offline_training.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import warnings\n",
    "warnings.simplefilter('ignore')\n",
    "\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "pd.set_option('max_columns', None)\n",
    "\n",
    "from tqdm.notebook import tqdm\n",
    "\n",
    "import gc\n",
    "from collections import defaultdict\n",
    "\n",
    "from sklearn.metrics import roc_auc_score\n",
    "\n",
    "from gensim.models import Word2Vec\n",
    "\n",
    "from bitarray import bitarray\n",
    "\n",
    "import lightgbm as lgb\n",
    "\n",
    "%matplotlib inline\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_pickle = '/home/user01/kaggle/riiid/cv1_train.pickle'\n",
    "valid_pickle = '/home/user01/kaggle/riiid/cv1_valid.pickle'\n",
    "\n",
    "question_file = '/home/user01/kaggle/riiid/questions.csv'\n",
    "lecture_file = '/home/user01/kaggle/riiid/lectures.csv'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "!mkdir -pv ./upload_cache_new"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(96817414, 9) (2453886, 9)\n",
      "CPU times: user 5.64 s, sys: 5.37 s, total: 11 s\n",
      "Wall time: 11 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "# read data\n",
    "\n",
    "field_needed = ['row_id', 'user_id', 'content_id', \n",
    "                'content_type_id', 'task_container_id',\n",
    "                'answered_correctly', 'prior_question_elapsed_time', \n",
    "                'prior_question_had_explanation', 'timestamp']\n",
    "\n",
    "train = pd.read_pickle(train_pickle)[field_needed]\n",
    "valid = pd.read_pickle(valid_pickle)[field_needed]\n",
    "    \n",
    "train = train.loc[train.content_type_id == False].reset_index(drop=True)\n",
    "valid = valid.loc[valid.content_type_id == False].reset_index(drop=True)\n",
    "\n",
    "train['prior_question_had_explanation'].fillna(False, inplace=True)\n",
    "valid['prior_question_had_explanation'].fillna(False, inplace=True)\n",
    "train['prior_question_had_explanation'] = train['prior_question_had_explanation'].astype('int8')\n",
    "valid['prior_question_had_explanation'] = valid['prior_question_had_explanation'].astype('int8')\n",
    "\n",
    "print(train.shape, valid.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>row_id</th>\n",
       "      <th>user_id</th>\n",
       "      <th>content_id</th>\n",
       "      <th>content_type_id</th>\n",
       "      <th>task_container_id</th>\n",
       "      <th>answered_correctly</th>\n",
       "      <th>prior_question_elapsed_time</th>\n",
       "      <th>prior_question_had_explanation</th>\n",
       "      <th>timestamp</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>32933156</td>\n",
       "      <td>705741139</td>\n",
       "      <td>128</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>32933157</td>\n",
       "      <td>705741139</td>\n",
       "      <td>7860</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>16000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>20666</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>32933158</td>\n",
       "      <td>705741139</td>\n",
       "      <td>7922</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>19000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>39172</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>32933159</td>\n",
       "      <td>705741139</td>\n",
       "      <td>156</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>17000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>58207</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>32933160</td>\n",
       "      <td>705741139</td>\n",
       "      <td>51</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>17000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>75779</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     row_id    user_id  content_id  content_type_id  task_container_id  \\\n",
       "0  32933156  705741139         128                0                  0   \n",
       "1  32933157  705741139        7860                0                  1   \n",
       "2  32933158  705741139        7922                0                  2   \n",
       "3  32933159  705741139         156                0                  3   \n",
       "4  32933160  705741139          51                0                  4   \n",
       "\n",
       "   answered_correctly  prior_question_elapsed_time  \\\n",
       "0                   1                          NaN   \n",
       "1                   1                      16000.0   \n",
       "2                   1                      19000.0   \n",
       "3                   1                      17000.0   \n",
       "4                   1                      17000.0   \n",
       "\n",
       "   prior_question_had_explanation  timestamp  \n",
       "0                               0          0  \n",
       "1                               0      20666  \n",
       "2                               0      39172  \n",
       "3                               0      58207  \n",
       "4                               0      75779  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>row_id</th>\n",
       "      <th>user_id</th>\n",
       "      <th>content_id</th>\n",
       "      <th>content_type_id</th>\n",
       "      <th>task_container_id</th>\n",
       "      <th>answered_correctly</th>\n",
       "      <th>prior_question_elapsed_time</th>\n",
       "      <th>prior_question_had_explanation</th>\n",
       "      <th>timestamp</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>45253</td>\n",
       "      <td>1186307</td>\n",
       "      <td>4451</td>\n",
       "      <td>0</td>\n",
       "      <td>370</td>\n",
       "      <td>1</td>\n",
       "      <td>10000.0</td>\n",
       "      <td>1</td>\n",
       "      <td>2146476402</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>41038234</td>\n",
       "      <td>871981876</td>\n",
       "      <td>1278</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>26000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>87403</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>56086085</td>\n",
       "      <td>1189905495</td>\n",
       "      <td>5604</td>\n",
       "      <td>0</td>\n",
       "      <td>1991</td>\n",
       "      <td>0</td>\n",
       "      <td>3000.0</td>\n",
       "      <td>1</td>\n",
       "      <td>41201916980</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>91995999</td>\n",
       "      <td>1951897185</td>\n",
       "      <td>8534</td>\n",
       "      <td>0</td>\n",
       "      <td>504</td>\n",
       "      <td>0</td>\n",
       "      <td>47000.0</td>\n",
       "      <td>1</td>\n",
       "      <td>4916060626</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>99237056</td>\n",
       "      <td>2105538840</td>\n",
       "      <td>5631</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>21000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>33806</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     row_id     user_id  content_id  content_type_id  task_container_id  \\\n",
       "0     45253     1186307        4451                0                370   \n",
       "1  41038234   871981876        1278                0                  3   \n",
       "2  56086085  1189905495        5604                0               1991   \n",
       "3  91995999  1951897185        8534                0                504   \n",
       "4  99237056  2105538840        5631                0                  1   \n",
       "\n",
       "   answered_correctly  prior_question_elapsed_time  \\\n",
       "0                   1                      10000.0   \n",
       "1                   0                      26000.0   \n",
       "2                   0                       3000.0   \n",
       "3                   0                      47000.0   \n",
       "4                   1                      21000.0   \n",
       "\n",
       "   prior_question_had_explanation    timestamp  \n",
       "0                               1   2146476402  \n",
       "1                               0        87403  \n",
       "2                               1  41201916980  \n",
       "3                               1   4916060626  \n",
       "4                               0        33806  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "valid.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 4.42 s, sys: 1.47 s, total: 5.89 s\n",
      "Wall time: 5.89 s\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>content_id</th>\n",
       "      <th>answered_correctly_avg_c</th>\n",
       "      <th>answered_correctly_var_c</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0.907334</td>\n",
       "      <td>0.084092</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0.890571</td>\n",
       "      <td>0.097468</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>0.554213</td>\n",
       "      <td>0.247067</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>0.779386</td>\n",
       "      <td>0.171951</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>0.613970</td>\n",
       "      <td>0.237019</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   content_id  answered_correctly_avg_c  answered_correctly_var_c\n",
       "0           0                  0.907334                  0.084092\n",
       "1           1                  0.890571                  0.097468\n",
       "2           2                  0.554213                  0.247067\n",
       "3           3                  0.779386                  0.171951\n",
       "4           4                  0.613970                  0.237019"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "# answered correctly average for each content\n",
    "\n",
    "content_df = train[['content_id', 'answered_correctly']].groupby(['content_id']).agg(['mean', 'var']).reset_index()\n",
    "content_df.columns = ['content_id', 'answered_correctly_avg_c', 'answered_correctly_var_c']\n",
    "content_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(13523, 4)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>content_id</th>\n",
       "      <th>answered_correctly_avg_c</th>\n",
       "      <th>answered_correctly_var_c</th>\n",
       "      <th>lecture_part</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0.907334</td>\n",
       "      <td>0.084092</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0.890571</td>\n",
       "      <td>0.097468</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>0.554213</td>\n",
       "      <td>0.247067</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>0.779386</td>\n",
       "      <td>0.171951</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>0.613970</td>\n",
       "      <td>0.237019</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   content_id  answered_correctly_avg_c  answered_correctly_var_c  \\\n",
       "0           0                  0.907334                  0.084092   \n",
       "1           1                  0.890571                  0.097468   \n",
       "2           2                  0.554213                  0.247067   \n",
       "3           3                  0.779386                  0.171951   \n",
       "4           4                  0.613970                  0.237019   \n",
       "\n",
       "   lecture_part  \n",
       "0           NaN  \n",
       "1           NaN  \n",
       "2           NaN  \n",
       "3           NaN  \n",
       "4           NaN  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# lectures\n",
    "\n",
    "lecture_df = pd.read_csv(lecture_file)\n",
    "lecture_df.rename(columns={'lecture_id': 'content_id', 'part': 'lecture_part'}, inplace=True)\n",
    "content_df = content_df.merge(lecture_df[['content_id', 'lecture_part']], how='left')\n",
    "\n",
    "print(content_df.shape)\n",
    "content_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 21.7 s, sys: 5.66 s, total: 27.3 s\n",
      "Wall time: 27.3 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "train = pd.merge(train, content_df, on=['content_id'], how=\"left\")\n",
    "valid = pd.merge(valid, content_df, on=['content_id'], how=\"left\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>row_id</th>\n",
       "      <th>user_id</th>\n",
       "      <th>content_id</th>\n",
       "      <th>content_type_id</th>\n",
       "      <th>task_container_id</th>\n",
       "      <th>answered_correctly</th>\n",
       "      <th>prior_question_elapsed_time</th>\n",
       "      <th>prior_question_had_explanation</th>\n",
       "      <th>timestamp</th>\n",
       "      <th>answered_correctly_avg_c</th>\n",
       "      <th>answered_correctly_var_c</th>\n",
       "      <th>lecture_part</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>32933156</td>\n",
       "      <td>705741139</td>\n",
       "      <td>128</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.966887</td>\n",
       "      <td>0.032018</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>32933157</td>\n",
       "      <td>705741139</td>\n",
       "      <td>7860</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>16000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>20666</td>\n",
       "      <td>0.954720</td>\n",
       "      <td>0.043232</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>32933158</td>\n",
       "      <td>705741139</td>\n",
       "      <td>7922</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>19000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>39172</td>\n",
       "      <td>0.953161</td>\n",
       "      <td>0.044647</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>32933159</td>\n",
       "      <td>705741139</td>\n",
       "      <td>156</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>17000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>58207</td>\n",
       "      <td>0.932764</td>\n",
       "      <td>0.062719</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>32933160</td>\n",
       "      <td>705741139</td>\n",
       "      <td>51</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>17000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>75779</td>\n",
       "      <td>0.930161</td>\n",
       "      <td>0.064964</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     row_id    user_id  content_id  content_type_id  task_container_id  \\\n",
       "0  32933156  705741139         128                0                  0   \n",
       "1  32933157  705741139        7860                0                  1   \n",
       "2  32933158  705741139        7922                0                  2   \n",
       "3  32933159  705741139         156                0                  3   \n",
       "4  32933160  705741139          51                0                  4   \n",
       "\n",
       "   answered_correctly  prior_question_elapsed_time  \\\n",
       "0                   1                          NaN   \n",
       "1                   1                      16000.0   \n",
       "2                   1                      19000.0   \n",
       "3                   1                      17000.0   \n",
       "4                   1                      17000.0   \n",
       "\n",
       "   prior_question_had_explanation  timestamp  answered_correctly_avg_c  \\\n",
       "0                               0          0                  0.966887   \n",
       "1                               0      20666                  0.954720   \n",
       "2                               0      39172                  0.953161   \n",
       "3                               0      58207                  0.932764   \n",
       "4                               0      75779                  0.930161   \n",
       "\n",
       "   answered_correctly_var_c  lecture_part  \n",
       "0                  0.032018           NaN  \n",
       "1                  0.043232           NaN  \n",
       "2                  0.044647           NaN  \n",
       "3                  0.062719           NaN  \n",
       "4                  0.064964           NaN  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(13523, 18)\n",
      "CPU times: user 18.8 s, sys: 4.84 s, total: 23.7 s\n",
      "Wall time: 23.7 s\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>content_id</th>\n",
       "      <th>correct_answer</th>\n",
       "      <th>part</th>\n",
       "      <th>tags</th>\n",
       "      <th>mean_answered_correctly_content</th>\n",
       "      <th>question_asked</th>\n",
       "      <th>right_answers</th>\n",
       "      <th>bundle_size</th>\n",
       "      <th>bundle_rignt_answers</th>\n",
       "      <th>bundle_questions_asked</th>\n",
       "      <th>bundle_accuracy</th>\n",
       "      <th>part_rignt_answers</th>\n",
       "      <th>part_questions_asked</th>\n",
       "      <th>part_accuracy</th>\n",
       "      <th>LorR</th>\n",
       "      <th>tag1</th>\n",
       "      <th>tag1_answered_correctly_mean</th>\n",
       "      <th>tag1_answered_correctly_var</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>51 131 162 38</td>\n",
       "      <td>0.907334</td>\n",
       "      <td>6777</td>\n",
       "      <td>6149.0</td>\n",
       "      <td>1</td>\n",
       "      <td>6149.0</td>\n",
       "      <td>6777</td>\n",
       "      <td>0.907334</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>51</td>\n",
       "      <td>0.838379</td>\n",
       "      <td>0.017258</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>131 36 81</td>\n",
       "      <td>0.890571</td>\n",
       "      <td>7265</td>\n",
       "      <td>6470.0</td>\n",
       "      <td>1</td>\n",
       "      <td>6470.0</td>\n",
       "      <td>7265</td>\n",
       "      <td>0.890571</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>0.850586</td>\n",
       "      <td>0.017212</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>131 101 162 92</td>\n",
       "      <td>0.554213</td>\n",
       "      <td>43781</td>\n",
       "      <td>24264.0</td>\n",
       "      <td>1</td>\n",
       "      <td>24264.0</td>\n",
       "      <td>43781</td>\n",
       "      <td>0.554213</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>0.850586</td>\n",
       "      <td>0.017212</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>131 149 162 29</td>\n",
       "      <td>0.779386</td>\n",
       "      <td>22451</td>\n",
       "      <td>17498.0</td>\n",
       "      <td>1</td>\n",
       "      <td>17498.0</td>\n",
       "      <td>22451</td>\n",
       "      <td>0.779386</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>0.850586</td>\n",
       "      <td>0.017212</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>131 5 162 38</td>\n",
       "      <td>0.613970</td>\n",
       "      <td>30881</td>\n",
       "      <td>18960.0</td>\n",
       "      <td>1</td>\n",
       "      <td>18960.0</td>\n",
       "      <td>30881</td>\n",
       "      <td>0.613970</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>0.850586</td>\n",
       "      <td>0.017212</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   content_id  correct_answer  part            tags  \\\n",
       "0           0               0     1   51 131 162 38   \n",
       "1           1               1     1       131 36 81   \n",
       "2           2               0     1  131 101 162 92   \n",
       "3           3               0     1  131 149 162 29   \n",
       "4           4               3     1    131 5 162 38   \n",
       "\n",
       "   mean_answered_correctly_content  question_asked  right_answers  \\\n",
       "0                         0.907334            6777         6149.0   \n",
       "1                         0.890571            7265         6470.0   \n",
       "2                         0.554213           43781        24264.0   \n",
       "3                         0.779386           22451        17498.0   \n",
       "4                         0.613970           30881        18960.0   \n",
       "\n",
       "   bundle_size  bundle_rignt_answers  bundle_questions_asked  bundle_accuracy  \\\n",
       "0            1                6149.0                    6777         0.907334   \n",
       "1            1                6470.0                    7265         0.890571   \n",
       "2            1               24264.0                   43781         0.554213   \n",
       "3            1               17498.0                   22451         0.779386   \n",
       "4            1               18960.0                   30881         0.613970   \n",
       "\n",
       "   part_rignt_answers  part_questions_asked  part_accuracy  LorR  tag1  \\\n",
       "0           5425569.0               7276830       0.745595     0    51   \n",
       "1           5425569.0               7276830       0.745595     0   131   \n",
       "2           5425569.0               7276830       0.745595     0   131   \n",
       "3           5425569.0               7276830       0.745595     0   131   \n",
       "4           5425569.0               7276830       0.745595     0   131   \n",
       "\n",
       "   tag1_answered_correctly_mean  tag1_answered_correctly_var  \n",
       "0                      0.838379                     0.017258  \n",
       "1                      0.850586                     0.017212  \n",
       "2                      0.850586                     0.017212  \n",
       "3                      0.850586                     0.017212  \n",
       "4                      0.850586                     0.017212  "
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "# question\n",
    "\n",
    "def addFeatures(_train, _questions):\n",
    "    \n",
    "    train_questions_only_df = _train[_train['answered_correctly']!=-1]\n",
    "    \n",
    "    grouped_by_content_df = train_questions_only_df.groupby('content_id')\n",
    "    content_answers_df = grouped_by_content_df.agg({'answered_correctly': ['mean', 'count'] }).copy()\n",
    "    content_answers_df.columns = ['mean_answered_correctly_content', 'question_asked']\n",
    "    \n",
    "    questions_df = _questions.merge(content_answers_df, left_on = 'question_id', right_on = 'content_id', how = 'left')\n",
    "    bundle_dict = questions_df['bundle_id'].value_counts().to_dict()\n",
    "    questions_df['right_answers'] = questions_df['mean_answered_correctly_content'] * questions_df['question_asked']\n",
    "    questions_df['bundle_size'] = questions_df['bundle_id'].apply(lambda x: bundle_dict[x])\n",
    "\n",
    "    grouped_by_bundle_df = questions_df.groupby('bundle_id')\n",
    "    bundle_answers_df = grouped_by_bundle_df.agg({'right_answers': 'sum', 'question_asked': 'sum'}).copy()\n",
    "    bundle_answers_df.columns = ['bundle_rignt_answers', 'bundle_questions_asked']\n",
    "    bundle_answers_df['bundle_accuracy'] = bundle_answers_df['bundle_rignt_answers'] / bundle_answers_df['bundle_questions_asked']\n",
    "\n",
    "    grouped_by_part_df = questions_df.groupby('part')\n",
    "    part_answers_df = grouped_by_part_df.agg({'right_answers': 'sum', 'question_asked': 'sum'}).copy()\n",
    "    part_answers_df.columns = ['part_rignt_answers', 'part_questions_asked']\n",
    "    part_answers_df['part_accuracy'] = part_answers_df['part_rignt_answers'] / part_answers_df['part_questions_asked']\n",
    "\n",
    "    return questions_df, bundle_answers_df, part_answers_df\n",
    "\n",
    "\n",
    "questions_df = pd.read_csv(question_file)\n",
    "questions_df, bundle_answers_df, part_answers_df = addFeatures(train, questions_df)\n",
    "\n",
    "questions_df = pd.merge(questions_df,\n",
    "                        bundle_answers_df,\n",
    "                        on='bundle_id',\n",
    "                        how='left')\n",
    "\n",
    "questions_df = pd.merge(questions_df,\n",
    "                        part_answers_df,\n",
    "                        on='part',\n",
    "                        how='left')\n",
    "\n",
    "questions_df.rename(columns={'question_id': 'content_id'}, inplace=True)\n",
    "\n",
    "questions_df['LorR'] = questions_df['part'].apply(lambda x: 1 if x >=5 else 0)\n",
    "\n",
    "tag = questions_df[\"tags\"].str.split(\" \", n=10, expand=True)\n",
    "tag.columns = ['tag1', 'tag2', 'tag3', 'tag4', 'tag5', 'tag6']\n",
    "questions_df = pd.concat([questions_df, tag], axis=1)\n",
    "questions_df['tag1'] = pd.to_numeric(questions_df['tag1'], errors='coerce')\n",
    "questions_df['tag1'] = questions_df['tag1'].fillna(-1).astype(int)\n",
    "questions_df.drop(['tag2', 'tag3', 'tag4', 'tag5', 'tag6', 'bundle_id'], axis=1, inplace=True)\n",
    "\n",
    "tag1_agg = questions_df.groupby('tag1')['mean_answered_correctly_content'].agg(['mean', 'var'])\n",
    "questions_df['tag1_answered_correctly_mean'] = questions_df['tag1'].map(tag1_agg['mean'])\n",
    "questions_df['tag1_answered_correctly_var'] = questions_df['tag1'].map(tag1_agg['var'])\n",
    "questions_df['tag1_answered_correctly_mean'] = questions_df.tag1_answered_correctly_mean.astype('float16')\n",
    "questions_df['tag1_answered_correctly_var'] = questions_df.tag1_answered_correctly_var.astype('float16')\n",
    "\n",
    "questions_df.set_index('content_id')\n",
    "\n",
    "print(questions_df.shape)\n",
    "questions_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>content_id</th>\n",
       "      <th>correct_answer</th>\n",
       "      <th>part</th>\n",
       "      <th>mean_answered_correctly_content</th>\n",
       "      <th>question_asked</th>\n",
       "      <th>right_answers</th>\n",
       "      <th>bundle_size</th>\n",
       "      <th>bundle_rignt_answers</th>\n",
       "      <th>bundle_questions_asked</th>\n",
       "      <th>bundle_accuracy</th>\n",
       "      <th>part_rignt_answers</th>\n",
       "      <th>part_questions_asked</th>\n",
       "      <th>part_accuracy</th>\n",
       "      <th>LorR</th>\n",
       "      <th>tag1</th>\n",
       "      <th>tag1_answered_correctly_mean</th>\n",
       "      <th>tag1_answered_correctly_var</th>\n",
       "      <th>tags_emb_0</th>\n",
       "      <th>tags_emb_1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.907334</td>\n",
       "      <td>6777</td>\n",
       "      <td>6149.0</td>\n",
       "      <td>1</td>\n",
       "      <td>6149.0</td>\n",
       "      <td>6777</td>\n",
       "      <td>0.907334</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>51</td>\n",
       "      <td>0.838379</td>\n",
       "      <td>0.017258</td>\n",
       "      <td>0.179532</td>\n",
       "      <td>0.141397</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0.890571</td>\n",
       "      <td>7265</td>\n",
       "      <td>6470.0</td>\n",
       "      <td>1</td>\n",
       "      <td>6470.0</td>\n",
       "      <td>7265</td>\n",
       "      <td>0.890571</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>0.850586</td>\n",
       "      <td>0.017212</td>\n",
       "      <td>-0.196008</td>\n",
       "      <td>-0.232887</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.554213</td>\n",
       "      <td>43781</td>\n",
       "      <td>24264.0</td>\n",
       "      <td>1</td>\n",
       "      <td>24264.0</td>\n",
       "      <td>43781</td>\n",
       "      <td>0.554213</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>0.850586</td>\n",
       "      <td>0.017212</td>\n",
       "      <td>0.034004</td>\n",
       "      <td>-0.229731</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.779386</td>\n",
       "      <td>22451</td>\n",
       "      <td>17498.0</td>\n",
       "      <td>1</td>\n",
       "      <td>17498.0</td>\n",
       "      <td>22451</td>\n",
       "      <td>0.779386</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>0.850586</td>\n",
       "      <td>0.017212</td>\n",
       "      <td>0.152637</td>\n",
       "      <td>-0.025414</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0.613970</td>\n",
       "      <td>30881</td>\n",
       "      <td>18960.0</td>\n",
       "      <td>1</td>\n",
       "      <td>18960.0</td>\n",
       "      <td>30881</td>\n",
       "      <td>0.613970</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>0.850586</td>\n",
       "      <td>0.017212</td>\n",
       "      <td>0.153832</td>\n",
       "      <td>-0.154031</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   content_id  correct_answer  part  mean_answered_correctly_content  \\\n",
       "0           0               0     1                         0.907334   \n",
       "1           1               1     1                         0.890571   \n",
       "2           2               0     1                         0.554213   \n",
       "3           3               0     1                         0.779386   \n",
       "4           4               3     1                         0.613970   \n",
       "\n",
       "   question_asked  right_answers  bundle_size  bundle_rignt_answers  \\\n",
       "0            6777         6149.0            1                6149.0   \n",
       "1            7265         6470.0            1                6470.0   \n",
       "2           43781        24264.0            1               24264.0   \n",
       "3           22451        17498.0            1               17498.0   \n",
       "4           30881        18960.0            1               18960.0   \n",
       "\n",
       "   bundle_questions_asked  bundle_accuracy  part_rignt_answers  \\\n",
       "0                    6777         0.907334           5425569.0   \n",
       "1                    7265         0.890571           5425569.0   \n",
       "2                   43781         0.554213           5425569.0   \n",
       "3                   22451         0.779386           5425569.0   \n",
       "4                   30881         0.613970           5425569.0   \n",
       "\n",
       "   part_questions_asked  part_accuracy  LorR  tag1  \\\n",
       "0               7276830       0.745595     0    51   \n",
       "1               7276830       0.745595     0   131   \n",
       "2               7276830       0.745595     0   131   \n",
       "3               7276830       0.745595     0   131   \n",
       "4               7276830       0.745595     0   131   \n",
       "\n",
       "   tag1_answered_correctly_mean  tag1_answered_correctly_var  tags_emb_0  \\\n",
       "0                      0.838379                     0.017258    0.179532   \n",
       "1                      0.850586                     0.017212   -0.196008   \n",
       "2                      0.850586                     0.017212    0.034004   \n",
       "3                      0.850586                     0.017212    0.152637   \n",
       "4                      0.850586                     0.017212    0.153832   \n",
       "\n",
       "   tags_emb_1  \n",
       "0    0.141397  \n",
       "1   -0.232887  \n",
       "2   -0.229731  \n",
       "3   -0.025414  \n",
       "4   -0.154031  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# embedding 函数\n",
    "\n",
    "def emb(df, c1, c2, emb_size=2):\n",
    "    \n",
    "    tmp = df.groupby(c1, as_index=False)[c2].agg(\n",
    "        {'{}_{}_list'.format(c1, c2): list})\n",
    "    sentences = tmp['{}_{}_list'.format(c1, c2)].values.tolist()\n",
    "    del tmp['{}_{}_list'.format(c1, c2)]\n",
    "\n",
    "    words = []\n",
    "    for i in range(len(sentences)):\n",
    "        words += sentences[i]\n",
    "        sentences[i] = [str(x) for x in sentences[i]]\n",
    "    words = list(set(words))\n",
    "\n",
    "    model = Word2Vec(sentences, \n",
    "                     size=emb_size, \n",
    "                     window=3,\n",
    "                     min_count=1, \n",
    "                     sg=0, \n",
    "                     hs=1, \n",
    "                     seed=2020)\n",
    "\n",
    "    emb_matrix = []\n",
    "    for w in words:\n",
    "        w = str(w)\n",
    "        if w in model:\n",
    "            emb_matrix.append(model[w])\n",
    "        else:\n",
    "            emb_matrix.append([0] * emb_size)\n",
    "\n",
    "    df_emb = pd.DataFrame(emb_matrix)\n",
    "    df_emb.columns = ['{}_emb_{}'.format(\n",
    "        c2, i) for i in range(emb_size)]\n",
    "\n",
    "    df_emb[c2] = words\n",
    "\n",
    "    del model, emb_matrix, sentences\n",
    "    return df_emb\n",
    "\n",
    "\n",
    "df_emb = emb(questions_df, 'content_id', 'tags')\n",
    "questions_df = questions_df.merge(df_emb, on='tags', how='left')\n",
    "questions_df.drop(['tags'], axis=1, inplace=True)\n",
    "\n",
    "del df_emb\n",
    "gc.collect()\n",
    "\n",
    "questions_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 35.9 s, sys: 21.1 s, total: 56.9 s\n",
      "Wall time: 56.9 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "qcols = [col for col in questions_df.columns if col not in ['correct_answer', \n",
    "                                                            'mean_answered_correctly_content']]\n",
    "\n",
    "train = pd.merge(train, \n",
    "                 questions_df[qcols], \n",
    "                 on='content_id', \n",
    "                 how='left')\n",
    "\n",
    "valid = pd.merge(valid, \n",
    "                 questions_df[qcols], \n",
    "                 on='content_id',\n",
    "                 how='left')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>row_id</th>\n",
       "      <th>user_id</th>\n",
       "      <th>content_id</th>\n",
       "      <th>content_type_id</th>\n",
       "      <th>task_container_id</th>\n",
       "      <th>answered_correctly</th>\n",
       "      <th>prior_question_elapsed_time</th>\n",
       "      <th>prior_question_had_explanation</th>\n",
       "      <th>timestamp</th>\n",
       "      <th>answered_correctly_avg_c</th>\n",
       "      <th>answered_correctly_var_c</th>\n",
       "      <th>lecture_part</th>\n",
       "      <th>part</th>\n",
       "      <th>question_asked</th>\n",
       "      <th>right_answers</th>\n",
       "      <th>bundle_size</th>\n",
       "      <th>bundle_rignt_answers</th>\n",
       "      <th>bundle_questions_asked</th>\n",
       "      <th>bundle_accuracy</th>\n",
       "      <th>part_rignt_answers</th>\n",
       "      <th>part_questions_asked</th>\n",
       "      <th>part_accuracy</th>\n",
       "      <th>LorR</th>\n",
       "      <th>tag1</th>\n",
       "      <th>tag1_answered_correctly_mean</th>\n",
       "      <th>tag1_answered_correctly_var</th>\n",
       "      <th>tags_emb_0</th>\n",
       "      <th>tags_emb_1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>32933156</td>\n",
       "      <td>705741139</td>\n",
       "      <td>128</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.966887</td>\n",
       "      <td>0.032018</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>19056</td>\n",
       "      <td>18425.0</td>\n",
       "      <td>1</td>\n",
       "      <td>18425.0</td>\n",
       "      <td>19056</td>\n",
       "      <td>0.966887</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>0.850586</td>\n",
       "      <td>0.017212</td>\n",
       "      <td>-0.007187</td>\n",
       "      <td>0.090353</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>32933157</td>\n",
       "      <td>705741139</td>\n",
       "      <td>7860</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>16000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>20666</td>\n",
       "      <td>0.954720</td>\n",
       "      <td>0.043232</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>21135</td>\n",
       "      <td>20178.0</td>\n",
       "      <td>1</td>\n",
       "      <td>20178.0</td>\n",
       "      <td>21135</td>\n",
       "      <td>0.954720</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>0.850586</td>\n",
       "      <td>0.017212</td>\n",
       "      <td>-0.038487</td>\n",
       "      <td>-0.097879</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>32933158</td>\n",
       "      <td>705741139</td>\n",
       "      <td>7922</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>19000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>39172</td>\n",
       "      <td>0.953161</td>\n",
       "      <td>0.044647</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>18852</td>\n",
       "      <td>17969.0</td>\n",
       "      <td>1</td>\n",
       "      <td>17969.0</td>\n",
       "      <td>18852</td>\n",
       "      <td>0.953161</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>0.850586</td>\n",
       "      <td>0.017212</td>\n",
       "      <td>-0.007187</td>\n",
       "      <td>0.090353</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>32933159</td>\n",
       "      <td>705741139</td>\n",
       "      <td>156</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>17000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>58207</td>\n",
       "      <td>0.932764</td>\n",
       "      <td>0.062719</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>21878</td>\n",
       "      <td>20407.0</td>\n",
       "      <td>1</td>\n",
       "      <td>20407.0</td>\n",
       "      <td>21878</td>\n",
       "      <td>0.932764</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>0.850586</td>\n",
       "      <td>0.017212</td>\n",
       "      <td>0.106040</td>\n",
       "      <td>-0.152273</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>32933160</td>\n",
       "      <td>705741139</td>\n",
       "      <td>51</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>17000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>75779</td>\n",
       "      <td>0.930161</td>\n",
       "      <td>0.064964</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>23454</td>\n",
       "      <td>21816.0</td>\n",
       "      <td>1</td>\n",
       "      <td>21816.0</td>\n",
       "      <td>23454</td>\n",
       "      <td>0.930161</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>0.850586</td>\n",
       "      <td>0.017212</td>\n",
       "      <td>-0.220193</td>\n",
       "      <td>0.188346</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     row_id    user_id  content_id  content_type_id  task_container_id  \\\n",
       "0  32933156  705741139         128                0                  0   \n",
       "1  32933157  705741139        7860                0                  1   \n",
       "2  32933158  705741139        7922                0                  2   \n",
       "3  32933159  705741139         156                0                  3   \n",
       "4  32933160  705741139          51                0                  4   \n",
       "\n",
       "   answered_correctly  prior_question_elapsed_time  \\\n",
       "0                   1                          NaN   \n",
       "1                   1                      16000.0   \n",
       "2                   1                      19000.0   \n",
       "3                   1                      17000.0   \n",
       "4                   1                      17000.0   \n",
       "\n",
       "   prior_question_had_explanation  timestamp  answered_correctly_avg_c  \\\n",
       "0                               0          0                  0.966887   \n",
       "1                               0      20666                  0.954720   \n",
       "2                               0      39172                  0.953161   \n",
       "3                               0      58207                  0.932764   \n",
       "4                               0      75779                  0.930161   \n",
       "\n",
       "   answered_correctly_var_c  lecture_part  part  question_asked  \\\n",
       "0                  0.032018           NaN     1           19056   \n",
       "1                  0.043232           NaN     1           21135   \n",
       "2                  0.044647           NaN     1           18852   \n",
       "3                  0.062719           NaN     1           21878   \n",
       "4                  0.064964           NaN     1           23454   \n",
       "\n",
       "   right_answers  bundle_size  bundle_rignt_answers  bundle_questions_asked  \\\n",
       "0        18425.0            1               18425.0                   19056   \n",
       "1        20178.0            1               20178.0                   21135   \n",
       "2        17969.0            1               17969.0                   18852   \n",
       "3        20407.0            1               20407.0                   21878   \n",
       "4        21816.0            1               21816.0                   23454   \n",
       "\n",
       "   bundle_accuracy  part_rignt_answers  part_questions_asked  part_accuracy  \\\n",
       "0         0.966887           5425569.0               7276830       0.745595   \n",
       "1         0.954720           5425569.0               7276830       0.745595   \n",
       "2         0.953161           5425569.0               7276830       0.745595   \n",
       "3         0.932764           5425569.0               7276830       0.745595   \n",
       "4         0.930161           5425569.0               7276830       0.745595   \n",
       "\n",
       "   LorR  tag1  tag1_answered_correctly_mean  tag1_answered_correctly_var  \\\n",
       "0     0   131                      0.850586                     0.017212   \n",
       "1     0   131                      0.850586                     0.017212   \n",
       "2     0   131                      0.850586                     0.017212   \n",
       "3     0   131                      0.850586                     0.017212   \n",
       "4     0   131                      0.850586                     0.017212   \n",
       "\n",
       "   tags_emb_0  tags_emb_1  \n",
       "0   -0.007187    0.090353  \n",
       "1   -0.038487   -0.097879  \n",
       "2   -0.007187    0.090353  \n",
       "3    0.106040   -0.152273  \n",
       "4   -0.220193    0.188346  "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(13523, 8)\n",
      "CPU times: user 1min 3s, sys: 17 s, total: 1min 20s\n",
      "Wall time: 1min 17s\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>user_id</th>\n",
       "      <th>content_id</th>\n",
       "      <th>part</th>\n",
       "      <th>tag1</th>\n",
       "      <th>c_question_count_percent</th>\n",
       "      <th>c_question_part_percent</th>\n",
       "      <th>c_question_tag1_percent</th>\n",
       "      <th>c_question_unique_users_seen</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>705741139</td>\n",
       "      <td>128</td>\n",
       "      <td>1</td>\n",
       "      <td>131</td>\n",
       "      <td>0.000198</td>\n",
       "      <td>0.075093</td>\n",
       "      <td>0.041839</td>\n",
       "      <td>19260</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>705741139</td>\n",
       "      <td>7860</td>\n",
       "      <td>1</td>\n",
       "      <td>131</td>\n",
       "      <td>0.000219</td>\n",
       "      <td>0.075093</td>\n",
       "      <td>0.041839</td>\n",
       "      <td>21261</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>705741139</td>\n",
       "      <td>7922</td>\n",
       "      <td>1</td>\n",
       "      <td>131</td>\n",
       "      <td>0.000196</td>\n",
       "      <td>0.075093</td>\n",
       "      <td>0.041839</td>\n",
       "      <td>19065</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>705741139</td>\n",
       "      <td>156</td>\n",
       "      <td>1</td>\n",
       "      <td>131</td>\n",
       "      <td>0.000227</td>\n",
       "      <td>0.075093</td>\n",
       "      <td>0.041839</td>\n",
       "      <td>21908</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>705741139</td>\n",
       "      <td>51</td>\n",
       "      <td>1</td>\n",
       "      <td>131</td>\n",
       "      <td>0.000243</td>\n",
       "      <td>0.075093</td>\n",
       "      <td>0.041839</td>\n",
       "      <td>23373</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     user_id  content_id  part  tag1  c_question_count_percent  \\\n",
       "0  705741139         128     1   131                  0.000198   \n",
       "1  705741139        7860     1   131                  0.000219   \n",
       "2  705741139        7922     1   131                  0.000196   \n",
       "3  705741139         156     1   131                  0.000227   \n",
       "4  705741139          51     1   131                  0.000243   \n",
       "\n",
       "   c_question_part_percent  c_question_tag1_percent  \\\n",
       "0                 0.075093                 0.041839   \n",
       "1                 0.075093                 0.041839   \n",
       "2                 0.075093                 0.041839   \n",
       "3                 0.075093                 0.041839   \n",
       "4                 0.075093                 0.041839   \n",
       "\n",
       "   c_question_unique_users_seen  \n",
       "0                         19260  \n",
       "1                         21261  \n",
       "2                         19065  \n",
       "3                         21908  \n",
       "4                         23373  "
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "# 静态特征\n",
    "\n",
    "use_cols = ['user_id', 'content_id', 'part', 'tag1']\n",
    "\n",
    "df = pd.concat([train[use_cols], valid[use_cols]])\n",
    "\n",
    "df['c_question_count_percent'] = df.groupby(['content_id'])['content_id'].transform('count') / len(df)\n",
    "df['c_question_part_percent'] = df.groupby(['part'])['part'].transform('count') / len(df)\n",
    "df['c_question_tag1_percent'] = df.groupby(['tag1'])['tag1'].transform('count') / len(df)\n",
    "df['c_question_unique_users_seen'] = df.groupby(['content_id'])['user_id'].transform('nunique')\n",
    "\n",
    "df = df.drop_duplicates(subset=['content_id'])\n",
    "gc.collect()\n",
    "\n",
    "print(df.shape)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(96817414, 32) (2453886, 32)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>row_id</th>\n",
       "      <th>user_id</th>\n",
       "      <th>content_id</th>\n",
       "      <th>content_type_id</th>\n",
       "      <th>task_container_id</th>\n",
       "      <th>answered_correctly</th>\n",
       "      <th>prior_question_elapsed_time</th>\n",
       "      <th>prior_question_had_explanation</th>\n",
       "      <th>timestamp</th>\n",
       "      <th>answered_correctly_avg_c</th>\n",
       "      <th>answered_correctly_var_c</th>\n",
       "      <th>lecture_part</th>\n",
       "      <th>part</th>\n",
       "      <th>question_asked</th>\n",
       "      <th>right_answers</th>\n",
       "      <th>bundle_size</th>\n",
       "      <th>bundle_rignt_answers</th>\n",
       "      <th>bundle_questions_asked</th>\n",
       "      <th>bundle_accuracy</th>\n",
       "      <th>part_rignt_answers</th>\n",
       "      <th>part_questions_asked</th>\n",
       "      <th>part_accuracy</th>\n",
       "      <th>LorR</th>\n",
       "      <th>tag1</th>\n",
       "      <th>tag1_answered_correctly_mean</th>\n",
       "      <th>tag1_answered_correctly_var</th>\n",
       "      <th>tags_emb_0</th>\n",
       "      <th>tags_emb_1</th>\n",
       "      <th>c_question_count_percent</th>\n",
       "      <th>c_question_part_percent</th>\n",
       "      <th>c_question_tag1_percent</th>\n",
       "      <th>c_question_unique_users_seen</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>32933156</td>\n",
       "      <td>705741139</td>\n",
       "      <td>128</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.966887</td>\n",
       "      <td>0.032018</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>19056</td>\n",
       "      <td>18425.0</td>\n",
       "      <td>1</td>\n",
       "      <td>18425.0</td>\n",
       "      <td>19056</td>\n",
       "      <td>0.966887</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>0.850586</td>\n",
       "      <td>0.017212</td>\n",
       "      <td>-0.007187</td>\n",
       "      <td>0.090353</td>\n",
       "      <td>0.000198</td>\n",
       "      <td>0.075093</td>\n",
       "      <td>0.041839</td>\n",
       "      <td>19260</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>32933157</td>\n",
       "      <td>705741139</td>\n",
       "      <td>7860</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>16000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>20666</td>\n",
       "      <td>0.954720</td>\n",
       "      <td>0.043232</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>21135</td>\n",
       "      <td>20178.0</td>\n",
       "      <td>1</td>\n",
       "      <td>20178.0</td>\n",
       "      <td>21135</td>\n",
       "      <td>0.954720</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>0.850586</td>\n",
       "      <td>0.017212</td>\n",
       "      <td>-0.038487</td>\n",
       "      <td>-0.097879</td>\n",
       "      <td>0.000219</td>\n",
       "      <td>0.075093</td>\n",
       "      <td>0.041839</td>\n",
       "      <td>21261</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>32933158</td>\n",
       "      <td>705741139</td>\n",
       "      <td>7922</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>19000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>39172</td>\n",
       "      <td>0.953161</td>\n",
       "      <td>0.044647</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>18852</td>\n",
       "      <td>17969.0</td>\n",
       "      <td>1</td>\n",
       "      <td>17969.0</td>\n",
       "      <td>18852</td>\n",
       "      <td>0.953161</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>0.850586</td>\n",
       "      <td>0.017212</td>\n",
       "      <td>-0.007187</td>\n",
       "      <td>0.090353</td>\n",
       "      <td>0.000196</td>\n",
       "      <td>0.075093</td>\n",
       "      <td>0.041839</td>\n",
       "      <td>19065</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>32933159</td>\n",
       "      <td>705741139</td>\n",
       "      <td>156</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>17000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>58207</td>\n",
       "      <td>0.932764</td>\n",
       "      <td>0.062719</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>21878</td>\n",
       "      <td>20407.0</td>\n",
       "      <td>1</td>\n",
       "      <td>20407.0</td>\n",
       "      <td>21878</td>\n",
       "      <td>0.932764</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>0.850586</td>\n",
       "      <td>0.017212</td>\n",
       "      <td>0.106040</td>\n",
       "      <td>-0.152273</td>\n",
       "      <td>0.000227</td>\n",
       "      <td>0.075093</td>\n",
       "      <td>0.041839</td>\n",
       "      <td>21908</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>32933160</td>\n",
       "      <td>705741139</td>\n",
       "      <td>51</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>17000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>75779</td>\n",
       "      <td>0.930161</td>\n",
       "      <td>0.064964</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>23454</td>\n",
       "      <td>21816.0</td>\n",
       "      <td>1</td>\n",
       "      <td>21816.0</td>\n",
       "      <td>23454</td>\n",
       "      <td>0.930161</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>0.850586</td>\n",
       "      <td>0.017212</td>\n",
       "      <td>-0.220193</td>\n",
       "      <td>0.188346</td>\n",
       "      <td>0.000243</td>\n",
       "      <td>0.075093</td>\n",
       "      <td>0.041839</td>\n",
       "      <td>23373</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     row_id    user_id  content_id  content_type_id  task_container_id  \\\n",
       "0  32933156  705741139         128                0                  0   \n",
       "1  32933157  705741139        7860                0                  1   \n",
       "2  32933158  705741139        7922                0                  2   \n",
       "3  32933159  705741139         156                0                  3   \n",
       "4  32933160  705741139          51                0                  4   \n",
       "\n",
       "   answered_correctly  prior_question_elapsed_time  \\\n",
       "0                   1                          NaN   \n",
       "1                   1                      16000.0   \n",
       "2                   1                      19000.0   \n",
       "3                   1                      17000.0   \n",
       "4                   1                      17000.0   \n",
       "\n",
       "   prior_question_had_explanation  timestamp  answered_correctly_avg_c  \\\n",
       "0                               0          0                  0.966887   \n",
       "1                               0      20666                  0.954720   \n",
       "2                               0      39172                  0.953161   \n",
       "3                               0      58207                  0.932764   \n",
       "4                               0      75779                  0.930161   \n",
       "\n",
       "   answered_correctly_var_c  lecture_part  part  question_asked  \\\n",
       "0                  0.032018           NaN     1           19056   \n",
       "1                  0.043232           NaN     1           21135   \n",
       "2                  0.044647           NaN     1           18852   \n",
       "3                  0.062719           NaN     1           21878   \n",
       "4                  0.064964           NaN     1           23454   \n",
       "\n",
       "   right_answers  bundle_size  bundle_rignt_answers  bundle_questions_asked  \\\n",
       "0        18425.0            1               18425.0                   19056   \n",
       "1        20178.0            1               20178.0                   21135   \n",
       "2        17969.0            1               17969.0                   18852   \n",
       "3        20407.0            1               20407.0                   21878   \n",
       "4        21816.0            1               21816.0                   23454   \n",
       "\n",
       "   bundle_accuracy  part_rignt_answers  part_questions_asked  part_accuracy  \\\n",
       "0         0.966887           5425569.0               7276830       0.745595   \n",
       "1         0.954720           5425569.0               7276830       0.745595   \n",
       "2         0.953161           5425569.0               7276830       0.745595   \n",
       "3         0.932764           5425569.0               7276830       0.745595   \n",
       "4         0.930161           5425569.0               7276830       0.745595   \n",
       "\n",
       "   LorR  tag1  tag1_answered_correctly_mean  tag1_answered_correctly_var  \\\n",
       "0     0   131                      0.850586                     0.017212   \n",
       "1     0   131                      0.850586                     0.017212   \n",
       "2     0   131                      0.850586                     0.017212   \n",
       "3     0   131                      0.850586                     0.017212   \n",
       "4     0   131                      0.850586                     0.017212   \n",
       "\n",
       "   tags_emb_0  tags_emb_1  c_question_count_percent  c_question_part_percent  \\\n",
       "0   -0.007187    0.090353                  0.000198                 0.075093   \n",
       "1   -0.038487   -0.097879                  0.000219                 0.075093   \n",
       "2   -0.007187    0.090353                  0.000196                 0.075093   \n",
       "3    0.106040   -0.152273                  0.000227                 0.075093   \n",
       "4   -0.220193    0.188346                  0.000243                 0.075093   \n",
       "\n",
       "   c_question_tag1_percent  c_question_unique_users_seen  \n",
       "0                 0.041839                         19260  \n",
       "1                 0.041839                         21261  \n",
       "2                 0.041839                         19065  \n",
       "3                 0.041839                         21908  \n",
       "4                 0.041839                         23373  "
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "use_cols = ['content_id', 'c_question_count_percent', \n",
    "            'c_question_part_percent', 'c_question_tag1_percent',\n",
    "            'c_question_unique_users_seen']\n",
    "\n",
    "train = pd.merge(train, df[use_cols], on='content_id', how='left')\n",
    "valid = pd.merge(valid, df[use_cols], on='content_id', how='left')\n",
    "\n",
    "print(train.shape, valid.shape)\n",
    "train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(13523, 23)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>content_id</th>\n",
       "      <th>correct_answer</th>\n",
       "      <th>part</th>\n",
       "      <th>mean_answered_correctly_content</th>\n",
       "      <th>question_asked</th>\n",
       "      <th>right_answers</th>\n",
       "      <th>bundle_size</th>\n",
       "      <th>bundle_rignt_answers</th>\n",
       "      <th>bundle_questions_asked</th>\n",
       "      <th>bundle_accuracy</th>\n",
       "      <th>part_rignt_answers</th>\n",
       "      <th>part_questions_asked</th>\n",
       "      <th>part_accuracy</th>\n",
       "      <th>LorR</th>\n",
       "      <th>tag1</th>\n",
       "      <th>tag1_answered_correctly_mean</th>\n",
       "      <th>tag1_answered_correctly_var</th>\n",
       "      <th>tags_emb_0</th>\n",
       "      <th>tags_emb_1</th>\n",
       "      <th>c_question_count_percent</th>\n",
       "      <th>c_question_part_percent</th>\n",
       "      <th>c_question_tag1_percent</th>\n",
       "      <th>c_question_unique_users_seen</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.907334</td>\n",
       "      <td>6777</td>\n",
       "      <td>6149.0</td>\n",
       "      <td>1</td>\n",
       "      <td>6149.0</td>\n",
       "      <td>6777</td>\n",
       "      <td>0.907334</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>51</td>\n",
       "      <td>0.838379</td>\n",
       "      <td>0.017258</td>\n",
       "      <td>0.179532</td>\n",
       "      <td>0.141397</td>\n",
       "      <td>0.000070</td>\n",
       "      <td>0.075093</td>\n",
       "      <td>0.003259</td>\n",
       "      <td>6380</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0.890571</td>\n",
       "      <td>7265</td>\n",
       "      <td>6470.0</td>\n",
       "      <td>1</td>\n",
       "      <td>6470.0</td>\n",
       "      <td>7265</td>\n",
       "      <td>0.890571</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>0.850586</td>\n",
       "      <td>0.017212</td>\n",
       "      <td>-0.196008</td>\n",
       "      <td>-0.232887</td>\n",
       "      <td>0.000075</td>\n",
       "      <td>0.075093</td>\n",
       "      <td>0.041839</td>\n",
       "      <td>6829</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.554213</td>\n",
       "      <td>43781</td>\n",
       "      <td>24264.0</td>\n",
       "      <td>1</td>\n",
       "      <td>24264.0</td>\n",
       "      <td>43781</td>\n",
       "      <td>0.554213</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>0.850586</td>\n",
       "      <td>0.017212</td>\n",
       "      <td>0.034004</td>\n",
       "      <td>-0.229731</td>\n",
       "      <td>0.000452</td>\n",
       "      <td>0.075093</td>\n",
       "      <td>0.041839</td>\n",
       "      <td>38322</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.779386</td>\n",
       "      <td>22451</td>\n",
       "      <td>17498.0</td>\n",
       "      <td>1</td>\n",
       "      <td>17498.0</td>\n",
       "      <td>22451</td>\n",
       "      <td>0.779386</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>0.850586</td>\n",
       "      <td>0.017212</td>\n",
       "      <td>0.152637</td>\n",
       "      <td>-0.025414</td>\n",
       "      <td>0.000231</td>\n",
       "      <td>0.075093</td>\n",
       "      <td>0.041839</td>\n",
       "      <td>20643</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0.613970</td>\n",
       "      <td>30881</td>\n",
       "      <td>18960.0</td>\n",
       "      <td>1</td>\n",
       "      <td>18960.0</td>\n",
       "      <td>30881</td>\n",
       "      <td>0.613970</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>0.850586</td>\n",
       "      <td>0.017212</td>\n",
       "      <td>0.153832</td>\n",
       "      <td>-0.154031</td>\n",
       "      <td>0.000320</td>\n",
       "      <td>0.075093</td>\n",
       "      <td>0.041839</td>\n",
       "      <td>29006</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   content_id  correct_answer  part  mean_answered_correctly_content  \\\n",
       "0           0               0     1                         0.907334   \n",
       "1           1               1     1                         0.890571   \n",
       "2           2               0     1                         0.554213   \n",
       "3           3               0     1                         0.779386   \n",
       "4           4               3     1                         0.613970   \n",
       "\n",
       "   question_asked  right_answers  bundle_size  bundle_rignt_answers  \\\n",
       "0            6777         6149.0            1                6149.0   \n",
       "1            7265         6470.0            1                6470.0   \n",
       "2           43781        24264.0            1               24264.0   \n",
       "3           22451        17498.0            1               17498.0   \n",
       "4           30881        18960.0            1               18960.0   \n",
       "\n",
       "   bundle_questions_asked  bundle_accuracy  part_rignt_answers  \\\n",
       "0                    6777         0.907334           5425569.0   \n",
       "1                    7265         0.890571           5425569.0   \n",
       "2                   43781         0.554213           5425569.0   \n",
       "3                   22451         0.779386           5425569.0   \n",
       "4                   30881         0.613970           5425569.0   \n",
       "\n",
       "   part_questions_asked  part_accuracy  LorR  tag1  \\\n",
       "0               7276830       0.745595     0    51   \n",
       "1               7276830       0.745595     0   131   \n",
       "2               7276830       0.745595     0   131   \n",
       "3               7276830       0.745595     0   131   \n",
       "4               7276830       0.745595     0   131   \n",
       "\n",
       "   tag1_answered_correctly_mean  tag1_answered_correctly_var  tags_emb_0  \\\n",
       "0                      0.838379                     0.017258    0.179532   \n",
       "1                      0.850586                     0.017212   -0.196008   \n",
       "2                      0.850586                     0.017212    0.034004   \n",
       "3                      0.850586                     0.017212    0.152637   \n",
       "4                      0.850586                     0.017212    0.153832   \n",
       "\n",
       "   tags_emb_1  c_question_count_percent  c_question_part_percent  \\\n",
       "0    0.141397                  0.000070                 0.075093   \n",
       "1   -0.232887                  0.000075                 0.075093   \n",
       "2   -0.229731                  0.000452                 0.075093   \n",
       "3   -0.025414                  0.000231                 0.075093   \n",
       "4   -0.154031                  0.000320                 0.075093   \n",
       "\n",
       "   c_question_tag1_percent  c_question_unique_users_seen  \n",
       "0                 0.003259                          6380  \n",
       "1                 0.041839                          6829  \n",
       "2                 0.041839                         38322  \n",
       "3                 0.041839                         20643  \n",
       "4                 0.041839                         29006  "
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "questions_df = pd.merge(questions_df, df[['content_id',\n",
    "                                          'c_question_count_percent', \n",
    "                                          'c_question_part_percent',\n",
    "                                          'c_question_tag1_percent',\n",
    "                                          'c_question_unique_users_seen']], on='content_id', how='left')\n",
    "\n",
    "print(questions_df.shape)\n",
    "questions_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "train.fillna(-1, inplace=True)\n",
    "valid.fillna(-1, inplace=True)\n",
    "\n",
    "use_cols = ['user_id', 'content_type_id', 'answered_correctly', \n",
    "            'part', 'tag1', 'timestamp', 'task_container_id',\n",
    "            'prior_question_had_explanation', 'LorR', 'content_id']\n",
    "\n",
    "train[use_cols] = train[use_cols].astype('int')\n",
    "valid[use_cols] = valid[use_cols].astype('int')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "from collections import OrderedDict\n",
    "\n",
    "class LRUCache(OrderedDict):\n",
    "\n",
    "    def __init__(self, capacity=8000000):\n",
    "        self.capacity = capacity\n",
    "        self.cache = OrderedDict()\n",
    "     \n",
    "\n",
    "    def get(self,key):\n",
    "        if key in self.cache:\n",
    "            value = self.cache.pop(key)\n",
    "            self.cache[key] = value\n",
    "        else:\n",
    "            value = -1\n",
    "         \n",
    "        return value\n",
    "     \n",
    "\n",
    "    def set(self,key,value):\n",
    "        if key in self.cache:\n",
    "            value = self.cache.pop(key)\n",
    "            self.cache[key] = value\n",
    "        else:\n",
    "            if len(self.cache) == self.capacity:\n",
    "                self.cache.popitem(last = False)\n",
    "                self.cache[key] = value\n",
    "            else:\n",
    "                self.cache[key] = value\n",
    "                \n",
    "                \n",
    "lru_cache = LRUCache()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "def add_user_feats(df,\n",
    "                   u_answered_correctly_count_dict,                     # 用户答对的次数\n",
    "                   u_answered_count_dict,                               # 用户作答的次数\n",
    "                   u_question_part_correctly_count_dict,                # 用户作答 part 类型题目且答对的次数\n",
    "                   u_question_part_count_dict,                          # 用户作答 part 类型题目的次数\n",
    "                   u_question_tag1_correctly_count_dict,                # 用户作答 tag1 标签题目且答对的次数\n",
    "                   u_question_tag1_count_dict,                          # 用户作答 tag1 标签题目的次数\n",
    "                   u_prior_question_correctly_timestamp_dict,           # 用户上次正确作答的时间戳\n",
    "                   u_prior2_question_correctly_timestamp_dict,          # 用户上次正确作答的时间戳\n",
    "                   u_prior3_question_correctly_timestamp_dict,          # 用户上次正确作答的时间戳\n",
    "                   u_prior_question_timestamp_dict,                     # 用户上次作答的时间戳\n",
    "                   u_prior2_question_timestamp_dict,                    # 用户上上次作答的时间戳\n",
    "                   u_prior3_question_timestamp_dict,                    # 用户上上上次作答的时间戳\n",
    "                   u_prior4_question_timestamp_dict,                    # 用户上上上上次作答的时间戳\n",
    "                   u_prior_lecture_timestamp_dict,                      # 用户上次观看 lecture 的时间戳\n",
    "                   u_prior2_lecture_timestamp_dict,                     # 用户上上次观看 lecture 的时间戳\n",
    "                   u_task_container_id_dict,                            # 用户 task id\n",
    "                   u_prior_question_explanation_count_dict,             # 用户上次查看问题解答的次数\n",
    "                   u_prior_question_explanation_correctly_count_dict,   # 用户上次查看问题解答后正确答对的次数\n",
    "                   u_question_listening_correctly_count_dict,           # 用户作答听力题目且答对的次数\n",
    "                   u_question_reading_correctly_count_dict,             # 用户作答阅读题目且答对的次数\n",
    "                   u_question_listening_count_dict,                     # 用户作答听力题目且答对的次数\n",
    "                   u_question_reading_count_dict,                       # 用户作答阅读题目且答对的次数\n",
    "                   u_question_incorrect_timestamp_dict,                 # 用户上次答错的时间戳\n",
    "                   u_question_incorrect_timestamp2_dict,                # 用户上上次答错的时间戳\n",
    "                   u_question_incorrect_timestamp3_dict,                # 用户上上上次答错的时间戳\n",
    "                   u_question_seen_dict,                                # 这道题用户是否做过\n",
    "                   u_question_part_correct_timestamp_dict,              # 用户上次该 part 答对的时间戳\n",
    "                   u_question_part_incorrect_timestamp_dict,            # 用户上次该 part 答错的时间戳\n",
    "                  ):\n",
    "    \n",
    "    uacc = np.zeros(len(df), dtype=np.int32)\n",
    "    uac = np.zeros(len(df), dtype=np.int32)\n",
    "    uqpcc = np.zeros(len(df), dtype=np.int32)\n",
    "    uqpc = np.zeros(len(df), dtype=np.int32)\n",
    "    uqt1cc = np.zeros(len(df), dtype=np.int32)\n",
    "    uqt1c = np.zeros(len(df), dtype=np.int32)\n",
    "    upqct = np.zeros(len(df), dtype=np.int32)\n",
    "    up2qct = np.zeros(len(df), dtype=np.int32)\n",
    "    up3qct = np.zeros(len(df), dtype=np.int32)\n",
    "    upqt = np.zeros(len(df), dtype=np.int32)\n",
    "    up2qt = np.zeros(len(df), dtype=np.int32)\n",
    "    up3qt = np.zeros(len(df), dtype=np.int32)\n",
    "    up4qt = np.zeros(len(df), dtype=np.int32)\n",
    "    uplt = np.zeros(len(df), dtype=np.int32)\n",
    "    up2lt = np.zeros(len(df), dtype=np.int32)\n",
    "    utci = np.zeros(len(df), dtype=np.int32)\n",
    "    upqec = np.zeros(len(df), dtype=np.int32)\n",
    "    upqecc = np.zeros(len(df), dtype=np.int32)\n",
    "    uqlcc = np.zeros(len(df), dtype=np.int32)\n",
    "    uqrcc = np.zeros(len(df), dtype=np.int32)\n",
    "    uqlc = np.zeros(len(df), dtype=np.int32)\n",
    "    uqrc = np.zeros(len(df), dtype=np.int32)\n",
    "    uqict = np.zeros(len(df), dtype=np.int32)\n",
    "    uqict2 = np.zeros(len(df), dtype=np.int32)\n",
    "    uqict3 = np.zeros(len(df), dtype=np.int32)\n",
    "    uqs = np.zeros(len(df), dtype=np.int32)\n",
    "    uqn = np.zeros(len(df), dtype=np.int32)        # 用户一共做了多少道不同的题目\n",
    "    uqpct = np.zeros(len(df), dtype=np.int32)\n",
    "    uqpict = np.zeros(len(df), dtype=np.int32)\n",
    "    uct = np.zeros(len(df), dtype=np.int32)\n",
    "    \n",
    "    for cnt, row in enumerate(tqdm(df[['user_id',\n",
    "                                       'content_type_id',\n",
    "                                       'answered_correctly', \n",
    "                                       'part',\n",
    "                                       'tag1',\n",
    "                                       'timestamp',\n",
    "                                       'task_container_id',\n",
    "                                       'prior_question_had_explanation',\n",
    "                                       'LorR',\n",
    "                                       'content_id']].values)):\n",
    "        user_id = row[0]\n",
    "        content_type = row[1]\n",
    "        answered_correctly = row[2]\n",
    "        question_part = row[3]\n",
    "        question_tag1 = row[4]\n",
    "        timestamp = row[5]\n",
    "        task_container_id = row[6]\n",
    "        prior_question_had_explanation = row[7]\n",
    "        LorR = row[8]\n",
    "        content_id = row[9]\n",
    "        \n",
    "        uid_part = str(user_id) + '_' + str(question_part)\n",
    "        uid_tag1 = str(user_id) + '_' + str(question_tag1)\n",
    "        uid_cid = int(user_id) + int(content_id)*10e10\n",
    "        \n",
    "        uct[cnt] = timestamp - lru_cache.get(uid_cid)\n",
    "        uacc[cnt] = u_answered_correctly_count_dict[user_id]\n",
    "        uac[cnt] = u_answered_count_dict[user_id]\n",
    "        uqpcc[cnt] = u_question_part_correctly_count_dict[uid_part]\n",
    "        uqpc[cnt] = u_question_part_count_dict[uid_part]\n",
    "        uqt1cc[cnt] = u_question_tag1_correctly_count_dict[uid_tag1]\n",
    "        uqt1c[cnt] = u_question_tag1_count_dict[uid_tag1]\n",
    "        upqct[cnt] = timestamp - u_prior_question_correctly_timestamp_dict[user_id]\n",
    "        up2qct[cnt] = timestamp - u_prior2_question_correctly_timestamp_dict[user_id]\n",
    "        up3qct[cnt] = timestamp - u_prior3_question_correctly_timestamp_dict[user_id]\n",
    "        upqt[cnt] = timestamp - u_prior_question_timestamp_dict[user_id]\n",
    "        up2qt[cnt] = timestamp - u_prior2_question_timestamp_dict[user_id]\n",
    "        up3qt[cnt] = timestamp - u_prior3_question_timestamp_dict[user_id]\n",
    "        up4qt[cnt] = timestamp - u_prior4_question_timestamp_dict[user_id]\n",
    "        uplt[cnt] = timestamp - u_prior_lecture_timestamp_dict[user_id]\n",
    "        up2lt[cnt] = timestamp - u_prior2_lecture_timestamp_dict[user_id]\n",
    "        utci[cnt] = task_container_id - u_task_container_id_dict[user_id]\n",
    "        upqec[cnt] = u_prior_question_explanation_count_dict[user_id]\n",
    "        upqecc[cnt] = u_prior_question_explanation_correctly_count_dict[user_id]\n",
    "        uqlcc[cnt] = u_question_listening_correctly_count_dict[user_id]\n",
    "        uqrcc[cnt] = u_question_reading_correctly_count_dict[user_id]\n",
    "        uqlc[cnt] = u_question_listening_count_dict[user_id]\n",
    "        uqrc[cnt] = u_question_reading_count_dict[user_id]\n",
    "        uqict[cnt] = timestamp - u_question_incorrect_timestamp_dict[user_id]\n",
    "        uqict2[cnt] = timestamp - u_question_incorrect_timestamp2_dict[user_id]\n",
    "        uqict3[cnt] = timestamp - u_question_incorrect_timestamp3_dict[user_id]\n",
    "        uqpct[cnt] = timestamp - u_question_part_correct_timestamp_dict[uid_part]\n",
    "        uqpict[cnt] = timestamp - u_question_part_incorrect_timestamp_dict[uid_part]\n",
    "        if content_type == 0:\n",
    "            if user_id not in u_question_seen_dict:\n",
    "                u_question_seen_dict[user_id] = bitarray('0'*14000, endian='little')\n",
    "            uqs[cnt] = u_question_seen_dict[user_id][content_id]\n",
    "            uqn[cnt] = u_question_seen_dict[user_id].count()\n",
    "        \n",
    "        if content_type == 0:\n",
    "                \n",
    "            if answered_correctly == 1:\n",
    "                u_prior3_question_correctly_timestamp_dict[user_id] = u_prior2_question_correctly_timestamp_dict[user_id]\n",
    "                u_prior2_question_correctly_timestamp_dict[user_id] = u_prior_question_correctly_timestamp_dict[user_id]\n",
    "                u_prior_question_correctly_timestamp_dict[user_id] = timestamp\n",
    "                u_prior_question_explanation_correctly_count_dict[user_id] += 1\n",
    "                u_question_part_correct_timestamp_dict[uid_part] = timestamp\n",
    "                if LorR == 0:\n",
    "                    u_question_listening_correctly_count_dict[user_id] += 1\n",
    "                else:\n",
    "                    u_question_reading_correctly_count_dict[user_id] += 1\n",
    "            else:\n",
    "                u_question_incorrect_timestamp3_dict[user_id] = u_question_incorrect_timestamp2_dict[user_id]\n",
    "                u_question_incorrect_timestamp2_dict[user_id] = u_question_incorrect_timestamp_dict[user_id]\n",
    "                u_question_incorrect_timestamp_dict[user_id] = timestamp\n",
    "                u_question_part_incorrect_timestamp_dict[uid_part] = timestamp\n",
    "                \n",
    "            u_answered_correctly_count_dict[user_id] += answered_correctly\n",
    "            u_answered_count_dict[user_id] += 1\n",
    "            u_question_part_correctly_count_dict[uid_part] += answered_correctly\n",
    "            u_question_part_count_dict[uid_part] += 1\n",
    "            u_question_tag1_correctly_count_dict[uid_tag1] += answered_correctly\n",
    "            u_question_tag1_count_dict[uid_tag1] += 1\n",
    "            u_prior4_question_timestamp_dict[user_id] = u_prior3_question_timestamp_dict[user_id]\n",
    "            u_prior3_question_timestamp_dict[user_id] = u_prior2_question_timestamp_dict[user_id]\n",
    "            u_prior2_question_timestamp_dict[user_id] = u_prior_question_timestamp_dict[user_id]\n",
    "            u_prior_question_timestamp_dict[user_id] = timestamp\n",
    "            u_task_container_id_dict[user_id] = task_container_id\n",
    "            u_prior_question_explanation_count_dict[user_id] += prior_question_had_explanation\n",
    "            if LorR == 0:\n",
    "                u_question_listening_count_dict[user_id] += 1\n",
    "            else:\n",
    "                u_question_reading_count_dict[user_id] += 1\n",
    "            u_question_seen_dict[user_id][content_id] = 1\n",
    "            lru_cache.set(uid_cid, timestamp)\n",
    "            \n",
    "        else:\n",
    "            u_prior2_lecture_timestamp_dict[user_id] = u_prior_lecture_timestamp_dict[user_id]\n",
    "            u_prior_lecture_timestamp_dict[user_id] = timestamp\n",
    "        \n",
    "        \n",
    "    user_feats_df = pd.DataFrame({'u_answered_correctly_count': uacc, \n",
    "                                  'u_answered_count': uac,\n",
    "                                  'u_question_part_correctly_count': uqpcc,\n",
    "                                  'u_question_part_count': uqpc,\n",
    "                                  'u_question_tag1_correctly_count': uqt1cc,\n",
    "                                  'u_question_tag1_count': uqt1c,\n",
    "                                  'u_prior_question_correctly_timestamp_diff': upqct,\n",
    "                                  'u_prior_question_correctly_timestamp_diff2': up2qct,\n",
    "                                  'u_prior_question_correctly_timestamp_diff3': up3qct,\n",
    "                                  'u_prior_question_timestamp_diff': upqt,\n",
    "                                  'u_prior_question_timestamp_diff2': up2qt,\n",
    "                                  'u_prior_question_timestamp_diff3': up3qt,\n",
    "                                  'u_prior_question_timestamp_diff4': up4qt,\n",
    "                                  'u_prior_lecture_timestamp_diff': uplt,\n",
    "                                  'u_prior2_lecture_timestamp_diff': up2lt,\n",
    "                                  'u_task_container_id_diff': utci,\n",
    "                                  'u_prior_question_explanation_count': upqec,\n",
    "                                  'u_prior_question_explanation_correctly_count': upqecc,\n",
    "                                  'u_question_listening_correctly_count': uqlcc,\n",
    "                                  'u_question_reading_correctly_count': uqrcc,\n",
    "                                  'u_question_listening_count': uqlc,\n",
    "                                  'u_question_reading_count': uqrc,\n",
    "                                  'u_question_incorrect_timestamp_diff': uqict,\n",
    "                                  'u_question_incorrect_timestamp_diff2': uqict2,\n",
    "                                  'u_question_incorrect_timestamp_diff3': uqict3,\n",
    "                                  'u_question_seen': uqs,\n",
    "                                  'u_question_nunique': uqn,\n",
    "                                  'u_question_part_correct_timestamp_diff': uqpct,\n",
    "                                  'u_question_part_incorrect_timestamp_diff': uqpict,\n",
    "                                  'uc_question_timestamp_diff': uct,\n",
    "                                 })\n",
    "    \n",
    "    user_feats_df['u_answered_correctly_avg'] = user_feats_df['u_answered_correctly_count'] / user_feats_df['u_answered_count']\n",
    "    user_feats_df['u_question_part_correctly_avg'] = user_feats_df['u_question_part_correctly_count'] / user_feats_df['u_question_part_count']\n",
    "    user_feats_df['u_question_tag1_correctly_avg'] = user_feats_df['u_question_tag1_correctly_count'] / user_feats_df['u_question_tag1_count']\n",
    "    user_feats_df['u_prior_question_timestamp_diff_2_1'] = user_feats_df['u_prior_question_timestamp_diff2'] - user_feats_df['u_prior_question_timestamp_diff']\n",
    "    user_feats_df['u_prior_question_timestamp_diff_3_2'] = user_feats_df['u_prior_question_timestamp_diff3'] - user_feats_df['u_prior_question_timestamp_diff2']\n",
    "    user_feats_df['u_prior_question_timestamp_diff_4_3'] = user_feats_df['u_prior_question_timestamp_diff4'] - user_feats_df['u_prior_question_timestamp_diff3']\n",
    "    \n",
    "    df = pd.concat([df, user_feats_df], axis=1)\n",
    "    df['u_question_timestamp_diff']  = df['u_prior_question_timestamp_diff_2_1'] - df['prior_question_elapsed_time']\n",
    "    df['u_question_timestamp_diff2']  = df['u_prior_question_timestamp_diff_3_2'] - df['prior_question_elapsed_time']\n",
    "    df['u_question_timestamp_diff3']  = df['u_prior_question_timestamp_diff_4_3'] - df['prior_question_elapsed_time']\n",
    "    df['u_listening_reading_ratio'] = df['u_question_listening_correctly_count'] / (1 + df['u_question_reading_correctly_count'])\n",
    "    df['u_listening_correctly_avg'] = df['u_question_listening_correctly_count'] / (1 + df['u_question_listening_count'])\n",
    "    df['u_reading_correctly_avg'] = df['u_question_reading_correctly_count'] / (1 + df['u_question_reading_count'])\n",
    "    \n",
    "    return df\n",
    "\n",
    "\n",
    "def add_user_feats_without_update(df, \n",
    "                                  u_answered_correctly_count_dict, \n",
    "                                  u_answered_count_dict, \n",
    "                                  u_question_part_correctly_count_dict,\n",
    "                                  u_question_part_count_dict,\n",
    "                                  u_question_tag1_correctly_count_dict,\n",
    "                                  u_question_tag1_count_dict,\n",
    "                                  u_prior_question_correctly_timestamp_dict,\n",
    "                                  u_prior2_question_correctly_timestamp_dict,\n",
    "                                  u_prior3_question_correctly_timestamp_dict,\n",
    "                                  u_prior_question_timestamp_dict,\n",
    "                                  u_prior2_question_timestamp_dict,\n",
    "                                  u_prior3_question_timestamp_dict,\n",
    "                                  u_prior4_question_timestamp_dict,\n",
    "                                  u_prior_lecture_timestamp_dict,\n",
    "                                  u_prior2_lecture_timestamp_dict,\n",
    "                                  u_task_container_id_dict,\n",
    "                                  u_prior_question_explanation_count_dict,\n",
    "                                  u_prior_question_explanation_correctly_count_dict,\n",
    "                                  u_question_listening_correctly_count_dict,\n",
    "                                  u_question_reading_correctly_count_dict,\n",
    "                                  u_question_listening_count_dict,\n",
    "                                  u_question_reading_count_dict,\n",
    "                                  u_question_incorrect_timestamp_dict,\n",
    "                                  u_question_incorrect_timestamp2_dict,\n",
    "                                  u_question_incorrect_timestamp3_dict,\n",
    "                                  u_question_seen_dict,\n",
    "                                  u_question_part_correct_timestamp_dict,\n",
    "                                  u_question_part_incorrect_timestamp_dict,\n",
    "                                 ):\n",
    "    \n",
    "    uacc = np.zeros(len(df), dtype=np.int32)\n",
    "    uac = np.zeros(len(df), dtype=np.int32)\n",
    "    uqpcc = np.zeros(len(df), dtype=np.int32)\n",
    "    uqpc = np.zeros(len(df), dtype=np.int32)\n",
    "    uqt1cc = np.zeros(len(df), dtype=np.int32)\n",
    "    uqt1c = np.zeros(len(df), dtype=np.int32)\n",
    "    upqct = np.zeros(len(df), dtype=np.int32)\n",
    "    up2qct = np.zeros(len(df), dtype=np.int32)\n",
    "    up3qct = np.zeros(len(df), dtype=np.int32)\n",
    "    upqt = np.zeros(len(df), dtype=np.int32)\n",
    "    up2qt = np.zeros(len(df), dtype=np.int32)\n",
    "    up3qt = np.zeros(len(df), dtype=np.int32)\n",
    "    up4qt = np.zeros(len(df), dtype=np.int32)\n",
    "    uplt = np.zeros(len(df), dtype=np.int32)\n",
    "    up2lt = np.zeros(len(df), dtype=np.int32)\n",
    "    utci = np.zeros(len(df), dtype=np.int32)\n",
    "    upqec = np.zeros(len(df), dtype=np.int32)\n",
    "    upqecc = np.zeros(len(df), dtype=np.int32)\n",
    "    uqlcc = np.zeros(len(df), dtype=np.int32)\n",
    "    uqrcc = np.zeros(len(df), dtype=np.int32)\n",
    "    uqlc = np.zeros(len(df), dtype=np.int32)\n",
    "    uqrc = np.zeros(len(df), dtype=np.int32)\n",
    "    uqict = np.zeros(len(df), dtype=np.int32)\n",
    "    uqict2 = np.zeros(len(df), dtype=np.int32)\n",
    "    uqict3 = np.zeros(len(df), dtype=np.int32)\n",
    "    uqs = np.zeros(len(df), dtype=np.int32)\n",
    "    uqn = np.zeros(len(df), dtype=np.int32)\n",
    "    uqpct = np.zeros(len(df), dtype=np.int32)\n",
    "    uqpict = np.zeros(len(df), dtype=np.int32)\n",
    "    uct = np.zeros(len(df), dtype=np.int32)\n",
    "    \n",
    "    for cnt, row in enumerate(df[['user_id',\n",
    "                                  'content_type_id', \n",
    "                                  'part',\n",
    "                                  'tag1',\n",
    "                                  'timestamp',\n",
    "                                  'task_container_id',\n",
    "                                  'prior_question_had_explanation',\n",
    "                                  'LorR',\n",
    "                                  'content_id']].values):\n",
    "        user_id = row[0]\n",
    "        content_type = row[1]\n",
    "        question_part = row[2]\n",
    "        question_tag1 = row[3]\n",
    "        timestamp = row[4]\n",
    "        task_container_id = row[5]\n",
    "        prior_question_had_explanation = row[6]\n",
    "        LorR = row[7]\n",
    "        content_id = row[8]\n",
    "        \n",
    "        uid_part = str(user_id) + '_' + str(question_part)\n",
    "        uid_tag1 = str(user_id) + '_' + str(question_tag1)\n",
    "        uid_cid = int(user_id) + int(content_id)*10e10\n",
    "        \n",
    "        uct[cnt] = timestamp - lru_cache.get(uid_cid)\n",
    "        uacc[cnt] = u_answered_correctly_count_dict[user_id]\n",
    "        uac[cnt] = u_answered_count_dict[user_id]\n",
    "        uqpcc[cnt] = u_question_part_correctly_count_dict[uid_part]\n",
    "        uqpc[cnt] = u_question_part_count_dict[uid_part]\n",
    "        uqt1cc[cnt] = u_question_tag1_correctly_count_dict[uid_tag1]\n",
    "        uqt1c[cnt] = u_question_tag1_count_dict[uid_tag1]\n",
    "        upqct[cnt] = timestamp - u_prior_question_correctly_timestamp_dict[user_id]\n",
    "        up2qct[cnt] = timestamp - u_prior2_question_correctly_timestamp_dict[user_id]\n",
    "        up3qct[cnt] = timestamp - u_prior3_question_correctly_timestamp_dict[user_id]\n",
    "        upqt[cnt] = timestamp - u_prior_question_timestamp_dict[user_id]\n",
    "        up2qt[cnt] = timestamp - u_prior2_question_timestamp_dict[user_id]\n",
    "        up3qt[cnt] = timestamp - u_prior3_question_timestamp_dict[user_id]\n",
    "        up4qt[cnt] = timestamp - u_prior4_question_timestamp_dict[user_id]\n",
    "        uplt[cnt] = timestamp - u_prior_lecture_timestamp_dict[user_id]\n",
    "        up2lt[cnt] = timestamp - u_prior2_lecture_timestamp_dict[user_id]\n",
    "        utci[cnt] = task_container_id - u_task_container_id_dict[user_id]\n",
    "        upqec[cnt] = u_prior_question_explanation_count_dict[user_id]\n",
    "        upqecc[cnt] = u_prior_question_explanation_correctly_count_dict[user_id]\n",
    "        uqlcc[cnt] = u_question_listening_correctly_count_dict[user_id]\n",
    "        uqrcc[cnt] = u_question_reading_correctly_count_dict[user_id]\n",
    "        uqlc[cnt] = u_question_listening_count_dict[user_id]\n",
    "        uqrc[cnt] = u_question_reading_count_dict[user_id]\n",
    "        uqict[cnt] = timestamp - u_question_incorrect_timestamp_dict[user_id]\n",
    "        uqict2[cnt] = timestamp - u_question_incorrect_timestamp2_dict[user_id]\n",
    "        uqict3[cnt] = timestamp - u_question_incorrect_timestamp3_dict[user_id]\n",
    "        uqpct[cnt] = timestamp - u_question_part_correct_timestamp_dict[uid_part]\n",
    "        uqpict[cnt] = timestamp - u_question_part_incorrect_timestamp_dict[uid_part]\n",
    "        if content_type == 0:\n",
    "            if user_id not in u_question_seen_dict:\n",
    "                u_question_seen_dict[user_id] = bitarray('0'*14000, endian='little')\n",
    "            uqs[cnt] = u_question_seen_dict[user_id][content_id]\n",
    "            uqn[cnt] = u_question_seen_dict[user_id].count()\n",
    "        \n",
    "    user_feats_df = pd.DataFrame({'u_answered_correctly_count': uacc, \n",
    "                                  'u_answered_count': uac,\n",
    "                                  'u_question_part_correctly_count': uqpcc,\n",
    "                                  'u_question_part_count': uqpc,\n",
    "                                  'u_question_tag1_correctly_count': uqt1cc,\n",
    "                                  'u_question_tag1_count': uqt1c,\n",
    "                                  'u_prior_question_correctly_timestamp_diff': upqct,\n",
    "                                  'u_prior_question_correctly_timestamp_diff2': up2qct,\n",
    "                                  'u_prior_question_correctly_timestamp_diff3': up3qct,\n",
    "                                  'u_prior_question_timestamp_diff': upqt,\n",
    "                                  'u_prior_question_timestamp_diff2': up2qt,\n",
    "                                  'u_prior_question_timestamp_diff3': up3qt,\n",
    "                                  'u_prior_question_timestamp_diff4': up4qt,\n",
    "                                  'u_prior_lecture_timestamp_diff': uplt,\n",
    "                                  'u_prior2_lecture_timestamp_diff': up2lt,\n",
    "                                  'u_task_container_id_diff': utci,\n",
    "                                  'u_prior_question_explanation_count': upqec,\n",
    "                                  'u_prior_question_explanation_correctly_count': upqecc,\n",
    "                                  'u_question_listening_correctly_count': uqlcc,\n",
    "                                  'u_question_reading_correctly_count': uqrcc,\n",
    "                                  'u_question_listening_count': uqlc,\n",
    "                                  'u_question_reading_count': uqrc,\n",
    "                                  'u_question_incorrect_timestamp_diff': uqict,\n",
    "                                  'u_question_incorrect_timestamp_diff2': uqict2,\n",
    "                                  'u_question_incorrect_timestamp_diff3': uqict3,\n",
    "                                  'u_question_seen': uqs,\n",
    "                                  'u_question_nunique': uqn,\n",
    "                                  'u_question_part_correct_timestamp_diff': uqpct,\n",
    "                                  'u_question_part_incorrect_timestamp_diff': uqpict,\n",
    "                                  'uc_question_timestamp_diff': uct,\n",
    "                                  })\n",
    "    \n",
    "    user_feats_df['u_answered_correctly_avg'] = user_feats_df['u_answered_correctly_count'] / user_feats_df['u_answered_count']\n",
    "    user_feats_df['u_question_part_correctly_avg'] = user_feats_df['u_question_part_correctly_count'] / user_feats_df['u_question_part_count']\n",
    "    user_feats_df['u_question_tag1_correctly_avg'] = user_feats_df['u_question_tag1_correctly_count'] / user_feats_df['u_question_tag1_count']\n",
    "    user_feats_df['u_prior_question_timestamp_diff_2_1'] = user_feats_df['u_prior_question_timestamp_diff2'] - user_feats_df['u_prior_question_timestamp_diff']\n",
    "    user_feats_df['u_prior_question_timestamp_diff_3_2'] = user_feats_df['u_prior_question_timestamp_diff3'] - user_feats_df['u_prior_question_timestamp_diff2']\n",
    "    user_feats_df['u_prior_question_timestamp_diff_4_3'] = user_feats_df['u_prior_question_timestamp_diff4'] - user_feats_df['u_prior_question_timestamp_diff3']\n",
    "    \n",
    "    df = pd.concat([df, user_feats_df], axis=1)\n",
    "    df['u_question_timestamp_diff']  = df['u_prior_question_timestamp_diff_2_1'] - df['prior_question_elapsed_time']\n",
    "    df['u_question_timestamp_diff2']  = df['u_prior_question_timestamp_diff_3_2'] - df['prior_question_elapsed_time']\n",
    "    df['u_question_timestamp_diff3']  = df['u_prior_question_timestamp_diff_4_3'] - df['prior_question_elapsed_time']\n",
    "    df['u_listening_reading_ratio'] = df['u_question_listening_correctly_count'] / (1 + df['u_question_reading_correctly_count'])\n",
    "    df['u_listening_correctly_avg'] = df['u_question_listening_correctly_count'] / (1 + df['u_question_listening_count'])\n",
    "    df['u_reading_correctly_avg'] = df['u_question_reading_correctly_count'] / (1 + df['u_question_reading_count'])\n",
    "    \n",
    "    return df\n",
    "\n",
    "\n",
    "def update_user_feats(df, \n",
    "                      u_answered_correctly_count_dict,\n",
    "                      u_answered_count_dict, \n",
    "                      u_question_part_correctly_count_dict,\n",
    "                      u_question_part_count_dict,\n",
    "                      u_question_tag1_correctly_count_dict,\n",
    "                      u_question_tag1_count_dict,\n",
    "                      u_prior_question_correctly_timestamp_dict,\n",
    "                      u_prior2_question_correctly_timestamp_dict,\n",
    "                      u_prior3_question_correctly_timestamp_dict,\n",
    "                      u_prior_question_timestamp_dict,\n",
    "                      u_prior2_question_timestamp_dict,\n",
    "                      u_prior3_question_timestamp_dict,\n",
    "                      u_prior4_question_timestamp_dict,\n",
    "                      u_prior_lecture_timestamp_dict, \n",
    "                      u_prior2_lecture_timestamp_dict,\n",
    "                      u_task_container_id_dict,\n",
    "                      u_prior_question_explanation_count_dict,\n",
    "                      u_prior_question_explanation_correctly_count_dict,\n",
    "                      u_question_listening_correctly_count_dict, \n",
    "                      u_question_reading_correctly_count_dict,\n",
    "                      u_question_listening_count_dict,\n",
    "                      u_question_reading_count_dict,\n",
    "                      u_question_incorrect_timestamp_dict,\n",
    "                      u_question_incorrect_timestamp2_dict,\n",
    "                      u_question_incorrect_timestamp3_dict,\n",
    "                      u_question_seen_dict,\n",
    "                      u_question_part_correct_timestamp_dict,\n",
    "                      u_question_part_incorrect_timestamp_dict,\n",
    "                     ):\n",
    "    \n",
    "    for cnt, row in enumerate(df[['user_id',\n",
    "                                  'content_type_id',\n",
    "                                  'answered_correctly', \n",
    "                                  'part',\n",
    "                                  'tag1',\n",
    "                                  'timestamp',\n",
    "                                  'task_container_id',\n",
    "                                  'prior_question_had_explanation',\n",
    "                                  'LorR',\n",
    "                                  'content_id']].values):\n",
    "        user_id = row[0]\n",
    "        content_type = row[1]\n",
    "        answered_correctly = row[2]\n",
    "        question_part = row[3]\n",
    "        question_tag1 = row[4]\n",
    "        timestamp = row[5]\n",
    "        task_container_id = row[6]\n",
    "        prior_question_had_explanation = row[7]\n",
    "        LorR = row[8]\n",
    "        content_id = row[9]\n",
    "        \n",
    "        uid_part = str(user_id) + '_' + str(question_part)\n",
    "        uid_tag1 = str(user_id) + '_' + str(question_tag1)\n",
    "        uid_cid = int(user_id) + int(content_id)*10e10\n",
    "        \n",
    "        if content_type == 0:\n",
    "            \n",
    "            if answered_correctly == 1:\n",
    "                u_prior3_question_correctly_timestamp_dict[user_id] = u_prior2_question_correctly_timestamp_dict[user_id]\n",
    "                u_prior2_question_correctly_timestamp_dict[user_id] = u_prior_question_correctly_timestamp_dict[user_id]\n",
    "                u_prior_question_correctly_timestamp_dict[user_id] = timestamp\n",
    "                u_prior_question_explanation_correctly_count_dict[user_id] += 1\n",
    "                u_question_part_correct_timestamp_dict[uid_part] = timestamp\n",
    "                if LorR == 0:\n",
    "                    u_question_listening_correctly_count_dict[user_id] += 1\n",
    "                else:\n",
    "                    u_question_reading_correctly_count_dict[user_id] += 1\n",
    "            else:\n",
    "                u_question_incorrect_timestamp3_dict[user_id] = u_question_incorrect_timestamp2_dict[user_id]\n",
    "                u_question_incorrect_timestamp2_dict[user_id] = u_question_incorrect_timestamp_dict[user_id]\n",
    "                u_question_incorrect_timestamp_dict[user_id] = timestamp\n",
    "                u_question_part_incorrect_timestamp_dict[uid_part] = timestamp\n",
    "                \n",
    "            u_answered_correctly_count_dict[user_id] += answered_correctly\n",
    "            u_answered_count_dict[user_id] += 1\n",
    "            u_question_part_correctly_count_dict[uid_part] += answered_correctly\n",
    "            u_question_part_count_dict[uid_part] += 1\n",
    "            u_question_tag1_correctly_count_dict[uid_tag1] += answered_correctly\n",
    "            u_question_tag1_count_dict[uid_tag1] += 1\n",
    "            u_prior4_question_timestamp_dict[user_id] = u_prior3_question_timestamp_dict[user_id]\n",
    "            u_prior3_question_timestamp_dict[user_id] = u_prior2_question_timestamp_dict[user_id]\n",
    "            u_prior2_question_timestamp_dict[user_id] = u_prior_question_timestamp_dict[user_id]\n",
    "            u_prior_question_timestamp_dict[user_id] = timestamp\n",
    "            u_task_container_id_dict[user_id] = task_container_id\n",
    "            u_prior_question_explanation_count_dict[user_id] += prior_question_had_explanation\n",
    "            if LorR == 0:\n",
    "                u_question_listening_count_dict[user_id] += 1\n",
    "            else:\n",
    "                u_question_reading_count_dict[user_id] += 1\n",
    "            u_question_seen_dict[user_id][content_id] = 1\n",
    "            lru_cache.set(uid_cid, timestamp)\n",
    "            \n",
    "        else:\n",
    "            u_prior2_lecture_timestamp_dict[user_id] = u_prior_lecture_timestamp_dict[user_id]\n",
    "            u_prior_lecture_timestamp_dict[user_id] = timestamp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "13e3080f9cf14f9b8cba1d96a6d09390",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=96817414.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f7ef13230ff74088baa520f5533d3ff0",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=2453886.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "CPU times: user 1h 28s, sys: 3min 34s, total: 1h 4min 3s\n",
      "Wall time: 1h 3min 27s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "# user stats features with loops\n",
    "\n",
    "u_answered_correctly_count_dict = defaultdict(int)\n",
    "u_answered_count_dict = defaultdict(int)\n",
    "u_question_part_correctly_count_dict = defaultdict(int)\n",
    "u_question_part_count_dict = defaultdict(int)\n",
    "u_question_tag1_correctly_count_dict = defaultdict(int)\n",
    "u_question_tag1_count_dict = defaultdict(int)\n",
    "u_question_tag2_correctly_count_dict = defaultdict(int)\n",
    "u_question_tag2_count_dict = defaultdict(int)\n",
    "u_prior_question_correctly_timestamp_dict = defaultdict(int)\n",
    "u_prior2_question_correctly_timestamp_dict = defaultdict(int)\n",
    "u_prior3_question_correctly_timestamp_dict = defaultdict(int)\n",
    "u_prior_question_timestamp_dict = defaultdict(int)\n",
    "u_prior2_question_timestamp_dict = defaultdict(int)\n",
    "u_prior3_question_timestamp_dict = defaultdict(int)\n",
    "u_prior4_question_timestamp_dict = defaultdict(int)\n",
    "u_prior_lecture_timestamp_dict = defaultdict(int)\n",
    "u_prior2_lecture_timestamp_dict = defaultdict(int)\n",
    "u_task_container_id_dict = defaultdict(int)\n",
    "u_prior_question_explanation_count_dict = defaultdict(int)\n",
    "u_prior_question_explanation_correctly_count_dict = defaultdict(int)\n",
    "u_question_listening_correctly_count_dict = defaultdict(int)\n",
    "u_question_reading_correctly_count_dict = defaultdict(int) \n",
    "u_question_listening_count_dict = defaultdict(int)\n",
    "u_question_reading_count_dict = defaultdict(int) \n",
    "u_question_incorrect_timestamp_dict = defaultdict(int)\n",
    "u_question_incorrect_timestamp2_dict = defaultdict(int)\n",
    "u_question_incorrect_timestamp3_dict = defaultdict(int)\n",
    "u_question_seen_dict = defaultdict(int)\n",
    "u_question_part_correct_timestamp_dict = defaultdict(int)\n",
    "u_question_part_incorrect_timestamp_dict = defaultdict(int)\n",
    "\n",
    "train = add_user_feats(train, \n",
    "                       u_answered_correctly_count_dict, u_answered_count_dict,\n",
    "                       u_question_part_correctly_count_dict, u_question_part_count_dict,\n",
    "                       u_question_tag1_correctly_count_dict, u_question_tag1_count_dict,\n",
    "                       u_prior_question_correctly_timestamp_dict, u_prior2_question_correctly_timestamp_dict, \n",
    "                       u_prior3_question_correctly_timestamp_dict, u_prior_question_timestamp_dict,\n",
    "                       u_prior2_question_timestamp_dict, u_prior3_question_timestamp_dict,\n",
    "                       u_prior4_question_timestamp_dict, u_prior_lecture_timestamp_dict,\n",
    "                       u_prior2_lecture_timestamp_dict, u_task_container_id_dict,\n",
    "                       u_prior_question_explanation_count_dict, u_prior_question_explanation_correctly_count_dict,\n",
    "                       u_question_listening_correctly_count_dict, u_question_reading_correctly_count_dict,\n",
    "                       u_question_listening_count_dict, u_question_reading_count_dict,\n",
    "                       u_question_incorrect_timestamp_dict, u_question_incorrect_timestamp2_dict,\n",
    "                       u_question_incorrect_timestamp3_dict, u_question_seen_dict,\n",
    "                       u_question_part_correct_timestamp_dict, u_question_part_incorrect_timestamp_dict,\n",
    "                      )\n",
    "\n",
    "valid = add_user_feats(valid, \n",
    "                       u_answered_correctly_count_dict, u_answered_count_dict,\n",
    "                       u_question_part_correctly_count_dict, u_question_part_count_dict,\n",
    "                       u_question_tag1_correctly_count_dict, u_question_tag1_count_dict,\n",
    "                       u_prior_question_correctly_timestamp_dict, u_prior2_question_correctly_timestamp_dict,\n",
    "                       u_prior3_question_correctly_timestamp_dict, u_prior_question_timestamp_dict,\n",
    "                       u_prior2_question_timestamp_dict, u_prior3_question_timestamp_dict,\n",
    "                       u_prior4_question_timestamp_dict, u_prior_lecture_timestamp_dict,\n",
    "                       u_prior2_lecture_timestamp_dict, u_task_container_id_dict,\n",
    "                       u_prior_question_explanation_count_dict, u_prior_question_explanation_correctly_count_dict,\n",
    "                       u_question_listening_correctly_count_dict, u_question_reading_correctly_count_dict,\n",
    "                       u_question_listening_count_dict, u_question_reading_count_dict,\n",
    "                       u_question_incorrect_timestamp_dict, u_question_incorrect_timestamp2_dict,\n",
    "                       u_question_incorrect_timestamp3_dict, u_question_seen_dict,\n",
    "                       u_question_part_correct_timestamp_dict, u_question_part_incorrect_timestamp_dict,\n",
    "                      )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "25340.273"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "prior_question_elapsed_time_mean = train.prior_question_elapsed_time.dropna().values.mean()\n",
    "prior_question_elapsed_time_mean"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "# fill with mean value for prior_question_elapsed_time\n",
    "train['prior_question_elapsed_time'] = train.prior_question_elapsed_time.fillna(prior_question_elapsed_time_mean)\n",
    "valid['prior_question_elapsed_time'] = valid.prior_question_elapsed_time.fillna(prior_question_elapsed_time_mean)\n",
    "\n",
    "# changing dtype to avoid lightgbm error\n",
    "# train['prior_question_had_explanation'] = train.prior_question_had_explanation.fillna(False).astype('int8')\n",
    "# valid['prior_question_had_explanation'] = valid.prior_question_had_explanation.fillna(False).astype('int8')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>row_id</th>\n",
       "      <th>user_id</th>\n",
       "      <th>content_id</th>\n",
       "      <th>content_type_id</th>\n",
       "      <th>task_container_id</th>\n",
       "      <th>answered_correctly</th>\n",
       "      <th>prior_question_elapsed_time</th>\n",
       "      <th>prior_question_had_explanation</th>\n",
       "      <th>timestamp</th>\n",
       "      <th>answered_correctly_avg_c</th>\n",
       "      <th>answered_correctly_var_c</th>\n",
       "      <th>lecture_part</th>\n",
       "      <th>part</th>\n",
       "      <th>question_asked</th>\n",
       "      <th>right_answers</th>\n",
       "      <th>bundle_size</th>\n",
       "      <th>bundle_rignt_answers</th>\n",
       "      <th>bundle_questions_asked</th>\n",
       "      <th>bundle_accuracy</th>\n",
       "      <th>part_rignt_answers</th>\n",
       "      <th>part_questions_asked</th>\n",
       "      <th>part_accuracy</th>\n",
       "      <th>LorR</th>\n",
       "      <th>tag1</th>\n",
       "      <th>tag1_answered_correctly_mean</th>\n",
       "      <th>tag1_answered_correctly_var</th>\n",
       "      <th>tags_emb_0</th>\n",
       "      <th>tags_emb_1</th>\n",
       "      <th>c_question_count_percent</th>\n",
       "      <th>c_question_part_percent</th>\n",
       "      <th>c_question_tag1_percent</th>\n",
       "      <th>c_question_unique_users_seen</th>\n",
       "      <th>u_answered_correctly_count</th>\n",
       "      <th>u_answered_count</th>\n",
       "      <th>u_question_part_correctly_count</th>\n",
       "      <th>u_question_part_count</th>\n",
       "      <th>u_question_tag1_correctly_count</th>\n",
       "      <th>u_question_tag1_count</th>\n",
       "      <th>u_prior_question_correctly_timestamp_diff</th>\n",
       "      <th>u_prior_question_correctly_timestamp_diff2</th>\n",
       "      <th>u_prior_question_correctly_timestamp_diff3</th>\n",
       "      <th>u_prior_question_timestamp_diff</th>\n",
       "      <th>u_prior_question_timestamp_diff2</th>\n",
       "      <th>u_prior_question_timestamp_diff3</th>\n",
       "      <th>u_prior_question_timestamp_diff4</th>\n",
       "      <th>u_prior_lecture_timestamp_diff</th>\n",
       "      <th>u_prior2_lecture_timestamp_diff</th>\n",
       "      <th>u_task_container_id_diff</th>\n",
       "      <th>u_prior_question_explanation_count</th>\n",
       "      <th>u_prior_question_explanation_correctly_count</th>\n",
       "      <th>u_question_listening_correctly_count</th>\n",
       "      <th>u_question_reading_correctly_count</th>\n",
       "      <th>u_question_listening_count</th>\n",
       "      <th>u_question_reading_count</th>\n",
       "      <th>u_question_incorrect_timestamp_diff</th>\n",
       "      <th>u_question_incorrect_timestamp_diff2</th>\n",
       "      <th>u_question_incorrect_timestamp_diff3</th>\n",
       "      <th>u_question_seen</th>\n",
       "      <th>u_question_nunique</th>\n",
       "      <th>u_question_part_correct_timestamp_diff</th>\n",
       "      <th>u_question_part_incorrect_timestamp_diff</th>\n",
       "      <th>uc_question_timestamp_diff</th>\n",
       "      <th>u_answered_correctly_avg</th>\n",
       "      <th>u_question_part_correctly_avg</th>\n",
       "      <th>u_question_tag1_correctly_avg</th>\n",
       "      <th>u_prior_question_timestamp_diff_2_1</th>\n",
       "      <th>u_prior_question_timestamp_diff_3_2</th>\n",
       "      <th>u_prior_question_timestamp_diff_4_3</th>\n",
       "      <th>u_question_timestamp_diff</th>\n",
       "      <th>u_question_timestamp_diff2</th>\n",
       "      <th>u_question_timestamp_diff3</th>\n",
       "      <th>u_listening_reading_ratio</th>\n",
       "      <th>u_listening_correctly_avg</th>\n",
       "      <th>u_reading_correctly_avg</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>32933156</td>\n",
       "      <td>705741139</td>\n",
       "      <td>128</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.966887</td>\n",
       "      <td>0.032018</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>19056</td>\n",
       "      <td>18425.0</td>\n",
       "      <td>1</td>\n",
       "      <td>18425.0</td>\n",
       "      <td>19056</td>\n",
       "      <td>0.966887</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>0.850586</td>\n",
       "      <td>0.017212</td>\n",
       "      <td>-0.007187</td>\n",
       "      <td>0.090353</td>\n",
       "      <td>0.000198</td>\n",
       "      <td>0.075093</td>\n",
       "      <td>0.041839</td>\n",
       "      <td>19260</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>32933157</td>\n",
       "      <td>705741139</td>\n",
       "      <td>7860</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>16000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>20666</td>\n",
       "      <td>0.954720</td>\n",
       "      <td>0.043232</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>21135</td>\n",
       "      <td>20178.0</td>\n",
       "      <td>1</td>\n",
       "      <td>20178.0</td>\n",
       "      <td>21135</td>\n",
       "      <td>0.954720</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>0.850586</td>\n",
       "      <td>0.017212</td>\n",
       "      <td>-0.038487</td>\n",
       "      <td>-0.097879</td>\n",
       "      <td>0.000219</td>\n",
       "      <td>0.075093</td>\n",
       "      <td>0.041839</td>\n",
       "      <td>21261</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>20666</td>\n",
       "      <td>20666</td>\n",
       "      <td>20666</td>\n",
       "      <td>20666</td>\n",
       "      <td>20666</td>\n",
       "      <td>20666</td>\n",
       "      <td>20666</td>\n",
       "      <td>20666</td>\n",
       "      <td>20666</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>20666</td>\n",
       "      <td>20666</td>\n",
       "      <td>20666</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>20666</td>\n",
       "      <td>20666</td>\n",
       "      <td>20667</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-16000.0</td>\n",
       "      <td>-16000.0</td>\n",
       "      <td>-16000.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>32933158</td>\n",
       "      <td>705741139</td>\n",
       "      <td>7922</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>19000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>39172</td>\n",
       "      <td>0.953161</td>\n",
       "      <td>0.044647</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>18852</td>\n",
       "      <td>17969.0</td>\n",
       "      <td>1</td>\n",
       "      <td>17969.0</td>\n",
       "      <td>18852</td>\n",
       "      <td>0.953161</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>0.850586</td>\n",
       "      <td>0.017212</td>\n",
       "      <td>-0.007187</td>\n",
       "      <td>0.090353</td>\n",
       "      <td>0.000196</td>\n",
       "      <td>0.075093</td>\n",
       "      <td>0.041839</td>\n",
       "      <td>19065</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>18506</td>\n",
       "      <td>39172</td>\n",
       "      <td>39172</td>\n",
       "      <td>18506</td>\n",
       "      <td>39172</td>\n",
       "      <td>39172</td>\n",
       "      <td>39172</td>\n",
       "      <td>39172</td>\n",
       "      <td>39172</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>39172</td>\n",
       "      <td>39172</td>\n",
       "      <td>39172</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>18506</td>\n",
       "      <td>39172</td>\n",
       "      <td>39173</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>20666</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1666.0</td>\n",
       "      <td>-19000.0</td>\n",
       "      <td>-19000.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>32933159</td>\n",
       "      <td>705741139</td>\n",
       "      <td>156</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>17000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>58207</td>\n",
       "      <td>0.932764</td>\n",
       "      <td>0.062719</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>21878</td>\n",
       "      <td>20407.0</td>\n",
       "      <td>1</td>\n",
       "      <td>20407.0</td>\n",
       "      <td>21878</td>\n",
       "      <td>0.932764</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>0.850586</td>\n",
       "      <td>0.017212</td>\n",
       "      <td>0.106040</td>\n",
       "      <td>-0.152273</td>\n",
       "      <td>0.000227</td>\n",
       "      <td>0.075093</td>\n",
       "      <td>0.041839</td>\n",
       "      <td>21908</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>19035</td>\n",
       "      <td>37541</td>\n",
       "      <td>58207</td>\n",
       "      <td>19035</td>\n",
       "      <td>37541</td>\n",
       "      <td>58207</td>\n",
       "      <td>58207</td>\n",
       "      <td>58207</td>\n",
       "      <td>58207</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>58207</td>\n",
       "      <td>58207</td>\n",
       "      <td>58207</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>19035</td>\n",
       "      <td>58207</td>\n",
       "      <td>58208</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>18506</td>\n",
       "      <td>20666</td>\n",
       "      <td>0</td>\n",
       "      <td>1506.0</td>\n",
       "      <td>3666.0</td>\n",
       "      <td>-17000.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.750000</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>32933160</td>\n",
       "      <td>705741139</td>\n",
       "      <td>51</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>17000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>75779</td>\n",
       "      <td>0.930161</td>\n",
       "      <td>0.064964</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>23454</td>\n",
       "      <td>21816.0</td>\n",
       "      <td>1</td>\n",
       "      <td>21816.0</td>\n",
       "      <td>23454</td>\n",
       "      <td>0.930161</td>\n",
       "      <td>5425569.0</td>\n",
       "      <td>7276830</td>\n",
       "      <td>0.745595</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>0.850586</td>\n",
       "      <td>0.017212</td>\n",
       "      <td>-0.220193</td>\n",
       "      <td>0.188346</td>\n",
       "      <td>0.000243</td>\n",
       "      <td>0.075093</td>\n",
       "      <td>0.041839</td>\n",
       "      <td>23373</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>17572</td>\n",
       "      <td>36607</td>\n",
       "      <td>55113</td>\n",
       "      <td>17572</td>\n",
       "      <td>36607</td>\n",
       "      <td>55113</td>\n",
       "      <td>75779</td>\n",
       "      <td>75779</td>\n",
       "      <td>75779</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>75779</td>\n",
       "      <td>75779</td>\n",
       "      <td>75779</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>17572</td>\n",
       "      <td>75779</td>\n",
       "      <td>75780</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>19035</td>\n",
       "      <td>18506</td>\n",
       "      <td>20666</td>\n",
       "      <td>2035.0</td>\n",
       "      <td>1506.0</td>\n",
       "      <td>3666.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.800000</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     row_id    user_id  content_id  content_type_id  task_container_id  \\\n",
       "0  32933156  705741139         128                0                  0   \n",
       "1  32933157  705741139        7860                0                  1   \n",
       "2  32933158  705741139        7922                0                  2   \n",
       "3  32933159  705741139         156                0                  3   \n",
       "4  32933160  705741139          51                0                  4   \n",
       "\n",
       "   answered_correctly  prior_question_elapsed_time  \\\n",
       "0                   1                         -1.0   \n",
       "1                   1                      16000.0   \n",
       "2                   1                      19000.0   \n",
       "3                   1                      17000.0   \n",
       "4                   1                      17000.0   \n",
       "\n",
       "   prior_question_had_explanation  timestamp  answered_correctly_avg_c  \\\n",
       "0                               0          0                  0.966887   \n",
       "1                               0      20666                  0.954720   \n",
       "2                               0      39172                  0.953161   \n",
       "3                               0      58207                  0.932764   \n",
       "4                               0      75779                  0.930161   \n",
       "\n",
       "   answered_correctly_var_c  lecture_part  part  question_asked  \\\n",
       "0                  0.032018          -1.0     1           19056   \n",
       "1                  0.043232          -1.0     1           21135   \n",
       "2                  0.044647          -1.0     1           18852   \n",
       "3                  0.062719          -1.0     1           21878   \n",
       "4                  0.064964          -1.0     1           23454   \n",
       "\n",
       "   right_answers  bundle_size  bundle_rignt_answers  bundle_questions_asked  \\\n",
       "0        18425.0            1               18425.0                   19056   \n",
       "1        20178.0            1               20178.0                   21135   \n",
       "2        17969.0            1               17969.0                   18852   \n",
       "3        20407.0            1               20407.0                   21878   \n",
       "4        21816.0            1               21816.0                   23454   \n",
       "\n",
       "   bundle_accuracy  part_rignt_answers  part_questions_asked  part_accuracy  \\\n",
       "0         0.966887           5425569.0               7276830       0.745595   \n",
       "1         0.954720           5425569.0               7276830       0.745595   \n",
       "2         0.953161           5425569.0               7276830       0.745595   \n",
       "3         0.932764           5425569.0               7276830       0.745595   \n",
       "4         0.930161           5425569.0               7276830       0.745595   \n",
       "\n",
       "   LorR  tag1  tag1_answered_correctly_mean  tag1_answered_correctly_var  \\\n",
       "0     0   131                      0.850586                     0.017212   \n",
       "1     0   131                      0.850586                     0.017212   \n",
       "2     0   131                      0.850586                     0.017212   \n",
       "3     0   131                      0.850586                     0.017212   \n",
       "4     0   131                      0.850586                     0.017212   \n",
       "\n",
       "   tags_emb_0  tags_emb_1  c_question_count_percent  c_question_part_percent  \\\n",
       "0   -0.007187    0.090353                  0.000198                 0.075093   \n",
       "1   -0.038487   -0.097879                  0.000219                 0.075093   \n",
       "2   -0.007187    0.090353                  0.000196                 0.075093   \n",
       "3    0.106040   -0.152273                  0.000227                 0.075093   \n",
       "4   -0.220193    0.188346                  0.000243                 0.075093   \n",
       "\n",
       "   c_question_tag1_percent  c_question_unique_users_seen  \\\n",
       "0                 0.041839                         19260   \n",
       "1                 0.041839                         21261   \n",
       "2                 0.041839                         19065   \n",
       "3                 0.041839                         21908   \n",
       "4                 0.041839                         23373   \n",
       "\n",
       "   u_answered_correctly_count  u_answered_count  \\\n",
       "0                           0                 0   \n",
       "1                           1                 1   \n",
       "2                           2                 2   \n",
       "3                           3                 3   \n",
       "4                           4                 4   \n",
       "\n",
       "   u_question_part_correctly_count  u_question_part_count  \\\n",
       "0                                0                      0   \n",
       "1                                1                      1   \n",
       "2                                2                      2   \n",
       "3                                3                      3   \n",
       "4                                4                      4   \n",
       "\n",
       "   u_question_tag1_correctly_count  u_question_tag1_count  \\\n",
       "0                                0                      0   \n",
       "1                                1                      1   \n",
       "2                                2                      2   \n",
       "3                                3                      3   \n",
       "4                                4                      4   \n",
       "\n",
       "   u_prior_question_correctly_timestamp_diff  \\\n",
       "0                                          0   \n",
       "1                                      20666   \n",
       "2                                      18506   \n",
       "3                                      19035   \n",
       "4                                      17572   \n",
       "\n",
       "   u_prior_question_correctly_timestamp_diff2  \\\n",
       "0                                           0   \n",
       "1                                       20666   \n",
       "2                                       39172   \n",
       "3                                       37541   \n",
       "4                                       36607   \n",
       "\n",
       "   u_prior_question_correctly_timestamp_diff3  \\\n",
       "0                                           0   \n",
       "1                                       20666   \n",
       "2                                       39172   \n",
       "3                                       58207   \n",
       "4                                       55113   \n",
       "\n",
       "   u_prior_question_timestamp_diff  u_prior_question_timestamp_diff2  \\\n",
       "0                                0                                 0   \n",
       "1                            20666                             20666   \n",
       "2                            18506                             39172   \n",
       "3                            19035                             37541   \n",
       "4                            17572                             36607   \n",
       "\n",
       "   u_prior_question_timestamp_diff3  u_prior_question_timestamp_diff4  \\\n",
       "0                                 0                                 0   \n",
       "1                             20666                             20666   \n",
       "2                             39172                             39172   \n",
       "3                             58207                             58207   \n",
       "4                             55113                             75779   \n",
       "\n",
       "   u_prior_lecture_timestamp_diff  u_prior2_lecture_timestamp_diff  \\\n",
       "0                               0                                0   \n",
       "1                           20666                            20666   \n",
       "2                           39172                            39172   \n",
       "3                           58207                            58207   \n",
       "4                           75779                            75779   \n",
       "\n",
       "   u_task_container_id_diff  u_prior_question_explanation_count  \\\n",
       "0                         0                                   0   \n",
       "1                         1                                   0   \n",
       "2                         1                                   0   \n",
       "3                         1                                   0   \n",
       "4                         1                                   0   \n",
       "\n",
       "   u_prior_question_explanation_correctly_count  \\\n",
       "0                                             0   \n",
       "1                                             1   \n",
       "2                                             2   \n",
       "3                                             3   \n",
       "4                                             4   \n",
       "\n",
       "   u_question_listening_correctly_count  u_question_reading_correctly_count  \\\n",
       "0                                     0                                   0   \n",
       "1                                     1                                   0   \n",
       "2                                     2                                   0   \n",
       "3                                     3                                   0   \n",
       "4                                     4                                   0   \n",
       "\n",
       "   u_question_listening_count  u_question_reading_count  \\\n",
       "0                           0                         0   \n",
       "1                           1                         0   \n",
       "2                           2                         0   \n",
       "3                           3                         0   \n",
       "4                           4                         0   \n",
       "\n",
       "   u_question_incorrect_timestamp_diff  u_question_incorrect_timestamp_diff2  \\\n",
       "0                                    0                                     0   \n",
       "1                                20666                                 20666   \n",
       "2                                39172                                 39172   \n",
       "3                                58207                                 58207   \n",
       "4                                75779                                 75779   \n",
       "\n",
       "   u_question_incorrect_timestamp_diff3  u_question_seen  u_question_nunique  \\\n",
       "0                                     0                0                   0   \n",
       "1                                 20666                0                   1   \n",
       "2                                 39172                0                   2   \n",
       "3                                 58207                0                   3   \n",
       "4                                 75779                0                   4   \n",
       "\n",
       "   u_question_part_correct_timestamp_diff  \\\n",
       "0                                       0   \n",
       "1                                   20666   \n",
       "2                                   18506   \n",
       "3                                   19035   \n",
       "4                                   17572   \n",
       "\n",
       "   u_question_part_incorrect_timestamp_diff  uc_question_timestamp_diff  \\\n",
       "0                                         0                           1   \n",
       "1                                     20666                       20667   \n",
       "2                                     39172                       39173   \n",
       "3                                     58207                       58208   \n",
       "4                                     75779                       75780   \n",
       "\n",
       "   u_answered_correctly_avg  u_question_part_correctly_avg  \\\n",
       "0                       NaN                            NaN   \n",
       "1                       1.0                            1.0   \n",
       "2                       1.0                            1.0   \n",
       "3                       1.0                            1.0   \n",
       "4                       1.0                            1.0   \n",
       "\n",
       "   u_question_tag1_correctly_avg  u_prior_question_timestamp_diff_2_1  \\\n",
       "0                            NaN                                    0   \n",
       "1                            1.0                                    0   \n",
       "2                            1.0                                20666   \n",
       "3                            1.0                                18506   \n",
       "4                            1.0                                19035   \n",
       "\n",
       "   u_prior_question_timestamp_diff_3_2  u_prior_question_timestamp_diff_4_3  \\\n",
       "0                                    0                                    0   \n",
       "1                                    0                                    0   \n",
       "2                                    0                                    0   \n",
       "3                                20666                                    0   \n",
       "4                                18506                                20666   \n",
       "\n",
       "   u_question_timestamp_diff  u_question_timestamp_diff2  \\\n",
       "0                        1.0                         1.0   \n",
       "1                   -16000.0                    -16000.0   \n",
       "2                     1666.0                    -19000.0   \n",
       "3                     1506.0                      3666.0   \n",
       "4                     2035.0                      1506.0   \n",
       "\n",
       "   u_question_timestamp_diff3  u_listening_reading_ratio  \\\n",
       "0                         1.0                        0.0   \n",
       "1                    -16000.0                        1.0   \n",
       "2                    -19000.0                        2.0   \n",
       "3                    -17000.0                        3.0   \n",
       "4                      3666.0                        4.0   \n",
       "\n",
       "   u_listening_correctly_avg  u_reading_correctly_avg  \n",
       "0                   0.000000                      0.0  \n",
       "1                   0.500000                      0.0  \n",
       "2                   0.666667                      0.0  \n",
       "3                   0.750000                      0.0  \n",
       "4                   0.800000                      0.0  "
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>row_id</th>\n",
       "      <th>user_id</th>\n",
       "      <th>content_id</th>\n",
       "      <th>content_type_id</th>\n",
       "      <th>task_container_id</th>\n",
       "      <th>answered_correctly</th>\n",
       "      <th>prior_question_elapsed_time</th>\n",
       "      <th>prior_question_had_explanation</th>\n",
       "      <th>timestamp</th>\n",
       "      <th>answered_correctly_avg_c</th>\n",
       "      <th>answered_correctly_var_c</th>\n",
       "      <th>lecture_part</th>\n",
       "      <th>part</th>\n",
       "      <th>question_asked</th>\n",
       "      <th>right_answers</th>\n",
       "      <th>bundle_size</th>\n",
       "      <th>bundle_rignt_answers</th>\n",
       "      <th>bundle_questions_asked</th>\n",
       "      <th>bundle_accuracy</th>\n",
       "      <th>part_rignt_answers</th>\n",
       "      <th>part_questions_asked</th>\n",
       "      <th>part_accuracy</th>\n",
       "      <th>LorR</th>\n",
       "      <th>tag1</th>\n",
       "      <th>tag1_answered_correctly_mean</th>\n",
       "      <th>tag1_answered_correctly_var</th>\n",
       "      <th>tags_emb_0</th>\n",
       "      <th>tags_emb_1</th>\n",
       "      <th>c_question_count_percent</th>\n",
       "      <th>c_question_part_percent</th>\n",
       "      <th>c_question_tag1_percent</th>\n",
       "      <th>c_question_unique_users_seen</th>\n",
       "      <th>u_answered_correctly_count</th>\n",
       "      <th>u_answered_count</th>\n",
       "      <th>u_question_part_correctly_count</th>\n",
       "      <th>u_question_part_count</th>\n",
       "      <th>u_question_tag1_correctly_count</th>\n",
       "      <th>u_question_tag1_count</th>\n",
       "      <th>u_prior_question_correctly_timestamp_diff</th>\n",
       "      <th>u_prior_question_correctly_timestamp_diff2</th>\n",
       "      <th>u_prior_question_correctly_timestamp_diff3</th>\n",
       "      <th>u_prior_question_timestamp_diff</th>\n",
       "      <th>u_prior_question_timestamp_diff2</th>\n",
       "      <th>u_prior_question_timestamp_diff3</th>\n",
       "      <th>u_prior_question_timestamp_diff4</th>\n",
       "      <th>u_prior_lecture_timestamp_diff</th>\n",
       "      <th>u_prior2_lecture_timestamp_diff</th>\n",
       "      <th>u_task_container_id_diff</th>\n",
       "      <th>u_prior_question_explanation_count</th>\n",
       "      <th>u_prior_question_explanation_correctly_count</th>\n",
       "      <th>u_question_listening_correctly_count</th>\n",
       "      <th>u_question_reading_correctly_count</th>\n",
       "      <th>u_question_listening_count</th>\n",
       "      <th>u_question_reading_count</th>\n",
       "      <th>u_question_incorrect_timestamp_diff</th>\n",
       "      <th>u_question_incorrect_timestamp_diff2</th>\n",
       "      <th>u_question_incorrect_timestamp_diff3</th>\n",
       "      <th>u_question_seen</th>\n",
       "      <th>u_question_nunique</th>\n",
       "      <th>u_question_part_correct_timestamp_diff</th>\n",
       "      <th>u_question_part_incorrect_timestamp_diff</th>\n",
       "      <th>uc_question_timestamp_diff</th>\n",
       "      <th>u_answered_correctly_avg</th>\n",
       "      <th>u_question_part_correctly_avg</th>\n",
       "      <th>u_question_tag1_correctly_avg</th>\n",
       "      <th>u_prior_question_timestamp_diff_2_1</th>\n",
       "      <th>u_prior_question_timestamp_diff_3_2</th>\n",
       "      <th>u_prior_question_timestamp_diff_4_3</th>\n",
       "      <th>u_question_timestamp_diff</th>\n",
       "      <th>u_question_timestamp_diff2</th>\n",
       "      <th>u_question_timestamp_diff3</th>\n",
       "      <th>u_listening_reading_ratio</th>\n",
       "      <th>u_listening_correctly_avg</th>\n",
       "      <th>u_reading_correctly_avg</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>45253</td>\n",
       "      <td>1186307</td>\n",
       "      <td>4451</td>\n",
       "      <td>0</td>\n",
       "      <td>370</td>\n",
       "      <td>1</td>\n",
       "      <td>10000.0</td>\n",
       "      <td>1</td>\n",
       "      <td>2146476402</td>\n",
       "      <td>0.511194</td>\n",
       "      <td>0.249882</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>5</td>\n",
       "      <td>32919</td>\n",
       "      <td>16828.0</td>\n",
       "      <td>1</td>\n",
       "      <td>16828.0</td>\n",
       "      <td>32919</td>\n",
       "      <td>0.511194</td>\n",
       "      <td>24363327.0</td>\n",
       "      <td>39909154</td>\n",
       "      <td>0.610470</td>\n",
       "      <td>1</td>\n",
       "      <td>175</td>\n",
       "      <td>0.562500</td>\n",
       "      <td>0.021606</td>\n",
       "      <td>-0.207543</td>\n",
       "      <td>0.186181</td>\n",
       "      <td>0.000340</td>\n",
       "      <td>0.412084</td>\n",
       "      <td>0.004473</td>\n",
       "      <td>26762</td>\n",
       "      <td>313</td>\n",
       "      <td>498</td>\n",
       "      <td>118</td>\n",
       "      <td>245</td>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>16898</td>\n",
       "      <td>151141</td>\n",
       "      <td>211368</td>\n",
       "      <td>16898</td>\n",
       "      <td>32512</td>\n",
       "      <td>82698</td>\n",
       "      <td>151141</td>\n",
       "      <td>2146476402</td>\n",
       "      <td>2146476402</td>\n",
       "      <td>1</td>\n",
       "      <td>480</td>\n",
       "      <td>313</td>\n",
       "      <td>62</td>\n",
       "      <td>251</td>\n",
       "      <td>88</td>\n",
       "      <td>410</td>\n",
       "      <td>32512</td>\n",
       "      <td>82698</td>\n",
       "      <td>238277</td>\n",
       "      <td>1</td>\n",
       "      <td>441</td>\n",
       "      <td>16898</td>\n",
       "      <td>32512</td>\n",
       "      <td>44938880</td>\n",
       "      <td>0.628514</td>\n",
       "      <td>0.481633</td>\n",
       "      <td>0.142857</td>\n",
       "      <td>15614</td>\n",
       "      <td>50186</td>\n",
       "      <td>68443</td>\n",
       "      <td>5614.0</td>\n",
       "      <td>40186.0</td>\n",
       "      <td>58443.0</td>\n",
       "      <td>0.246032</td>\n",
       "      <td>0.696629</td>\n",
       "      <td>0.610706</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>41038234</td>\n",
       "      <td>871981876</td>\n",
       "      <td>1278</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>26000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>87403</td>\n",
       "      <td>0.649971</td>\n",
       "      <td>0.227510</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>164278</td>\n",
       "      <td>106776.0</td>\n",
       "      <td>1</td>\n",
       "      <td>106776.0</td>\n",
       "      <td>164278</td>\n",
       "      <td>0.649971</td>\n",
       "      <td>12965289.0</td>\n",
       "      <td>18291408</td>\n",
       "      <td>0.708819</td>\n",
       "      <td>0</td>\n",
       "      <td>143</td>\n",
       "      <td>0.767090</td>\n",
       "      <td>0.017136</td>\n",
       "      <td>0.193045</td>\n",
       "      <td>0.165625</td>\n",
       "      <td>0.001724</td>\n",
       "      <td>0.188810</td>\n",
       "      <td>0.064636</td>\n",
       "      <td>165489</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>87403</td>\n",
       "      <td>87403</td>\n",
       "      <td>87403</td>\n",
       "      <td>22899</td>\n",
       "      <td>51871</td>\n",
       "      <td>87403</td>\n",
       "      <td>87403</td>\n",
       "      <td>87403</td>\n",
       "      <td>87403</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>22899</td>\n",
       "      <td>51871</td>\n",
       "      <td>87403</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>87403</td>\n",
       "      <td>87403</td>\n",
       "      <td>87404</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>28972</td>\n",
       "      <td>35532</td>\n",
       "      <td>0</td>\n",
       "      <td>2972.0</td>\n",
       "      <td>9532.0</td>\n",
       "      <td>-26000.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>56086085</td>\n",
       "      <td>1189905495</td>\n",
       "      <td>5604</td>\n",
       "      <td>0</td>\n",
       "      <td>1991</td>\n",
       "      <td>0</td>\n",
       "      <td>3000.0</td>\n",
       "      <td>1</td>\n",
       "      <td>41201916980</td>\n",
       "      <td>0.667658</td>\n",
       "      <td>0.221909</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>5</td>\n",
       "      <td>12108</td>\n",
       "      <td>8084.0</td>\n",
       "      <td>1</td>\n",
       "      <td>8084.0</td>\n",
       "      <td>12108</td>\n",
       "      <td>0.667658</td>\n",
       "      <td>24363327.0</td>\n",
       "      <td>39909154</td>\n",
       "      <td>0.610470</td>\n",
       "      <td>1</td>\n",
       "      <td>48</td>\n",
       "      <td>0.644043</td>\n",
       "      <td>0.019333</td>\n",
       "      <td>0.111092</td>\n",
       "      <td>-0.185174</td>\n",
       "      <td>0.000125</td>\n",
       "      <td>0.412084</td>\n",
       "      <td>0.006243</td>\n",
       "      <td>10843</td>\n",
       "      <td>1375</td>\n",
       "      <td>2102</td>\n",
       "      <td>712</td>\n",
       "      <td>1089</td>\n",
       "      <td>12</td>\n",
       "      <td>17</td>\n",
       "      <td>27258</td>\n",
       "      <td>62834</td>\n",
       "      <td>118494</td>\n",
       "      <td>27258</td>\n",
       "      <td>39668</td>\n",
       "      <td>62834</td>\n",
       "      <td>77536</td>\n",
       "      <td>-1747755980</td>\n",
       "      <td>-1747755980</td>\n",
       "      <td>1</td>\n",
       "      <td>2091</td>\n",
       "      <td>1375</td>\n",
       "      <td>610</td>\n",
       "      <td>765</td>\n",
       "      <td>917</td>\n",
       "      <td>1185</td>\n",
       "      <td>39668</td>\n",
       "      <td>77536</td>\n",
       "      <td>92497</td>\n",
       "      <td>0</td>\n",
       "      <td>1835</td>\n",
       "      <td>27258</td>\n",
       "      <td>39668</td>\n",
       "      <td>-1747755979</td>\n",
       "      <td>0.654139</td>\n",
       "      <td>0.653811</td>\n",
       "      <td>0.705882</td>\n",
       "      <td>12410</td>\n",
       "      <td>23166</td>\n",
       "      <td>14702</td>\n",
       "      <td>9410.0</td>\n",
       "      <td>20166.0</td>\n",
       "      <td>11702.0</td>\n",
       "      <td>0.796345</td>\n",
       "      <td>0.664488</td>\n",
       "      <td>0.645025</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>91995999</td>\n",
       "      <td>1951897185</td>\n",
       "      <td>8534</td>\n",
       "      <td>0</td>\n",
       "      <td>504</td>\n",
       "      <td>0</td>\n",
       "      <td>47000.0</td>\n",
       "      <td>1</td>\n",
       "      <td>4916060626</td>\n",
       "      <td>0.424576</td>\n",
       "      <td>0.244340</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>5</td>\n",
       "      <td>8618</td>\n",
       "      <td>3659.0</td>\n",
       "      <td>1</td>\n",
       "      <td>3659.0</td>\n",
       "      <td>8618</td>\n",
       "      <td>0.424576</td>\n",
       "      <td>24363327.0</td>\n",
       "      <td>39909154</td>\n",
       "      <td>0.610470</td>\n",
       "      <td>1</td>\n",
       "      <td>24</td>\n",
       "      <td>0.529785</td>\n",
       "      <td>0.030197</td>\n",
       "      <td>0.228403</td>\n",
       "      <td>-0.249632</td>\n",
       "      <td>0.000089</td>\n",
       "      <td>0.412084</td>\n",
       "      <td>0.002585</td>\n",
       "      <td>7566</td>\n",
       "      <td>279</td>\n",
       "      <td>500</td>\n",
       "      <td>187</td>\n",
       "      <td>325</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>210001732</td>\n",
       "      <td>210061483</td>\n",
       "      <td>210192481</td>\n",
       "      <td>57387</td>\n",
       "      <td>210001732</td>\n",
       "      <td>210061483</td>\n",
       "      <td>210129010</td>\n",
       "      <td>621093330</td>\n",
       "      <td>621093330</td>\n",
       "      <td>1</td>\n",
       "      <td>492</td>\n",
       "      <td>279</td>\n",
       "      <td>92</td>\n",
       "      <td>187</td>\n",
       "      <td>175</td>\n",
       "      <td>325</td>\n",
       "      <td>57387</td>\n",
       "      <td>210129010</td>\n",
       "      <td>210522304</td>\n",
       "      <td>0</td>\n",
       "      <td>408</td>\n",
       "      <td>235586768</td>\n",
       "      <td>57387</td>\n",
       "      <td>621093331</td>\n",
       "      <td>0.558000</td>\n",
       "      <td>0.575385</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>209944345</td>\n",
       "      <td>59751</td>\n",
       "      <td>67527</td>\n",
       "      <td>209897344.0</td>\n",
       "      <td>12751.0</td>\n",
       "      <td>20527.0</td>\n",
       "      <td>0.489362</td>\n",
       "      <td>0.522727</td>\n",
       "      <td>0.573620</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>99237056</td>\n",
       "      <td>2105538840</td>\n",
       "      <td>5631</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>21000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>33806</td>\n",
       "      <td>0.474163</td>\n",
       "      <td>0.249358</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>5</td>\n",
       "      <td>9676</td>\n",
       "      <td>4588.0</td>\n",
       "      <td>1</td>\n",
       "      <td>4588.0</td>\n",
       "      <td>9676</td>\n",
       "      <td>0.474163</td>\n",
       "      <td>24363327.0</td>\n",
       "      <td>39909154</td>\n",
       "      <td>0.610470</td>\n",
       "      <td>1</td>\n",
       "      <td>53</td>\n",
       "      <td>0.654785</td>\n",
       "      <td>0.032288</td>\n",
       "      <td>-0.139049</td>\n",
       "      <td>-0.157287</td>\n",
       "      <td>0.000100</td>\n",
       "      <td>0.412084</td>\n",
       "      <td>0.029988</td>\n",
       "      <td>7853</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>33806</td>\n",
       "      <td>33806</td>\n",
       "      <td>33806</td>\n",
       "      <td>33806</td>\n",
       "      <td>33806</td>\n",
       "      <td>33806</td>\n",
       "      <td>33806</td>\n",
       "      <td>33806</td>\n",
       "      <td>33806</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>33806</td>\n",
       "      <td>33806</td>\n",
       "      <td>33806</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>33806</td>\n",
       "      <td>33806</td>\n",
       "      <td>33807</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-21000.0</td>\n",
       "      <td>-21000.0</td>\n",
       "      <td>-21000.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.500000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     row_id     user_id  content_id  content_type_id  task_container_id  \\\n",
       "0     45253     1186307        4451                0                370   \n",
       "1  41038234   871981876        1278                0                  3   \n",
       "2  56086085  1189905495        5604                0               1991   \n",
       "3  91995999  1951897185        8534                0                504   \n",
       "4  99237056  2105538840        5631                0                  1   \n",
       "\n",
       "   answered_correctly  prior_question_elapsed_time  \\\n",
       "0                   1                      10000.0   \n",
       "1                   0                      26000.0   \n",
       "2                   0                       3000.0   \n",
       "3                   0                      47000.0   \n",
       "4                   1                      21000.0   \n",
       "\n",
       "   prior_question_had_explanation    timestamp  answered_correctly_avg_c  \\\n",
       "0                               1   2146476402                  0.511194   \n",
       "1                               0        87403                  0.649971   \n",
       "2                               1  41201916980                  0.667658   \n",
       "3                               1   4916060626                  0.424576   \n",
       "4                               0        33806                  0.474163   \n",
       "\n",
       "   answered_correctly_var_c  lecture_part  part  question_asked  \\\n",
       "0                  0.249882          -1.0     5           32919   \n",
       "1                  0.227510          -1.0     2          164278   \n",
       "2                  0.221909          -1.0     5           12108   \n",
       "3                  0.244340          -1.0     5            8618   \n",
       "4                  0.249358          -1.0     5            9676   \n",
       "\n",
       "   right_answers  bundle_size  bundle_rignt_answers  bundle_questions_asked  \\\n",
       "0        16828.0            1               16828.0                   32919   \n",
       "1       106776.0            1              106776.0                  164278   \n",
       "2         8084.0            1                8084.0                   12108   \n",
       "3         3659.0            1                3659.0                    8618   \n",
       "4         4588.0            1                4588.0                    9676   \n",
       "\n",
       "   bundle_accuracy  part_rignt_answers  part_questions_asked  part_accuracy  \\\n",
       "0         0.511194          24363327.0              39909154       0.610470   \n",
       "1         0.649971          12965289.0              18291408       0.708819   \n",
       "2         0.667658          24363327.0              39909154       0.610470   \n",
       "3         0.424576          24363327.0              39909154       0.610470   \n",
       "4         0.474163          24363327.0              39909154       0.610470   \n",
       "\n",
       "   LorR  tag1  tag1_answered_correctly_mean  tag1_answered_correctly_var  \\\n",
       "0     1   175                      0.562500                     0.021606   \n",
       "1     0   143                      0.767090                     0.017136   \n",
       "2     1    48                      0.644043                     0.019333   \n",
       "3     1    24                      0.529785                     0.030197   \n",
       "4     1    53                      0.654785                     0.032288   \n",
       "\n",
       "   tags_emb_0  tags_emb_1  c_question_count_percent  c_question_part_percent  \\\n",
       "0   -0.207543    0.186181                  0.000340                 0.412084   \n",
       "1    0.193045    0.165625                  0.001724                 0.188810   \n",
       "2    0.111092   -0.185174                  0.000125                 0.412084   \n",
       "3    0.228403   -0.249632                  0.000089                 0.412084   \n",
       "4   -0.139049   -0.157287                  0.000100                 0.412084   \n",
       "\n",
       "   c_question_tag1_percent  c_question_unique_users_seen  \\\n",
       "0                 0.004473                         26762   \n",
       "1                 0.064636                        165489   \n",
       "2                 0.006243                         10843   \n",
       "3                 0.002585                          7566   \n",
       "4                 0.029988                          7853   \n",
       "\n",
       "   u_answered_correctly_count  u_answered_count  \\\n",
       "0                         313               498   \n",
       "1                           0                 3   \n",
       "2                        1375              2102   \n",
       "3                         279               500   \n",
       "4                           1                 1   \n",
       "\n",
       "   u_question_part_correctly_count  u_question_part_count  \\\n",
       "0                              118                    245   \n",
       "1                                0                      0   \n",
       "2                              712                   1089   \n",
       "3                              187                    325   \n",
       "4                                1                      1   \n",
       "\n",
       "   u_question_tag1_correctly_count  u_question_tag1_count  \\\n",
       "0                                1                      7   \n",
       "1                                0                      0   \n",
       "2                               12                     17   \n",
       "3                                0                      1   \n",
       "4                                0                      0   \n",
       "\n",
       "   u_prior_question_correctly_timestamp_diff  \\\n",
       "0                                      16898   \n",
       "1                                      87403   \n",
       "2                                      27258   \n",
       "3                                  210001732   \n",
       "4                                      33806   \n",
       "\n",
       "   u_prior_question_correctly_timestamp_diff2  \\\n",
       "0                                      151141   \n",
       "1                                       87403   \n",
       "2                                       62834   \n",
       "3                                   210061483   \n",
       "4                                       33806   \n",
       "\n",
       "   u_prior_question_correctly_timestamp_diff3  \\\n",
       "0                                      211368   \n",
       "1                                       87403   \n",
       "2                                      118494   \n",
       "3                                   210192481   \n",
       "4                                       33806   \n",
       "\n",
       "   u_prior_question_timestamp_diff  u_prior_question_timestamp_diff2  \\\n",
       "0                            16898                             32512   \n",
       "1                            22899                             51871   \n",
       "2                            27258                             39668   \n",
       "3                            57387                         210001732   \n",
       "4                            33806                             33806   \n",
       "\n",
       "   u_prior_question_timestamp_diff3  u_prior_question_timestamp_diff4  \\\n",
       "0                             82698                            151141   \n",
       "1                             87403                             87403   \n",
       "2                             62834                             77536   \n",
       "3                         210061483                         210129010   \n",
       "4                             33806                             33806   \n",
       "\n",
       "   u_prior_lecture_timestamp_diff  u_prior2_lecture_timestamp_diff  \\\n",
       "0                      2146476402                       2146476402   \n",
       "1                           87403                            87403   \n",
       "2                     -1747755980                      -1747755980   \n",
       "3                       621093330                        621093330   \n",
       "4                           33806                            33806   \n",
       "\n",
       "   u_task_container_id_diff  u_prior_question_explanation_count  \\\n",
       "0                         1                                 480   \n",
       "1                         1                                   0   \n",
       "2                         1                                2091   \n",
       "3                         1                                 492   \n",
       "4                         1                                   0   \n",
       "\n",
       "   u_prior_question_explanation_correctly_count  \\\n",
       "0                                           313   \n",
       "1                                             0   \n",
       "2                                          1375   \n",
       "3                                           279   \n",
       "4                                             1   \n",
       "\n",
       "   u_question_listening_correctly_count  u_question_reading_correctly_count  \\\n",
       "0                                    62                                 251   \n",
       "1                                     0                                   0   \n",
       "2                                   610                                 765   \n",
       "3                                    92                                 187   \n",
       "4                                     0                                   1   \n",
       "\n",
       "   u_question_listening_count  u_question_reading_count  \\\n",
       "0                          88                       410   \n",
       "1                           3                         0   \n",
       "2                         917                      1185   \n",
       "3                         175                       325   \n",
       "4                           0                         1   \n",
       "\n",
       "   u_question_incorrect_timestamp_diff  u_question_incorrect_timestamp_diff2  \\\n",
       "0                                32512                                 82698   \n",
       "1                                22899                                 51871   \n",
       "2                                39668                                 77536   \n",
       "3                                57387                             210129010   \n",
       "4                                33806                                 33806   \n",
       "\n",
       "   u_question_incorrect_timestamp_diff3  u_question_seen  u_question_nunique  \\\n",
       "0                                238277                1                 441   \n",
       "1                                 87403                0                   3   \n",
       "2                                 92497                0                1835   \n",
       "3                             210522304                0                 408   \n",
       "4                                 33806                0                   1   \n",
       "\n",
       "   u_question_part_correct_timestamp_diff  \\\n",
       "0                                   16898   \n",
       "1                                   87403   \n",
       "2                                   27258   \n",
       "3                               235586768   \n",
       "4                                   33806   \n",
       "\n",
       "   u_question_part_incorrect_timestamp_diff  uc_question_timestamp_diff  \\\n",
       "0                                     32512                    44938880   \n",
       "1                                     87403                       87404   \n",
       "2                                     39668                 -1747755979   \n",
       "3                                     57387                   621093331   \n",
       "4                                     33806                       33807   \n",
       "\n",
       "   u_answered_correctly_avg  u_question_part_correctly_avg  \\\n",
       "0                  0.628514                       0.481633   \n",
       "1                  0.000000                            NaN   \n",
       "2                  0.654139                       0.653811   \n",
       "3                  0.558000                       0.575385   \n",
       "4                  1.000000                       1.000000   \n",
       "\n",
       "   u_question_tag1_correctly_avg  u_prior_question_timestamp_diff_2_1  \\\n",
       "0                       0.142857                                15614   \n",
       "1                            NaN                                28972   \n",
       "2                       0.705882                                12410   \n",
       "3                       0.000000                            209944345   \n",
       "4                            NaN                                    0   \n",
       "\n",
       "   u_prior_question_timestamp_diff_3_2  u_prior_question_timestamp_diff_4_3  \\\n",
       "0                                50186                                68443   \n",
       "1                                35532                                    0   \n",
       "2                                23166                                14702   \n",
       "3                                59751                                67527   \n",
       "4                                    0                                    0   \n",
       "\n",
       "   u_question_timestamp_diff  u_question_timestamp_diff2  \\\n",
       "0                     5614.0                     40186.0   \n",
       "1                     2972.0                      9532.0   \n",
       "2                     9410.0                     20166.0   \n",
       "3                209897344.0                     12751.0   \n",
       "4                   -21000.0                    -21000.0   \n",
       "\n",
       "   u_question_timestamp_diff3  u_listening_reading_ratio  \\\n",
       "0                     58443.0                   0.246032   \n",
       "1                    -26000.0                   0.000000   \n",
       "2                     11702.0                   0.796345   \n",
       "3                     20527.0                   0.489362   \n",
       "4                    -21000.0                   0.000000   \n",
       "\n",
       "   u_listening_correctly_avg  u_reading_correctly_avg  \n",
       "0                   0.696629                 0.610706  \n",
       "1                   0.000000                 0.000000  \n",
       "2                   0.664488                 0.645025  \n",
       "3                   0.522727                 0.573620  \n",
       "4                   0.000000                 0.500000  "
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "valid.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(96817414, 74) (2453886, 74)\n"
     ]
    }
   ],
   "source": [
    "print(train.shape, valid.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "66\n",
      "['content_id', 'prior_question_elapsed_time', 'prior_question_had_explanation', 'answered_correctly_avg_c', 'answered_correctly_var_c', 'part', 'question_asked', 'right_answers', 'bundle_size', 'bundle_rignt_answers', 'bundle_questions_asked', 'bundle_accuracy', 'part_rignt_answers', 'part_questions_asked', 'part_accuracy', 'tag1', 'tag1_answered_correctly_mean', 'tag1_answered_correctly_var', 'tags_emb_0', 'tags_emb_1', 'c_question_count_percent', 'c_question_part_percent', 'c_question_tag1_percent', 'c_question_unique_users_seen', 'u_answered_correctly_count', 'u_answered_count', 'u_question_part_correctly_count', 'u_question_part_count', 'u_question_tag1_correctly_count', 'u_question_tag1_count', 'u_prior_question_correctly_timestamp_diff', 'u_prior_question_correctly_timestamp_diff2', 'u_prior_question_correctly_timestamp_diff3', 'u_prior_question_timestamp_diff', 'u_prior_question_timestamp_diff2', 'u_prior_question_timestamp_diff3', 'u_prior_question_timestamp_diff4', 'u_prior_lecture_timestamp_diff', 'u_prior2_lecture_timestamp_diff', 'u_task_container_id_diff', 'u_prior_question_explanation_count', 'u_prior_question_explanation_correctly_count', 'u_question_listening_correctly_count', 'u_question_reading_correctly_count', 'u_question_listening_count', 'u_question_reading_count', 'u_question_incorrect_timestamp_diff', 'u_question_incorrect_timestamp_diff2', 'u_question_incorrect_timestamp_diff3', 'u_question_seen', 'u_question_nunique', 'u_question_part_correct_timestamp_diff', 'u_question_part_incorrect_timestamp_diff', 'uc_question_timestamp_diff', 'u_answered_correctly_avg', 'u_question_part_correctly_avg', 'u_question_tag1_correctly_avg', 'u_prior_question_timestamp_diff_2_1', 'u_prior_question_timestamp_diff_3_2', 'u_prior_question_timestamp_diff_4_3', 'u_question_timestamp_diff', 'u_question_timestamp_diff2', 'u_question_timestamp_diff3', 'u_listening_reading_ratio', 'u_listening_correctly_avg', 'u_reading_correctly_avg']\n"
     ]
    }
   ],
   "source": [
    "TARGET = 'answered_correctly'\n",
    "FEATS = [col for col in train.columns if col not in ['row_id',\n",
    "                                                     'user_id',\n",
    "#                                                      'content_id',\n",
    "                                                     'task_container_id',\n",
    "                                                     'question_id',\n",
    "                                                     'content_type_id',\n",
    "                                                     'timestamp',\n",
    "                                                     'answered_correctly',\n",
    "                                                     'lecture_part',\n",
    "                                                     'LorR',\n",
    "                                                    ]]\n",
    "\n",
    "print(len(FEATS))\n",
    "print(FEATS)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "105"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "drop_cols = list(set(train.columns) - set(FEATS))\n",
    "\n",
    "y_tr = train[TARGET]\n",
    "y_va = valid[TARGET]\n",
    "\n",
    "train.drop(drop_cols, axis=1, inplace=True)\n",
    "valid.drop(drop_cols, axis=1, inplace=True)\n",
    "\n",
    "gc.collect()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "20"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lgb_train = lgb.Dataset(train[FEATS], y_tr)\n",
    "lgb_valid = lgb.Dataset(valid[FEATS], y_va)\n",
    "\n",
    "del train, y_tr\n",
    "gc.collect()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[LightGBM] [Info] Number of positive: 63669896, number of negative: 33147518\n",
      "[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 7.090313 seconds.\n",
      "You can set `force_row_wise=true` to remove the overhead.\n",
      "And if memory is not enough, you can set `force_col_wise=true`.\n",
      "[LightGBM] [Info] Total Bins 14140\n",
      "[LightGBM] [Info] Number of data points in the train set: 96817414, number of used features: 66\n",
      "[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.657629 -> initscore=0.652744\n",
      "[LightGBM] [Info] Start training from score 0.652744\n",
      "Training until validation scores don't improve for 120 rounds\n",
      "[100]\ttraining's binary_logloss: 0.525853\tvalid_1's binary_logloss: 0.530868\n",
      "[200]\ttraining's binary_logloss: 0.523067\tvalid_1's binary_logloss: 0.528054\n",
      "[300]\ttraining's binary_logloss: 0.521762\tvalid_1's binary_logloss: 0.526736\n",
      "[400]\ttraining's binary_logloss: 0.520973\tvalid_1's binary_logloss: 0.525945\n",
      "[500]\ttraining's binary_logloss: 0.520294\tvalid_1's binary_logloss: 0.525304\n",
      "[600]\ttraining's binary_logloss: 0.519757\tvalid_1's binary_logloss: 0.524761\n",
      "[700]\ttraining's binary_logloss: 0.519317\tvalid_1's binary_logloss: 0.524336\n",
      "[800]\ttraining's binary_logloss: 0.518955\tvalid_1's binary_logloss: 0.523985\n",
      "[900]\ttraining's binary_logloss: 0.518588\tvalid_1's binary_logloss: 0.523643\n",
      "[1000]\ttraining's binary_logloss: 0.518291\tvalid_1's binary_logloss: 0.523353\n",
      "[1100]\ttraining's binary_logloss: 0.518028\tvalid_1's binary_logloss: 0.523109\n",
      "[1200]\ttraining's binary_logloss: 0.517765\tvalid_1's binary_logloss: 0.522869\n",
      "[1300]\ttraining's binary_logloss: 0.51751\tvalid_1's binary_logloss: 0.522651\n",
      "[1400]\ttraining's binary_logloss: 0.517275\tvalid_1's binary_logloss: 0.522444\n",
      "[1500]\ttraining's binary_logloss: 0.517069\tvalid_1's binary_logloss: 0.522266\n",
      "[1600]\ttraining's binary_logloss: 0.516864\tvalid_1's binary_logloss: 0.522086\n",
      "[1700]\ttraining's binary_logloss: 0.51666\tvalid_1's binary_logloss: 0.52191\n",
      "[1800]\ttraining's binary_logloss: 0.516481\tvalid_1's binary_logloss: 0.521755\n",
      "[1900]\ttraining's binary_logloss: 0.516297\tvalid_1's binary_logloss: 0.521593\n",
      "[2000]\ttraining's binary_logloss: 0.516125\tvalid_1's binary_logloss: 0.521443\n",
      "[2100]\ttraining's binary_logloss: 0.515974\tvalid_1's binary_logloss: 0.521318\n",
      "[2200]\ttraining's binary_logloss: 0.515799\tvalid_1's binary_logloss: 0.521165\n",
      "[2300]\ttraining's binary_logloss: 0.515652\tvalid_1's binary_logloss: 0.521042\n",
      "[2400]\ttraining's binary_logloss: 0.515513\tvalid_1's binary_logloss: 0.520931\n",
      "[2500]\ttraining's binary_logloss: 0.51537\tvalid_1's binary_logloss: 0.52081\n",
      "[2600]\ttraining's binary_logloss: 0.515239\tvalid_1's binary_logloss: 0.520713\n",
      "[2700]\ttraining's binary_logloss: 0.515127\tvalid_1's binary_logloss: 0.520625\n",
      "[2800]\ttraining's binary_logloss: 0.515004\tvalid_1's binary_logloss: 0.520524\n",
      "[2900]\ttraining's binary_logloss: 0.514879\tvalid_1's binary_logloss: 0.520422\n",
      "[3000]\ttraining's binary_logloss: 0.514749\tvalid_1's binary_logloss: 0.52031\n",
      "[3100]\ttraining's binary_logloss: 0.514626\tvalid_1's binary_logloss: 0.520219\n",
      "[3200]\ttraining's binary_logloss: 0.514496\tvalid_1's binary_logloss: 0.520106\n",
      "[3300]\ttraining's binary_logloss: 0.514388\tvalid_1's binary_logloss: 0.520019\n",
      "[3400]\ttraining's binary_logloss: 0.514284\tvalid_1's binary_logloss: 0.51994\n",
      "[3500]\ttraining's binary_logloss: 0.51418\tvalid_1's binary_logloss: 0.519862\n",
      "[3600]\ttraining's binary_logloss: 0.514087\tvalid_1's binary_logloss: 0.519795\n",
      "[3700]\ttraining's binary_logloss: 0.513976\tvalid_1's binary_logloss: 0.519709\n",
      "[3800]\ttraining's binary_logloss: 0.513875\tvalid_1's binary_logloss: 0.519641\n",
      "[3900]\ttraining's binary_logloss: 0.513779\tvalid_1's binary_logloss: 0.519571\n",
      "[4000]\ttraining's binary_logloss: 0.5137\tvalid_1's binary_logloss: 0.519515\n",
      "[4100]\ttraining's binary_logloss: 0.513615\tvalid_1's binary_logloss: 0.519449\n",
      "[4200]\ttraining's binary_logloss: 0.513517\tvalid_1's binary_logloss: 0.519372\n",
      "[4300]\ttraining's binary_logloss: 0.513437\tvalid_1's binary_logloss: 0.519313\n",
      "[4400]\ttraining's binary_logloss: 0.513347\tvalid_1's binary_logloss: 0.519247\n",
      "[4500]\ttraining's binary_logloss: 0.513251\tvalid_1's binary_logloss: 0.519177\n",
      "[4600]\ttraining's binary_logloss: 0.513174\tvalid_1's binary_logloss: 0.519127\n",
      "[4700]\ttraining's binary_logloss: 0.513089\tvalid_1's binary_logloss: 0.519066\n",
      "[4800]\ttraining's binary_logloss: 0.513009\tvalid_1's binary_logloss: 0.519015\n",
      "[4900]\ttraining's binary_logloss: 0.512923\tvalid_1's binary_logloss: 0.518956\n",
      "[5000]\ttraining's binary_logloss: 0.512841\tvalid_1's binary_logloss: 0.518899\n",
      "[5100]\ttraining's binary_logloss: 0.512772\tvalid_1's binary_logloss: 0.518856\n",
      "[5200]\ttraining's binary_logloss: 0.512704\tvalid_1's binary_logloss: 0.518817\n",
      "[5300]\ttraining's binary_logloss: 0.512627\tvalid_1's binary_logloss: 0.518768\n",
      "[5400]\ttraining's binary_logloss: 0.512553\tvalid_1's binary_logloss: 0.518722\n",
      "[5500]\ttraining's binary_logloss: 0.512479\tvalid_1's binary_logloss: 0.518677\n",
      "[5600]\ttraining's binary_logloss: 0.512408\tvalid_1's binary_logloss: 0.518633\n",
      "[5700]\ttraining's binary_logloss: 0.512341\tvalid_1's binary_logloss: 0.518597\n",
      "[5800]\ttraining's binary_logloss: 0.512277\tvalid_1's binary_logloss: 0.518568\n",
      "[5900]\ttraining's binary_logloss: 0.512211\tvalid_1's binary_logloss: 0.518533\n",
      "[6000]\ttraining's binary_logloss: 0.512143\tvalid_1's binary_logloss: 0.518495\n",
      "[6100]\ttraining's binary_logloss: 0.512083\tvalid_1's binary_logloss: 0.518468\n",
      "[6200]\ttraining's binary_logloss: 0.512024\tvalid_1's binary_logloss: 0.518438\n",
      "[6300]\ttraining's binary_logloss: 0.511962\tvalid_1's binary_logloss: 0.518405\n",
      "[6400]\ttraining's binary_logloss: 0.511895\tvalid_1's binary_logloss: 0.518368\n",
      "[6500]\ttraining's binary_logloss: 0.511829\tvalid_1's binary_logloss: 0.518328\n",
      "[6600]\ttraining's binary_logloss: 0.511769\tvalid_1's binary_logloss: 0.518299\n",
      "[6700]\ttraining's binary_logloss: 0.511705\tvalid_1's binary_logloss: 0.518262\n",
      "[6800]\ttraining's binary_logloss: 0.511637\tvalid_1's binary_logloss: 0.518221\n",
      "[6900]\ttraining's binary_logloss: 0.511587\tvalid_1's binary_logloss: 0.518201\n",
      "[7000]\ttraining's binary_logloss: 0.511531\tvalid_1's binary_logloss: 0.518177\n",
      "[7100]\ttraining's binary_logloss: 0.511472\tvalid_1's binary_logloss: 0.518148\n",
      "[7200]\ttraining's binary_logloss: 0.511409\tvalid_1's binary_logloss: 0.518113\n",
      "[7300]\ttraining's binary_logloss: 0.511357\tvalid_1's binary_logloss: 0.518093\n",
      "[7400]\ttraining's binary_logloss: 0.511293\tvalid_1's binary_logloss: 0.518052\n",
      "[7500]\ttraining's binary_logloss: 0.511237\tvalid_1's binary_logloss: 0.518023\n",
      "[7600]\ttraining's binary_logloss: 0.511182\tvalid_1's binary_logloss: 0.517996\n",
      "[7700]\ttraining's binary_logloss: 0.511129\tvalid_1's binary_logloss: 0.51797\n",
      "[7800]\ttraining's binary_logloss: 0.511075\tvalid_1's binary_logloss: 0.517942\n",
      "[7900]\ttraining's binary_logloss: 0.511018\tvalid_1's binary_logloss: 0.51791\n",
      "[8000]\ttraining's binary_logloss: 0.510969\tvalid_1's binary_logloss: 0.517894\n",
      "[8100]\ttraining's binary_logloss: 0.510908\tvalid_1's binary_logloss: 0.517858\n",
      "[8200]\ttraining's binary_logloss: 0.510852\tvalid_1's binary_logloss: 0.51783\n",
      "[8300]\ttraining's binary_logloss: 0.510797\tvalid_1's binary_logloss: 0.517804\n",
      "[8400]\ttraining's binary_logloss: 0.510739\tvalid_1's binary_logloss: 0.517772\n",
      "[8500]\ttraining's binary_logloss: 0.510681\tvalid_1's binary_logloss: 0.517748\n",
      "[8600]\ttraining's binary_logloss: 0.510625\tvalid_1's binary_logloss: 0.517718\n",
      "[8700]\ttraining's binary_logloss: 0.510572\tvalid_1's binary_logloss: 0.517694\n",
      "[8800]\ttraining's binary_logloss: 0.510518\tvalid_1's binary_logloss: 0.517673\n",
      "[8900]\ttraining's binary_logloss: 0.510467\tvalid_1's binary_logloss: 0.517652\n",
      "[9000]\ttraining's binary_logloss: 0.510417\tvalid_1's binary_logloss: 0.517634\n",
      "[9100]\ttraining's binary_logloss: 0.510367\tvalid_1's binary_logloss: 0.517612\n",
      "[9200]\ttraining's binary_logloss: 0.510313\tvalid_1's binary_logloss: 0.517586\n",
      "[9300]\ttraining's binary_logloss: 0.510264\tvalid_1's binary_logloss: 0.517565\n",
      "[9400]\ttraining's binary_logloss: 0.510211\tvalid_1's binary_logloss: 0.517543\n",
      "[9500]\ttraining's binary_logloss: 0.510162\tvalid_1's binary_logloss: 0.517522\n",
      "[9600]\ttraining's binary_logloss: 0.510114\tvalid_1's binary_logloss: 0.517505\n",
      "[9700]\ttraining's binary_logloss: 0.510062\tvalid_1's binary_logloss: 0.517485\n",
      "[9800]\ttraining's binary_logloss: 0.510014\tvalid_1's binary_logloss: 0.517467\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[9900]\ttraining's binary_logloss: 0.509963\tvalid_1's binary_logloss: 0.517445\n",
      "[10000]\ttraining's binary_logloss: 0.509913\tvalid_1's binary_logloss: 0.517427\n",
      "[10100]\ttraining's binary_logloss: 0.509863\tvalid_1's binary_logloss: 0.517406\n",
      "[10200]\ttraining's binary_logloss: 0.509817\tvalid_1's binary_logloss: 0.517389\n",
      "[10300]\ttraining's binary_logloss: 0.509772\tvalid_1's binary_logloss: 0.517376\n",
      "[10400]\ttraining's binary_logloss: 0.509723\tvalid_1's binary_logloss: 0.517354\n",
      "[10500]\ttraining's binary_logloss: 0.509678\tvalid_1's binary_logloss: 0.517336\n",
      "[10600]\ttraining's binary_logloss: 0.509634\tvalid_1's binary_logloss: 0.51732\n",
      "[10700]\ttraining's binary_logloss: 0.509588\tvalid_1's binary_logloss: 0.517304\n",
      "[10800]\ttraining's binary_logloss: 0.509543\tvalid_1's binary_logloss: 0.51729\n",
      "[10900]\ttraining's binary_logloss: 0.5095\tvalid_1's binary_logloss: 0.517276\n",
      "[11000]\ttraining's binary_logloss: 0.509452\tvalid_1's binary_logloss: 0.517259\n",
      "[11100]\ttraining's binary_logloss: 0.509405\tvalid_1's binary_logloss: 0.517244\n",
      "[11200]\ttraining's binary_logloss: 0.509361\tvalid_1's binary_logloss: 0.517231\n",
      "[11300]\ttraining's binary_logloss: 0.509321\tvalid_1's binary_logloss: 0.517221\n",
      "[11400]\ttraining's binary_logloss: 0.509275\tvalid_1's binary_logloss: 0.517205\n",
      "[11500]\ttraining's binary_logloss: 0.509229\tvalid_1's binary_logloss: 0.517189\n",
      "[11600]\ttraining's binary_logloss: 0.509187\tvalid_1's binary_logloss: 0.517181\n",
      "[11700]\ttraining's binary_logloss: 0.509145\tvalid_1's binary_logloss: 0.517165\n",
      "[11800]\ttraining's binary_logloss: 0.509107\tvalid_1's binary_logloss: 0.517158\n",
      "[11900]\ttraining's binary_logloss: 0.509067\tvalid_1's binary_logloss: 0.517146\n",
      "[12000]\ttraining's binary_logloss: 0.509021\tvalid_1's binary_logloss: 0.517131\n",
      "[12100]\ttraining's binary_logloss: 0.508976\tvalid_1's binary_logloss: 0.517115\n",
      "[12200]\ttraining's binary_logloss: 0.508934\tvalid_1's binary_logloss: 0.517103\n",
      "[12300]\ttraining's binary_logloss: 0.50889\tvalid_1's binary_logloss: 0.517088\n",
      "[12400]\ttraining's binary_logloss: 0.508847\tvalid_1's binary_logloss: 0.517074\n",
      "[12500]\ttraining's binary_logloss: 0.508804\tvalid_1's binary_logloss: 0.517063\n",
      "[12600]\ttraining's binary_logloss: 0.508761\tvalid_1's binary_logloss: 0.517052\n",
      "[12700]\ttraining's binary_logloss: 0.508718\tvalid_1's binary_logloss: 0.517042\n",
      "[12800]\ttraining's binary_logloss: 0.508678\tvalid_1's binary_logloss: 0.517035\n",
      "[12900]\ttraining's binary_logloss: 0.508637\tvalid_1's binary_logloss: 0.517026\n",
      "[13000]\ttraining's binary_logloss: 0.508595\tvalid_1's binary_logloss: 0.517015\n",
      "[13100]\ttraining's binary_logloss: 0.508551\tvalid_1's binary_logloss: 0.517001\n",
      "[13200]\ttraining's binary_logloss: 0.508507\tvalid_1's binary_logloss: 0.516992\n",
      "[13300]\ttraining's binary_logloss: 0.508468\tvalid_1's binary_logloss: 0.516983\n",
      "[13400]\ttraining's binary_logloss: 0.50843\tvalid_1's binary_logloss: 0.516978\n",
      "[13500]\ttraining's binary_logloss: 0.508384\tvalid_1's binary_logloss: 0.516962\n",
      "[13600]\ttraining's binary_logloss: 0.508344\tvalid_1's binary_logloss: 0.516952\n",
      "[13700]\ttraining's binary_logloss: 0.508302\tvalid_1's binary_logloss: 0.51694\n",
      "[13800]\ttraining's binary_logloss: 0.508264\tvalid_1's binary_logloss: 0.516935\n",
      "[13900]\ttraining's binary_logloss: 0.508226\tvalid_1's binary_logloss: 0.516929\n",
      "[14000]\ttraining's binary_logloss: 0.508183\tvalid_1's binary_logloss: 0.516917\n",
      "[14100]\ttraining's binary_logloss: 0.508142\tvalid_1's binary_logloss: 0.516904\n",
      "[14200]\ttraining's binary_logloss: 0.508101\tvalid_1's binary_logloss: 0.516898\n",
      "[14300]\ttraining's binary_logloss: 0.508062\tvalid_1's binary_logloss: 0.51689\n",
      "[14400]\ttraining's binary_logloss: 0.508018\tvalid_1's binary_logloss: 0.516875\n",
      "[14500]\ttraining's binary_logloss: 0.507975\tvalid_1's binary_logloss: 0.516859\n",
      "[14600]\ttraining's binary_logloss: 0.507933\tvalid_1's binary_logloss: 0.516845\n",
      "[14700]\ttraining's binary_logloss: 0.507893\tvalid_1's binary_logloss: 0.516836\n",
      "[14800]\ttraining's binary_logloss: 0.50785\tvalid_1's binary_logloss: 0.516825\n",
      "[14900]\ttraining's binary_logloss: 0.507812\tvalid_1's binary_logloss: 0.516817\n",
      "[15000]\ttraining's binary_logloss: 0.507775\tvalid_1's binary_logloss: 0.516812\n",
      "[15100]\ttraining's binary_logloss: 0.507734\tvalid_1's binary_logloss: 0.516802\n",
      "[15200]\ttraining's binary_logloss: 0.507695\tvalid_1's binary_logloss: 0.516793\n",
      "[15300]\ttraining's binary_logloss: 0.507653\tvalid_1's binary_logloss: 0.516782\n",
      "[15400]\ttraining's binary_logloss: 0.507612\tvalid_1's binary_logloss: 0.51677\n",
      "[15500]\ttraining's binary_logloss: 0.507574\tvalid_1's binary_logloss: 0.516763\n",
      "[15600]\ttraining's binary_logloss: 0.507535\tvalid_1's binary_logloss: 0.516754\n",
      "[15700]\ttraining's binary_logloss: 0.507498\tvalid_1's binary_logloss: 0.516746\n",
      "[15800]\ttraining's binary_logloss: 0.507457\tvalid_1's binary_logloss: 0.516731\n",
      "[15900]\ttraining's binary_logloss: 0.50742\tvalid_1's binary_logloss: 0.516723\n",
      "[16000]\ttraining's binary_logloss: 0.507381\tvalid_1's binary_logloss: 0.516715\n",
      "[16100]\ttraining's binary_logloss: 0.507343\tvalid_1's binary_logloss: 0.516708\n",
      "[16200]\ttraining's binary_logloss: 0.507302\tvalid_1's binary_logloss: 0.516698\n",
      "[16300]\ttraining's binary_logloss: 0.507263\tvalid_1's binary_logloss: 0.516688\n",
      "[16400]\ttraining's binary_logloss: 0.507224\tvalid_1's binary_logloss: 0.516679\n",
      "[16500]\ttraining's binary_logloss: 0.507185\tvalid_1's binary_logloss: 0.516669\n",
      "[16600]\ttraining's binary_logloss: 0.507148\tvalid_1's binary_logloss: 0.516662\n",
      "[16700]\ttraining's binary_logloss: 0.507111\tvalid_1's binary_logloss: 0.516656\n",
      "[16800]\ttraining's binary_logloss: 0.507074\tvalid_1's binary_logloss: 0.516648\n",
      "[16900]\ttraining's binary_logloss: 0.507037\tvalid_1's binary_logloss: 0.516644\n",
      "[17000]\ttraining's binary_logloss: 0.506998\tvalid_1's binary_logloss: 0.516636\n",
      "[17100]\ttraining's binary_logloss: 0.506964\tvalid_1's binary_logloss: 0.516631\n",
      "[17200]\ttraining's binary_logloss: 0.506927\tvalid_1's binary_logloss: 0.516624\n",
      "[17300]\ttraining's binary_logloss: 0.506888\tvalid_1's binary_logloss: 0.516615\n",
      "[17400]\ttraining's binary_logloss: 0.50685\tvalid_1's binary_logloss: 0.516607\n",
      "[17500]\ttraining's binary_logloss: 0.506815\tvalid_1's binary_logloss: 0.516601\n",
      "[17600]\ttraining's binary_logloss: 0.506779\tvalid_1's binary_logloss: 0.516595\n",
      "[17700]\ttraining's binary_logloss: 0.506741\tvalid_1's binary_logloss: 0.51659\n",
      "[17800]\ttraining's binary_logloss: 0.506702\tvalid_1's binary_logloss: 0.516583\n",
      "[17900]\ttraining's binary_logloss: 0.506667\tvalid_1's binary_logloss: 0.516578\n",
      "[18000]\ttraining's binary_logloss: 0.506631\tvalid_1's binary_logloss: 0.516574\n",
      "[18100]\ttraining's binary_logloss: 0.506593\tvalid_1's binary_logloss: 0.516566\n",
      "[18200]\ttraining's binary_logloss: 0.506555\tvalid_1's binary_logloss: 0.516557\n",
      "[18300]\ttraining's binary_logloss: 0.506519\tvalid_1's binary_logloss: 0.51655\n",
      "[18400]\ttraining's binary_logloss: 0.506486\tvalid_1's binary_logloss: 0.516547\n",
      "[18500]\ttraining's binary_logloss: 0.506448\tvalid_1's binary_logloss: 0.516539\n",
      "[18600]\ttraining's binary_logloss: 0.506413\tvalid_1's binary_logloss: 0.516533\n",
      "[18700]\ttraining's binary_logloss: 0.506378\tvalid_1's binary_logloss: 0.51653\n",
      "[18800]\ttraining's binary_logloss: 0.506343\tvalid_1's binary_logloss: 0.516524\n",
      "[18900]\ttraining's binary_logloss: 0.506308\tvalid_1's binary_logloss: 0.516515\n",
      "[19000]\ttraining's binary_logloss: 0.506269\tvalid_1's binary_logloss: 0.516507\n",
      "[19100]\ttraining's binary_logloss: 0.506231\tvalid_1's binary_logloss: 0.516501\n",
      "[19200]\ttraining's binary_logloss: 0.506193\tvalid_1's binary_logloss: 0.516494\n",
      "[19300]\ttraining's binary_logloss: 0.506154\tvalid_1's binary_logloss: 0.516484\n",
      "[19400]\ttraining's binary_logloss: 0.506119\tvalid_1's binary_logloss: 0.516477\n",
      "[19500]\ttraining's binary_logloss: 0.506084\tvalid_1's binary_logloss: 0.516471\n",
      "[19600]\ttraining's binary_logloss: 0.506049\tvalid_1's binary_logloss: 0.516465\n",
      "[19700]\ttraining's binary_logloss: 0.506012\tvalid_1's binary_logloss: 0.516457\n",
      "[19800]\ttraining's binary_logloss: 0.505976\tvalid_1's binary_logloss: 0.51645\n",
      "[19900]\ttraining's binary_logloss: 0.50594\tvalid_1's binary_logloss: 0.516445\n",
      "[20000]\ttraining's binary_logloss: 0.505905\tvalid_1's binary_logloss: 0.516439\n",
      "[20100]\ttraining's binary_logloss: 0.50587\tvalid_1's binary_logloss: 0.516434\n",
      "[20200]\ttraining's binary_logloss: 0.505837\tvalid_1's binary_logloss: 0.51643\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[20300]\ttraining's binary_logloss: 0.505801\tvalid_1's binary_logloss: 0.516427\n",
      "[20400]\ttraining's binary_logloss: 0.505764\tvalid_1's binary_logloss: 0.51642\n",
      "[20500]\ttraining's binary_logloss: 0.505731\tvalid_1's binary_logloss: 0.516417\n",
      "[20600]\ttraining's binary_logloss: 0.505698\tvalid_1's binary_logloss: 0.516413\n",
      "[20700]\ttraining's binary_logloss: 0.505664\tvalid_1's binary_logloss: 0.516412\n",
      "[20800]\ttraining's binary_logloss: 0.50563\tvalid_1's binary_logloss: 0.516406\n",
      "[20900]\ttraining's binary_logloss: 0.505599\tvalid_1's binary_logloss: 0.516404\n",
      "[21000]\ttraining's binary_logloss: 0.505566\tvalid_1's binary_logloss: 0.516404\n",
      "[21100]\ttraining's binary_logloss: 0.505532\tvalid_1's binary_logloss: 0.5164\n",
      "[21200]\ttraining's binary_logloss: 0.505499\tvalid_1's binary_logloss: 0.516398\n",
      "[21300]\ttraining's binary_logloss: 0.505465\tvalid_1's binary_logloss: 0.516395\n",
      "[21400]\ttraining's binary_logloss: 0.50543\tvalid_1's binary_logloss: 0.516391\n",
      "[21500]\ttraining's binary_logloss: 0.505396\tvalid_1's binary_logloss: 0.516389\n",
      "[21600]\ttraining's binary_logloss: 0.505361\tvalid_1's binary_logloss: 0.516386\n",
      "[21700]\ttraining's binary_logloss: 0.505325\tvalid_1's binary_logloss: 0.516381\n",
      "[21800]\ttraining's binary_logloss: 0.50529\tvalid_1's binary_logloss: 0.516373\n",
      "[21900]\ttraining's binary_logloss: 0.505254\tvalid_1's binary_logloss: 0.516366\n",
      "[22000]\ttraining's binary_logloss: 0.505219\tvalid_1's binary_logloss: 0.51636\n",
      "[22100]\ttraining's binary_logloss: 0.505185\tvalid_1's binary_logloss: 0.516357\n",
      "[22200]\ttraining's binary_logloss: 0.505153\tvalid_1's binary_logloss: 0.516353\n",
      "[22300]\ttraining's binary_logloss: 0.50512\tvalid_1's binary_logloss: 0.516351\n",
      "[22400]\ttraining's binary_logloss: 0.505086\tvalid_1's binary_logloss: 0.516347\n",
      "[22500]\ttraining's binary_logloss: 0.50505\tvalid_1's binary_logloss: 0.516344\n",
      "[22600]\ttraining's binary_logloss: 0.505017\tvalid_1's binary_logloss: 0.516341\n",
      "[22700]\ttraining's binary_logloss: 0.504982\tvalid_1's binary_logloss: 0.516337\n",
      "[22800]\ttraining's binary_logloss: 0.504949\tvalid_1's binary_logloss: 0.516333\n",
      "[22900]\ttraining's binary_logloss: 0.504915\tvalid_1's binary_logloss: 0.516331\n",
      "[23000]\ttraining's binary_logloss: 0.504879\tvalid_1's binary_logloss: 0.516323\n",
      "[23100]\ttraining's binary_logloss: 0.504845\tvalid_1's binary_logloss: 0.51632\n",
      "[23200]\ttraining's binary_logloss: 0.504812\tvalid_1's binary_logloss: 0.51632\n",
      "[23300]\ttraining's binary_logloss: 0.504779\tvalid_1's binary_logloss: 0.516317\n",
      "[23400]\ttraining's binary_logloss: 0.504745\tvalid_1's binary_logloss: 0.516312\n",
      "[23500]\ttraining's binary_logloss: 0.504711\tvalid_1's binary_logloss: 0.516308\n",
      "[23600]\ttraining's binary_logloss: 0.504675\tvalid_1's binary_logloss: 0.516306\n",
      "[23700]\ttraining's binary_logloss: 0.504641\tvalid_1's binary_logloss: 0.516302\n",
      "[23800]\ttraining's binary_logloss: 0.504608\tvalid_1's binary_logloss: 0.516299\n",
      "[23900]\ttraining's binary_logloss: 0.504575\tvalid_1's binary_logloss: 0.516295\n",
      "[24000]\ttraining's binary_logloss: 0.50454\tvalid_1's binary_logloss: 0.516289\n",
      "[24100]\ttraining's binary_logloss: 0.504506\tvalid_1's binary_logloss: 0.516284\n",
      "[24200]\ttraining's binary_logloss: 0.504471\tvalid_1's binary_logloss: 0.516279\n",
      "[24300]\ttraining's binary_logloss: 0.504439\tvalid_1's binary_logloss: 0.516276\n",
      "[24400]\ttraining's binary_logloss: 0.504407\tvalid_1's binary_logloss: 0.516271\n",
      "[24500]\ttraining's binary_logloss: 0.504376\tvalid_1's binary_logloss: 0.516268\n",
      "[24600]\ttraining's binary_logloss: 0.504343\tvalid_1's binary_logloss: 0.516265\n",
      "[24700]\ttraining's binary_logloss: 0.504311\tvalid_1's binary_logloss: 0.516262\n",
      "[24800]\ttraining's binary_logloss: 0.504278\tvalid_1's binary_logloss: 0.516259\n",
      "[24900]\ttraining's binary_logloss: 0.504246\tvalid_1's binary_logloss: 0.516256\n",
      "[25000]\ttraining's binary_logloss: 0.504214\tvalid_1's binary_logloss: 0.516256\n",
      "[25100]\ttraining's binary_logloss: 0.504181\tvalid_1's binary_logloss: 0.516252\n",
      "[25200]\ttraining's binary_logloss: 0.504146\tvalid_1's binary_logloss: 0.516248\n",
      "[25300]\ttraining's binary_logloss: 0.504112\tvalid_1's binary_logloss: 0.516245\n",
      "[25400]\ttraining's binary_logloss: 0.504081\tvalid_1's binary_logloss: 0.516242\n",
      "[25500]\ttraining's binary_logloss: 0.504049\tvalid_1's binary_logloss: 0.516241\n",
      "[25600]\ttraining's binary_logloss: 0.504015\tvalid_1's binary_logloss: 0.516237\n",
      "[25700]\ttraining's binary_logloss: 0.503983\tvalid_1's binary_logloss: 0.516234\n",
      "[25800]\ttraining's binary_logloss: 0.503951\tvalid_1's binary_logloss: 0.516232\n",
      "[25900]\ttraining's binary_logloss: 0.503919\tvalid_1's binary_logloss: 0.516229\n",
      "[26000]\ttraining's binary_logloss: 0.503887\tvalid_1's binary_logloss: 0.516226\n",
      "[26100]\ttraining's binary_logloss: 0.503855\tvalid_1's binary_logloss: 0.516226\n",
      "Early stopping, best iteration is:\n",
      "[26053]\ttraining's binary_logloss: 0.503869\tvalid_1's binary_logloss: 0.516225\n"
     ]
    }
   ],
   "source": [
    "model = lgb.train({'objective': 'binary'}, \n",
    "                  lgb_train,\n",
    "                  valid_sets=[lgb_train, lgb_valid],\n",
    "                  verbose_eval=100,\n",
    "                  num_boost_round=30000,\n",
    "                  early_stopping_rounds=120)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "auc: 0.7959698296462356\n"
     ]
    }
   ],
   "source": [
    "print('auc:', roc_auc_score(y_va, model.predict(valid[FEATS])))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f4df05afcd0>"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x1152 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "plt.rcParams[\"figure.figsize\"] = (16, 16)\n",
    "lgb.plot_importance(model, importance_type='gain')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f4c52354760>"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x1152 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.rcParams[\"figure.figsize\"] = (16, 16)\n",
    "lgb.plot_importance(model, importance_type='split')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>row_id</th>\n",
       "      <th>answered_correctly</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0.400318</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0.871046</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>0.656222</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>0.853126</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>0.301258</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5</td>\n",
       "      <td>0.666266</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>6</td>\n",
       "      <td>0.519784</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>7</td>\n",
       "      <td>0.510012</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>8</td>\n",
       "      <td>0.743665</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>9</td>\n",
       "      <td>0.511642</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>10</td>\n",
       "      <td>0.815667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>11</td>\n",
       "      <td>0.539803</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>12</td>\n",
       "      <td>0.467004</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>13</td>\n",
       "      <td>0.732516</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>14</td>\n",
       "      <td>0.325141</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>15</td>\n",
       "      <td>0.784822</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>16</td>\n",
       "      <td>0.527455</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>17</td>\n",
       "      <td>0.934914</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>18</td>\n",
       "      <td>0.714859</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>19</td>\n",
       "      <td>0.418227</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>20</td>\n",
       "      <td>0.769338</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>21</td>\n",
       "      <td>0.651102</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>22</td>\n",
       "      <td>0.391399</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>23</td>\n",
       "      <td>0.344272</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>24</td>\n",
       "      <td>0.442560</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>25</td>\n",
       "      <td>0.954531</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>26</td>\n",
       "      <td>0.788641</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>27</td>\n",
       "      <td>0.079429</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>28</td>\n",
       "      <td>0.537424</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>29</td>\n",
       "      <td>0.737749</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    row_id  answered_correctly\n",
       "0        0            0.400318\n",
       "1        1            0.871046\n",
       "2        2            0.656222\n",
       "3        3            0.853126\n",
       "4        4            0.301258\n",
       "5        5            0.666266\n",
       "6        6            0.519784\n",
       "7        7            0.510012\n",
       "8        8            0.743665\n",
       "9        9            0.511642\n",
       "10      10            0.815667\n",
       "11      11            0.539803\n",
       "12      12            0.467004\n",
       "13      13            0.732516\n",
       "14      14            0.325141\n",
       "15      15            0.784822\n",
       "16      16            0.527455\n",
       "17      17            0.934914\n",
       "18      18            0.714859\n",
       "19      19            0.418227\n",
       "20      20            0.769338\n",
       "21      21            0.651102\n",
       "22      22            0.391399\n",
       "23      23            0.344272\n",
       "24      24            0.442560\n",
       "25      25            0.954531\n",
       "26      26            0.788641\n",
       "27      27            0.079429\n",
       "28      28            0.537424\n",
       "29      29            0.737749"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 检查推理阶段是否正常 (重点看 index=4 和 index=11)\n",
    "\n",
    "test_df = pd.read_csv('/home/user01/kaggle/riiid/example_test.csv')\n",
    "test_df = test_df[test_df['content_type_id'] == 0].reset_index(drop=True)\n",
    "test_df['prior_question_had_explanation'] = test_df.prior_question_had_explanation.fillna(False).astype('int8')\n",
    "test_df = pd.merge(test_df, questions_df, on='content_id', how='left')\n",
    "test_df = pd.merge(test_df, content_df, on='content_id',  how='left')\n",
    "test_df['prior_question_elapsed_time'] = test_df.prior_question_elapsed_time.fillna(prior_question_elapsed_time_mean)\n",
    "test_df.fillna(-1, inplace=True)\n",
    "use_cols = ['user_id', 'content_type_id', 'part', 'tag1',\n",
    "            'timestamp', 'task_container_id', 'prior_question_had_explanation', \n",
    "            'LorR', 'content_id']\n",
    "test_df[use_cols] = test_df[use_cols].astype('int')\n",
    "test_df = add_user_feats_without_update(test_df,\n",
    "                                        u_answered_correctly_count_dict, \n",
    "                                        u_answered_count_dict, \n",
    "                                        u_question_part_correctly_count_dict,\n",
    "                                        u_question_part_count_dict,\n",
    "                                        u_question_tag1_correctly_count_dict,\n",
    "                                        u_question_tag1_count_dict,\n",
    "                                        u_prior_question_correctly_timestamp_dict,\n",
    "                                        u_prior2_question_correctly_timestamp_dict,\n",
    "                                        u_prior3_question_correctly_timestamp_dict,\n",
    "                                        u_prior_question_timestamp_dict,\n",
    "                                        u_prior2_question_timestamp_dict,\n",
    "                                        u_prior3_question_timestamp_dict,\n",
    "                                        u_prior4_question_timestamp_dict,\n",
    "                                        u_prior_lecture_timestamp_dict,\n",
    "                                        u_prior2_lecture_timestamp_dict,\n",
    "                                        u_task_container_id_dict,\n",
    "                                        u_prior_question_explanation_count_dict,\n",
    "                                        u_prior_question_explanation_correctly_count_dict,\n",
    "                                        u_question_listening_correctly_count_dict, \n",
    "                                        u_question_reading_correctly_count_dict,\n",
    "                                        u_question_listening_count_dict, \n",
    "                                        u_question_reading_count_dict,\n",
    "                                        u_question_incorrect_timestamp_dict,\n",
    "                                        u_question_incorrect_timestamp2_dict,\n",
    "                                        u_question_incorrect_timestamp3_dict,\n",
    "                                        u_question_seen_dict,\n",
    "                                        u_question_part_correct_timestamp_dict, \n",
    "                                        u_question_part_incorrect_timestamp_dict)\n",
    "test_df[TARGET] = model.predict(test_df[FEATS])\n",
    "\n",
    "test_df[['row_id', TARGET]].head(30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 保存到上传目录\n",
    "\n",
    "upload_dir = './upload_cache_new'\n",
    "\n",
    "# Kaggle 上的 pickle.HIGHEST_PROTOCOL = 4\n",
    "\n",
    "content_df.to_pickle(f'{upload_dir}/content_df.pickle', protocol=4)\n",
    "questions_df.to_pickle(f'{upload_dir}/questions_df.pickle', protocol=4)\n",
    "\n",
    "model.save_model(f'{upload_dir}/lgb_model.txt')\n",
    "\n",
    "import pickle   # Kaggle 上的 pickle.HIGHEST_PROTOCOL = 4\n",
    "    \n",
    "with open(f'{upload_dir}/u_question_seen_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_question_seen_dict, handle, protocol=4)\n",
    "\n",
    "with open(f'{upload_dir}/u_answered_correctly_count_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_answered_correctly_count_dict, handle, protocol=4)\n",
    "    \n",
    "with open(f'{upload_dir}/u_answered_count_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_answered_count_dict, handle, protocol=4)\n",
    "    \n",
    "with open(f'{upload_dir}/u_question_part_correctly_count_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_question_part_correctly_count_dict, handle, protocol=4)\n",
    "    \n",
    "with open(f'{upload_dir}/u_question_part_count_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_question_part_count_dict, handle, protocol=4)\n",
    "    \n",
    "with open(f'{upload_dir}/u_question_tag1_correctly_count_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_question_tag1_correctly_count_dict, handle, protocol=4)\n",
    "    \n",
    "with open(f'{upload_dir}/u_question_tag1_count_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_question_tag1_count_dict, handle, protocol=4)  \n",
    "\n",
    "with open(f'{upload_dir}/u_prior_question_correctly_timestamp_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_prior_question_correctly_timestamp_dict, handle, protocol=4) \n",
    "    \n",
    "with open(f'{upload_dir}/u_prior2_question_correctly_timestamp_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_prior2_question_correctly_timestamp_dict, handle, protocol=4) \n",
    "    \n",
    "with open(f'{upload_dir}/u_prior3_question_correctly_timestamp_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_prior3_question_correctly_timestamp_dict, handle, protocol=4) \n",
    "    \n",
    "with open(f'{upload_dir}/u_prior_question_timestamp_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_prior_question_timestamp_dict, handle, protocol=4)\n",
    "    \n",
    "with open(f'{upload_dir}/u_prior2_question_timestamp_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_prior2_question_timestamp_dict, handle, protocol=4)\n",
    "    \n",
    "with open(f'{upload_dir}/u_prior3_question_timestamp_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_prior3_question_timestamp_dict, handle, protocol=4)\n",
    "    \n",
    "with open(f'{upload_dir}/u_prior4_question_timestamp_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_prior4_question_timestamp_dict, handle, protocol=4)\n",
    "    \n",
    "with open(f'{upload_dir}/u_prior_lecture_timestamp_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_prior_lecture_timestamp_dict, handle, protocol=4)\n",
    "    \n",
    "with open(f'{upload_dir}/u_prior2_lecture_timestamp_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_prior2_lecture_timestamp_dict, handle, protocol=4)\n",
    "    \n",
    "with open(f'{upload_dir}/u_task_container_id_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_task_container_id_dict, handle, protocol=4)\n",
    "    \n",
    "with open(f'{upload_dir}/u_prior_question_explanation_count_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_prior_question_explanation_count_dict, handle, protocol=4)\n",
    "    \n",
    "with open(f'{upload_dir}/u_prior_question_explanation_correctly_count_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_prior_question_explanation_correctly_count_dict, handle, protocol=4)\n",
    "    \n",
    "with open(f'{upload_dir}/u_question_listening_correctly_count_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_question_listening_correctly_count_dict, handle, protocol=4)\n",
    "    \n",
    "with open(f'{upload_dir}/u_question_reading_correctly_count_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_question_reading_correctly_count_dict, handle, protocol=4)\n",
    "    \n",
    "with open(f'{upload_dir}/u_question_listening_count_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_question_listening_count_dict, handle, protocol=4)\n",
    "    \n",
    "with open(f'{upload_dir}/u_question_reading_count_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_question_reading_count_dict, handle, protocol=4)\n",
    "    \n",
    "with open(f'{upload_dir}/u_question_incorrect_timestamp_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_question_incorrect_timestamp_dict, handle, protocol=4)\n",
    "    \n",
    "with open(f'{upload_dir}/u_question_incorrect_timestamp2_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_question_incorrect_timestamp2_dict, handle, protocol=4)\n",
    "    \n",
    "with open(f'{upload_dir}/u_question_incorrect_timestamp3_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_question_incorrect_timestamp3_dict, handle, protocol=4)\n",
    "    \n",
    "with open(f'{upload_dir}/u_question_part_correct_timestamp_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_question_part_correct_timestamp_dict, handle, protocol=4)\n",
    "    \n",
    "with open(f'{upload_dir}/u_question_part_incorrect_timestamp_dict.pickle', 'wb') as handle:\n",
    "    pickle.dump(u_question_part_incorrect_timestamp_dict, handle, protocol=4)\n",
    "    \n",
    "with open(f'{upload_dir}/uc_cache.pickle', 'wb') as handle:\n",
    "    pickle.dump(lru_cache, handle, protocol=4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "total 2.2G\r\n",
      "-rw-rw-r-- 1 user01 user01 530K Jan  2 07:52 content_df.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01  124 Jan  1 07:54 dataset-metadata.json\r\n",
      "-rw-rw-r-- 1 user01 user01  88M Jan  2 07:52 lgb_model.txt\r\n",
      "-rw-rw-r-- 1 user01 user01 2.4M Jan  2 07:52 questions_df.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01  15M Jan  2 07:52 u_answered_correctly_count_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01 8.0M Jan  2 07:52 u_answered_count_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01 7.9M Jan  2 07:55 u_prior2_lecture_timestamp_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01  15M Jan  2 07:54 u_prior2_question_correctly_timestamp_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01  15M Jan  2 07:54 u_prior2_question_timestamp_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01  15M Jan  2 07:54 u_prior3_question_correctly_timestamp_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01  15M Jan  2 07:55 u_prior3_question_timestamp_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01  15M Jan  2 07:55 u_prior4_question_timestamp_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01 7.9M Jan  2 07:55 u_prior_lecture_timestamp_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01  15M Jan  2 07:54 u_prior_question_correctly_timestamp_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01 8.0M Jan  2 07:55 u_prior_question_explanation_correctly_count_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01  15M Jan  2 07:55 u_prior_question_explanation_count_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01  15M Jan  2 07:54 u_prior_question_timestamp_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01  15M Jan  2 07:55 u_question_incorrect_timestamp2_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01  15M Jan  2 07:55 u_question_incorrect_timestamp3_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01  15M Jan  2 07:55 u_question_incorrect_timestamp_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01 8.0M Jan  2 07:55 u_question_listening_correctly_count_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01 8.0M Jan  2 07:55 u_question_listening_count_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01  56M Jan  2 07:55 u_question_part_correct_timestamp_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01  59M Jan  2 07:53 u_question_part_correctly_count_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01  29M Jan  2 07:53 u_question_part_count_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01  56M Jan  2 07:56 u_question_part_incorrect_timestamp_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01 8.0M Jan  2 07:55 u_question_reading_correctly_count_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01 8.0M Jan  2 07:55 u_question_reading_count_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01 674M Jan  2 07:52 u_question_seen_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01 520M Jan  2 07:54 u_question_tag1_correctly_count_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01 266M Jan  2 07:54 u_question_tag1_count_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01  15M Jan  2 07:55 u_task_container_id_dict.pickle\r\n",
      "-rw-rw-r-- 1 user01 user01 214M Jan  2 07:56 uc_cache.pickle\r\n"
     ]
    }
   ],
   "source": [
    "!ls -lh {upload_dir}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}