(x.columns.values))))\n", "time_t = time.time()\n", "print(time_t - time_s)\n", "obj_lgb.save_model('output/lgb_cat.txt')" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "collapsed": true }, "outputs": [], "source": [] } ], "metadata": { "kernelspec": { "display_name": "Python 3", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.5.2" } }, "nbformat": 4, "nbformat_minor": 2 }