02. 다른 Features 추가, 새로운 DF 생성 (데이터 가공 & 모델링).ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "e54158bf",
   "metadata": {},
   "source": [
    "- STARTDATE, ENDDATE 분리 후 합치기만 함\n",
    "- **STARTDATE ~ ENDDATE 사이 날짜 삽입 해야함.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "7731b399",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings(action='ignore')\n",
    "import tensorflow as tf"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "eae19327",
   "metadata": {},
   "source": [
    "# Data Load && "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "3b6a5934",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "lab = pd.read_csv('/data/MIMIC_III/LABEVENTS.csv')\n",
    "pre = pd.read_csv('/data/MIMIC_III/PRESCRIPTIONS.csv')\n",
    "pro = pd.read_csv('/data/MIMIC_III/PROCEDUREEVENTS_MV.csv')\n",
    "폐렴환자 = pd.read_csv('/project/data/폐렴환자lab.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "efe33b69",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "690"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lab['ITEMID'].nunique()\n",
    "폐렴환자['ITEMID'].nunique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "10d20784",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "7799"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#폐렴환자 id 추출\n",
    "subid = 폐렴환자['SUBJECT_ID'].unique()\n",
    "len(subid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "8f85e61f",
   "metadata": {},
   "outputs": [],
   "source": [
    "import datetime as dt\n",
    "\n",
    "#datetime - date만 추출하는 함수 생성\n",
    "\n",
    "def only_date(time):\n",
    "    time = pd.to_datetime(time)\n",
    "    time = time.dt.date\n",
    "    return time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "fb6fe7c3",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "46894f66",
   "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>SUBJECT_ID</th>\n",
       "      <th>HADM_ID</th>\n",
       "      <th>ITEMID</th>\n",
       "      <th>CHARTTIME</th>\n",
       "      <th>VALUE</th>\n",
       "      <th>VALUENUM</th>\n",
       "      <th>VALUEUOM</th>\n",
       "      <th>FLAG</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>281</td>\n",
       "      <td>3</td>\n",
       "      <td>NaN</td>\n",
       "      <td>50820</td>\n",
       "      <td>2101-10-12 16:07:00</td>\n",
       "      <td>7.39</td>\n",
       "      <td>7.39</td>\n",
       "      <td>units</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>282</td>\n",
       "      <td>3</td>\n",
       "      <td>NaN</td>\n",
       "      <td>50800</td>\n",
       "      <td>2101-10-12 18:17:00</td>\n",
       "      <td>ART</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>283</td>\n",
       "      <td>3</td>\n",
       "      <td>NaN</td>\n",
       "      <td>50802</td>\n",
       "      <td>2101-10-12 18:17:00</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1.00</td>\n",
       "      <td>mEq/L</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>284</td>\n",
       "      <td>3</td>\n",
       "      <td>NaN</td>\n",
       "      <td>50804</td>\n",
       "      <td>2101-10-12 18:17:00</td>\n",
       "      <td>22</td>\n",
       "      <td>22.00</td>\n",
       "      <td>mEq/L</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>285</td>\n",
       "      <td>3</td>\n",
       "      <td>NaN</td>\n",
       "      <td>50808</td>\n",
       "      <td>2101-10-12 18:17:00</td>\n",
       "      <td>0.93</td>\n",
       "      <td>0.93</td>\n",
       "      <td>mmol/L</td>\n",
       "      <td>abnormal</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27854050</th>\n",
       "      <td>27428435</td>\n",
       "      <td>96443</td>\n",
       "      <td>103219.0</td>\n",
       "      <td>50882</td>\n",
       "      <td>2109-12-30 01:40:00</td>\n",
       "      <td>26</td>\n",
       "      <td>26.00</td>\n",
       "      <td>mEq/L</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27854051</th>\n",
       "      <td>27428436</td>\n",
       "      <td>96443</td>\n",
       "      <td>103219.0</td>\n",
       "      <td>50885</td>\n",
       "      <td>2109-12-30 01:40:00</td>\n",
       "      <td>2.1</td>\n",
       "      <td>2.10</td>\n",
       "      <td>mg/dL</td>\n",
       "      <td>abnormal</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27854052</th>\n",
       "      <td>27428437</td>\n",
       "      <td>96443</td>\n",
       "      <td>103219.0</td>\n",
       "      <td>50902</td>\n",
       "      <td>2109-12-30 01:40:00</td>\n",
       "      <td>97</td>\n",
       "      <td>97.00</td>\n",
       "      <td>mEq/L</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27854053</th>\n",
       "      <td>27428438</td>\n",
       "      <td>96443</td>\n",
       "      <td>103219.0</td>\n",
       "      <td>50911</td>\n",
       "      <td>2109-12-30 01:40:00</td>\n",
       "      <td>2</td>\n",
       "      <td>2.00</td>\n",
       "      <td>ng/mL</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27854054</th>\n",
       "      <td>27428439</td>\n",
       "      <td>96443</td>\n",
       "      <td>103219.0</td>\n",
       "      <td>50912</td>\n",
       "      <td>2109-12-30 01:40:00</td>\n",
       "      <td>1.6</td>\n",
       "      <td>1.60</td>\n",
       "      <td>mg/dL</td>\n",
       "      <td>abnormal</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>27854055 rows × 9 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            ROW_ID  SUBJECT_ID   HADM_ID  ITEMID            CHARTTIME VALUE  \\\n",
       "0              281           3       NaN   50820  2101-10-12 16:07:00  7.39   \n",
       "1              282           3       NaN   50800  2101-10-12 18:17:00   ART   \n",
       "2              283           3       NaN   50802  2101-10-12 18:17:00    -1   \n",
       "3              284           3       NaN   50804  2101-10-12 18:17:00    22   \n",
       "4              285           3       NaN   50808  2101-10-12 18:17:00  0.93   \n",
       "...            ...         ...       ...     ...                  ...   ...   \n",
       "27854050  27428435       96443  103219.0   50882  2109-12-30 01:40:00    26   \n",
       "27854051  27428436       96443  103219.0   50885  2109-12-30 01:40:00   2.1   \n",
       "27854052  27428437       96443  103219.0   50902  2109-12-30 01:40:00    97   \n",
       "27854053  27428438       96443  103219.0   50911  2109-12-30 01:40:00     2   \n",
       "27854054  27428439       96443  103219.0   50912  2109-12-30 01:40:00   1.6   \n",
       "\n",
       "          VALUENUM VALUEUOM      FLAG  \n",
       "0             7.39    units       NaN  \n",
       "1              NaN      NaN       NaN  \n",
       "2            -1.00    mEq/L       NaN  \n",
       "3            22.00    mEq/L       NaN  \n",
       "4             0.93   mmol/L  abnormal  \n",
       "...            ...      ...       ...  \n",
       "27854050     26.00    mEq/L       NaN  \n",
       "27854051      2.10    mg/dL  abnormal  \n",
       "27854052     97.00    mEq/L       NaN  \n",
       "27854053      2.00    ng/mL       NaN  \n",
       "27854054      1.60    mg/dL  abnormal  \n",
       "\n",
       "[27854055 rows x 9 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lab"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "e5013616",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DRUG\n",
       "Potassium Chloride             192993\n",
       "Insulin                        143465\n",
       "D5W                            142241\n",
       "Furosemide                     133122\n",
       "0.9% Sodium Chloride           130147\n",
       "NS                             129731\n",
       "Magnesium Sulfate               90427\n",
       "Iso-Osmotic Dextrose            87005\n",
       "Sodium Chloride 0.9%  Flush     83392\n",
       "Acetaminophen                   78768\n",
       "Name: SUBJECT_ID, dtype: int64"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pre.groupby(['DRUG'])['SUBJECT_ID'].count().sort_values(ascending=False)[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "a934adeb",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ROW_ID</th>\n",
       "      <th>SUBJECT_ID</th>\n",
       "      <th>HADM_ID</th>\n",
       "      <th>ICUSTAY_ID</th>\n",
       "      <th>STARTTIME</th>\n",
       "      <th>ENDTIME</th>\n",
       "      <th>ITEMID</th>\n",
       "      <th>VALUE</th>\n",
       "      <th>VALUEUOM</th>\n",
       "      <th>LOCATION</th>\n",
       "      <th>...</th>\n",
       "      <th>ORDERCATEGORYNAME</th>\n",
       "      <th>SECONDARYORDERCATEGORYNAME</th>\n",
       "      <th>ORDERCATEGORYDESCRIPTION</th>\n",
       "      <th>ISOPENBAG</th>\n",
       "      <th>CONTINUEINNEXTDEPT</th>\n",
       "      <th>CANCELREASON</th>\n",
       "      <th>STATUSDESCRIPTION</th>\n",
       "      <th>COMMENTS_EDITEDBY</th>\n",
       "      <th>COMMENTS_CANCELEDBY</th>\n",
       "      <th>COMMENTS_DATE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>379</td>\n",
       "      <td>29070</td>\n",
       "      <td>115071</td>\n",
       "      <td>232563.0</td>\n",
       "      <td>2145-03-12 23:04:00</td>\n",
       "      <td>2145-03-12 23:05:00</td>\n",
       "      <td>225401</td>\n",
       "      <td>1.0</td>\n",
       "      <td>None</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>Procedures</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Electrolytes</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>FinishedRunning</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>380</td>\n",
       "      <td>29070</td>\n",
       "      <td>115071</td>\n",
       "      <td>232563.0</td>\n",
       "      <td>2145-03-12 23:04:00</td>\n",
       "      <td>2145-03-12 23:05:00</td>\n",
       "      <td>225454</td>\n",
       "      <td>1.0</td>\n",
       "      <td>None</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>Procedures</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Electrolytes</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>FinishedRunning</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>381</td>\n",
       "      <td>29070</td>\n",
       "      <td>115071</td>\n",
       "      <td>232563.0</td>\n",
       "      <td>2145-03-12 23:05:00</td>\n",
       "      <td>2145-03-18 20:01:00</td>\n",
       "      <td>225792</td>\n",
       "      <td>8456.0</td>\n",
       "      <td>hour</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>Ventilation</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Task</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>FinishedRunning</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>382</td>\n",
       "      <td>29070</td>\n",
       "      <td>115071</td>\n",
       "      <td>232563.0</td>\n",
       "      <td>2145-03-12 23:36:00</td>\n",
       "      <td>2145-03-12 23:37:00</td>\n",
       "      <td>225402</td>\n",
       "      <td>1.0</td>\n",
       "      <td>None</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>Procedures</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Electrolytes</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>FinishedRunning</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>383</td>\n",
       "      <td>29070</td>\n",
       "      <td>115071</td>\n",
       "      <td>232563.0</td>\n",
       "      <td>2145-03-13 01:27:00</td>\n",
       "      <td>2145-03-16 16:00:00</td>\n",
       "      <td>224560</td>\n",
       "      <td>5193.0</td>\n",
       "      <td>min</td>\n",
       "      <td>Right IJ</td>\n",
       "      <td>...</td>\n",
       "      <td>Invasive Lines</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Task</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>FinishedRunning</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>258061</th>\n",
       "      <td>257337</td>\n",
       "      <td>41958</td>\n",
       "      <td>182711</td>\n",
       "      <td>246007.0</td>\n",
       "      <td>2155-08-09 01:51:00</td>\n",
       "      <td>2155-08-10 11:20:00</td>\n",
       "      <td>225204</td>\n",
       "      <td>2009.0</td>\n",
       "      <td>min</td>\n",
       "      <td>Right Antecube</td>\n",
       "      <td>...</td>\n",
       "      <td>Invasive Lines</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Task</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>FinishedRunning</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>258062</th>\n",
       "      <td>257338</td>\n",
       "      <td>41958</td>\n",
       "      <td>182711</td>\n",
       "      <td>246007.0</td>\n",
       "      <td>2155-08-09 01:57:00</td>\n",
       "      <td>2155-08-09 01:58:00</td>\n",
       "      <td>225402</td>\n",
       "      <td>1.0</td>\n",
       "      <td>None</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>Procedures</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Electrolytes</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>FinishedRunning</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>258063</th>\n",
       "      <td>257339</td>\n",
       "      <td>41958</td>\n",
       "      <td>182711</td>\n",
       "      <td>246007.0</td>\n",
       "      <td>2155-08-09 01:57:00</td>\n",
       "      <td>2155-08-09 01:58:00</td>\n",
       "      <td>225459</td>\n",
       "      <td>1.0</td>\n",
       "      <td>None</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>Imaging</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Electrolytes</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>FinishedRunning</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>258064</th>\n",
       "      <td>257340</td>\n",
       "      <td>41958</td>\n",
       "      <td>182711</td>\n",
       "      <td>246007.0</td>\n",
       "      <td>2155-08-09 01:57:00</td>\n",
       "      <td>2155-08-09 01:58:00</td>\n",
       "      <td>225966</td>\n",
       "      <td>1.0</td>\n",
       "      <td>None</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>Procedures</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Electrolytes</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>Rewritten</td>\n",
       "      <td>NaN</td>\n",
       "      <td>RN</td>\n",
       "      <td>2155-08-09 02:20:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>258065</th>\n",
       "      <td>257341</td>\n",
       "      <td>41958</td>\n",
       "      <td>182711</td>\n",
       "      <td>246007.0</td>\n",
       "      <td>2155-08-09 02:20:00</td>\n",
       "      <td>2155-08-09 02:21:00</td>\n",
       "      <td>225401</td>\n",
       "      <td>1.0</td>\n",
       "      <td>None</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>Procedures</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Electrolytes</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>FinishedRunning</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>258066 rows × 25 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        ROW_ID  SUBJECT_ID  HADM_ID  ICUSTAY_ID            STARTTIME  \\\n",
       "0          379       29070   115071    232563.0  2145-03-12 23:04:00   \n",
       "1          380       29070   115071    232563.0  2145-03-12 23:04:00   \n",
       "2          381       29070   115071    232563.0  2145-03-12 23:05:00   \n",
       "3          382       29070   115071    232563.0  2145-03-12 23:36:00   \n",
       "4          383       29070   115071    232563.0  2145-03-13 01:27:00   \n",
       "...        ...         ...      ...         ...                  ...   \n",
       "258061  257337       41958   182711    246007.0  2155-08-09 01:51:00   \n",
       "258062  257338       41958   182711    246007.0  2155-08-09 01:57:00   \n",
       "258063  257339       41958   182711    246007.0  2155-08-09 01:57:00   \n",
       "258064  257340       41958   182711    246007.0  2155-08-09 01:57:00   \n",
       "258065  257341       41958   182711    246007.0  2155-08-09 02:20:00   \n",
       "\n",
       "                    ENDTIME  ITEMID   VALUE VALUEUOM         LOCATION  ...  \\\n",
       "0       2145-03-12 23:05:00  225401     1.0     None              NaN  ...   \n",
       "1       2145-03-12 23:05:00  225454     1.0     None              NaN  ...   \n",
       "2       2145-03-18 20:01:00  225792  8456.0     hour              NaN  ...   \n",
       "3       2145-03-12 23:37:00  225402     1.0     None              NaN  ...   \n",
       "4       2145-03-16 16:00:00  224560  5193.0      min         Right IJ  ...   \n",
       "...                     ...     ...     ...      ...              ...  ...   \n",
       "258061  2155-08-10 11:20:00  225204  2009.0      min  Right Antecube   ...   \n",
       "258062  2155-08-09 01:58:00  225402     1.0     None              NaN  ...   \n",
       "258063  2155-08-09 01:58:00  225459     1.0     None              NaN  ...   \n",
       "258064  2155-08-09 01:58:00  225966     1.0     None              NaN  ...   \n",
       "258065  2155-08-09 02:21:00  225401     1.0     None              NaN  ...   \n",
       "\n",
       "       ORDERCATEGORYNAME SECONDARYORDERCATEGORYNAME  ORDERCATEGORYDESCRIPTION  \\\n",
       "0             Procedures                        NaN              Electrolytes   \n",
       "1             Procedures                        NaN              Electrolytes   \n",
       "2            Ventilation                        NaN                      Task   \n",
       "3             Procedures                        NaN              Electrolytes   \n",
       "4         Invasive Lines                        NaN                      Task   \n",
       "...                  ...                        ...                       ...   \n",
       "258061    Invasive Lines                        NaN                      Task   \n",
       "258062        Procedures                        NaN              Electrolytes   \n",
       "258063           Imaging                        NaN              Electrolytes   \n",
       "258064        Procedures                        NaN              Electrolytes   \n",
       "258065        Procedures                        NaN              Electrolytes   \n",
       "\n",
       "        ISOPENBAG  CONTINUEINNEXTDEPT CANCELREASON  STATUSDESCRIPTION  \\\n",
       "0               0                   0            0    FinishedRunning   \n",
       "1               0                   0            0    FinishedRunning   \n",
       "2               1                   0            0    FinishedRunning   \n",
       "3               0                   0            0    FinishedRunning   \n",
       "4               1                   0            0    FinishedRunning   \n",
       "...           ...                 ...          ...                ...   \n",
       "258061          1                   0            0    FinishedRunning   \n",
       "258062          0                   0            0    FinishedRunning   \n",
       "258063          0                   0            0    FinishedRunning   \n",
       "258064          0                   0            1          Rewritten   \n",
       "258065          0                   0            0    FinishedRunning   \n",
       "\n",
       "       COMMENTS_EDITEDBY  COMMENTS_CANCELEDBY        COMMENTS_DATE  \n",
       "0                    NaN                  NaN                  NaN  \n",
       "1                    NaN                  NaN                  NaN  \n",
       "2                    NaN                  NaN                  NaN  \n",
       "3                    NaN                  NaN                  NaN  \n",
       "4                    NaN                  NaN                  NaN  \n",
       "...                  ...                  ...                  ...  \n",
       "258061               NaN                  NaN                  NaN  \n",
       "258062               NaN                  NaN                  NaN  \n",
       "258063               NaN                  NaN                  NaN  \n",
       "258064               NaN                   RN  2155-08-09 02:20:00  \n",
       "258065               NaN                  NaN                  NaN  \n",
       "\n",
       "[258066 rows x 25 columns]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pro"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a8960603",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "id": "881da966",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[SHORT_TITLE\n",
       " Op skl base fx-deep coma    5\n",
       " Opn skul fx NEC-brf coma    5\n",
       " Cl skl w oth fx-mod coma    5\n",
       " Cl skl w oth fx-coma NOS    5\n",
       " Cl skl w oth fx-brf coma    5\n",
       " Cl skl w oth fx w/o coma    5\n",
       " Opn skul/oth fx-coma NOS    5\n",
       " Opn skul/oth fx-brf coma    5\n",
       " Opn skul/oth fx w/o coma    5\n",
       " Cl skl vlt fx-proln coma    5\n",
       " Name: ICD9_CODE, dtype: int64]"
      ]
     },
     "execution_count": 94,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "d_diag = pd.read_csv('/project/data/D_ICD_DIAGNOSES.csv') #선행\n",
    "[d_diag.groupby(['SHORT_TITLE'])['ICD9_CODE'].count().sort_values(ascending = False)[:10]]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "91e21c53",
   "metadata": {},
   "source": [
    "# LAB2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "62116016",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "7799"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(subid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "b6a16b2f",
   "metadata": {},
   "outputs": [],
   "source": [
    "lab2 = lab[(lab['SUBJECT_ID'].isin(subid))&(lab['FLAG']=='abnormal')].reset_index(drop=True)\n",
    "lab2 = lab2[['SUBJECT_ID','ITEMID','CHARTTIME']]\n",
    "lab2['Type'] = 'LAB'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "d4fa15f5",
   "metadata": {},
   "outputs": [],
   "source": [
    "lab2['CHARTTIME'] = only_date(lab2['CHARTTIME'])\n",
    "lab2 = lab2.rename(columns={'CHARTTIME':'DATE', 'ITEMID':'ITEM'})\n",
    "lab2 = lab2[['SUBJECT_ID','DATE','ITEM','Type']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "0356d0f2",
   "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>SUBJECT_ID</th>\n",
       "      <th>DATE</th>\n",
       "      <th>ITEM</th>\n",
       "      <th>Type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>9</td>\n",
       "      <td>2149-11-14</td>\n",
       "      <td>50821</td>\n",
       "      <td>LAB</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>9</td>\n",
       "      <td>2149-11-14</td>\n",
       "      <td>50910</td>\n",
       "      <td>LAB</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>9</td>\n",
       "      <td>2149-11-14</td>\n",
       "      <td>50911</td>\n",
       "      <td>LAB</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>9</td>\n",
       "      <td>2149-11-14</td>\n",
       "      <td>50964</td>\n",
       "      <td>LAB</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>9</td>\n",
       "      <td>2149-11-14</td>\n",
       "      <td>50821</td>\n",
       "      <td>LAB</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3406604</th>\n",
       "      <td>96442</td>\n",
       "      <td>2115-06-29</td>\n",
       "      <td>51221</td>\n",
       "      <td>LAB</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3406605</th>\n",
       "      <td>96442</td>\n",
       "      <td>2115-06-29</td>\n",
       "      <td>51222</td>\n",
       "      <td>LAB</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3406606</th>\n",
       "      <td>96442</td>\n",
       "      <td>2115-06-29</td>\n",
       "      <td>51244</td>\n",
       "      <td>LAB</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3406607</th>\n",
       "      <td>96442</td>\n",
       "      <td>2115-06-29</td>\n",
       "      <td>51254</td>\n",
       "      <td>LAB</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3406608</th>\n",
       "      <td>96442</td>\n",
       "      <td>2115-06-29</td>\n",
       "      <td>51256</td>\n",
       "      <td>LAB</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3406609 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         SUBJECT_ID        DATE   ITEM Type\n",
       "0                 9  2149-11-14  50821  LAB\n",
       "1                 9  2149-11-14  50910  LAB\n",
       "2                 9  2149-11-14  50911  LAB\n",
       "3                 9  2149-11-14  50964  LAB\n",
       "4                 9  2149-11-14  50821  LAB\n",
       "...             ...         ...    ...  ...\n",
       "3406604       96442  2115-06-29  51221  LAB\n",
       "3406605       96442  2115-06-29  51222  LAB\n",
       "3406606       96442  2115-06-29  51244  LAB\n",
       "3406607       96442  2115-06-29  51254  LAB\n",
       "3406608       96442  2115-06-29  51256  LAB\n",
       "\n",
       "[3406609 rows x 4 columns]"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lab2"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9e25d3f6",
   "metadata": {},
   "source": [
    "# PRE2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "8e4feb16",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1329162, 4)"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pre2 = pre[pre['SUBJECT_ID'].isin(subid)].reset_index(drop=True)\n",
    "pre2 = pre2[['SUBJECT_ID','STARTDATE','ENDDATE','NDC']]\n",
    "\n",
    "#time drop, yyyy-mm-dd로 변경\n",
    "pre2['STARTDATE'], pre2['ENDDATE'] = only_date(pre2['STARTDATE']), only_date(pre2['ENDDATE'])\n",
    "\n",
    "pre2.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "21111b7c",
   "metadata": {},
   "outputs": [],
   "source": [
    "#enddate가 null인 값은 startdate 삽입\n",
    "end_null = pre2[(pre2['ENDDATE'].isnull())&(pre2['STARTDATE'].notnull())]\n",
    "end_null['ENDDATE'] = end_null['STARTDATE']\n",
    "\n",
    "start_null = pre2[(pre2['STARTDATE'].isnull())&(pre2['ENDDATE'].notnull())]\n",
    "start_null['STARTDATE'] = start_null['ENDDATE']\n",
    "\n",
    "pre2.loc[end_null.index] = end_null\n",
    "pre2.loc[start_null.index] = start_null"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "d392680a",
   "metadata": {},
   "outputs": [],
   "source": [
    "#drop null\n",
    "# pre2 = pre2.drop(pre2[pre2['STARTDATE'].isnull()].index).sort_values(['SUBJECT_ID','STARTDATE']).reset_index(drop=True)\n",
    "pre2 = pre2.dropna().sort_values(['SUBJECT_ID','STARTDATE']).reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "e95f893d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1327249, 4)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "SUBJECT_ID    0\n",
       "STARTDATE     0\n",
       "ENDDATE       0\n",
       "NDC           0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(pre2.shape)\n",
    "pre2.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "17ff7f53",
   "metadata": {},
   "outputs": [],
   "source": [
    "from datetime import datetime, timedelta\n",
    "\n",
    "def date_range(start, end):\n",
    "    dates = [(start + timedelta(days=i)).strftime(\"%Y-%m-%d\") for i in range((end-start).days+1)]\n",
    "    return dates"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "7315ec75",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(datetime.date(2149, 11, 9), datetime.date(2149, 11, 11))"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pre2['STARTDATE'][10], pre2['ENDDATE'][10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "456a124d",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['2149-11-09', '2149-11-10', '2149-11-11']"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "date_range(pre2['STARTDATE'][10], pre2['ENDDATE'][10])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "05da2e34",
   "metadata": {},
   "outputs": [],
   "source": [
    "# pre2.to_csv('./pre2.csv', index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f5a793a0",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "77db71c1",
   "metadata": {},
   "outputs": [],
   "source": [
    "# pre_dict={}\n",
    "# lst_time = []\n",
    "\n",
    "# with open('./pre2.csv', 'r') as f:\n",
    "#     line=f.readline()\n",
    "#     for line in f:\n",
    "#         sub, start, end, ndc = line[0], line[1], line[2], line[3]\n",
    "#         d = date_range(pre2['STARTDATE'][i], pre2['ENDDATE'][i])\n",
    "#         pre2_dic[i] = d"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e870b5d4",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "fe8cfaee",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# index에 맞춰서 itemid 넣어주려고 dic 생성, key: index, value: 기간\n",
    "pre2_dic = {}\n",
    "for i in range(len(pre2)):\n",
    "    d = date_range(pre2['STARTDATE'][i], pre2['ENDDATE'][i])\n",
    "    pre2_dic[i] = d\n",
    "    \n",
    "lst_time = []\n",
    "for i in range(len(pre2)):\n",
    "    lst_time.extend(pre2_dic[i][:])\n",
    "    \n",
    "# pre2 NDC열에 들어갈 lst\n",
    "lst_itemid = []\n",
    "for i in range(len(pre2)):\n",
    "    for j in pre2_dic[i]:\n",
    "        lst_itemid.append(pre2['NDC'][i])\n",
    "        \n",
    "# pre2 subid열에 들어갈 lst\n",
    "lst_subid = []\n",
    "for i in range(len(pre2)):\n",
    "    for j in pre2_dic[i]:\n",
    "        lst_subid.append(pre2['SUBJECT_ID'][i])\n",
    "\n",
    "pre2sub = pd.DataFrame(lst_subid)\n",
    "pre2time = pd.DataFrame(lst_time)\n",
    "pre2item = pd.DataFrame(lst_itemid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "6ff3d01e",
   "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>SUBJECT_ID</th>\n",
       "      <th>DATE</th>\n",
       "      <th>ITEM</th>\n",
       "      <th>Type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>9</td>\n",
       "      <td>2149-11-09</td>\n",
       "      <td>0.0</td>\n",
       "      <td>PRE</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>9</td>\n",
       "      <td>2149-11-09</td>\n",
       "      <td>85036207.0</td>\n",
       "      <td>PRE</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>9</td>\n",
       "      <td>2149-11-09</td>\n",
       "      <td>456066270.0</td>\n",
       "      <td>PRE</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>9</td>\n",
       "      <td>2149-11-09</td>\n",
       "      <td>456066270.0</td>\n",
       "      <td>PRE</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>9</td>\n",
       "      <td>2149-11-09</td>\n",
       "      <td>338001702.0</td>\n",
       "      <td>PRE</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5650364</th>\n",
       "      <td>99985</td>\n",
       "      <td>2181-02-10</td>\n",
       "      <td>0.0</td>\n",
       "      <td>PRE</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5650365</th>\n",
       "      <td>99985</td>\n",
       "      <td>2181-02-11</td>\n",
       "      <td>0.0</td>\n",
       "      <td>PRE</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5650366</th>\n",
       "      <td>99985</td>\n",
       "      <td>2181-02-12</td>\n",
       "      <td>0.0</td>\n",
       "      <td>PRE</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5650367</th>\n",
       "      <td>99985</td>\n",
       "      <td>2181-02-11</td>\n",
       "      <td>781305714.0</td>\n",
       "      <td>PRE</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5650368</th>\n",
       "      <td>99985</td>\n",
       "      <td>2181-02-12</td>\n",
       "      <td>781305714.0</td>\n",
       "      <td>PRE</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5650369 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         SUBJECT_ID        DATE         ITEM Type\n",
       "0                 9  2149-11-09          0.0  PRE\n",
       "1                 9  2149-11-09   85036207.0  PRE\n",
       "2                 9  2149-11-09  456066270.0  PRE\n",
       "3                 9  2149-11-09  456066270.0  PRE\n",
       "4                 9  2149-11-09  338001702.0  PRE\n",
       "...             ...         ...          ...  ...\n",
       "5650364       99985  2181-02-10          0.0  PRE\n",
       "5650365       99985  2181-02-11          0.0  PRE\n",
       "5650366       99985  2181-02-12          0.0  PRE\n",
       "5650367       99985  2181-02-11  781305714.0  PRE\n",
       "5650368       99985  2181-02-12  781305714.0  PRE\n",
       "\n",
       "[5650369 rows x 4 columns]"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pre2_completed = pd.concat([pre2sub, pre2time, pre2item], axis = 1)\n",
    "pre2_completed.columns = ['SUBJECT_ID', 'DATE', 'ITEM']\n",
    "pre2_completed['Type'] = 'PRE'\n",
    "pre2_completed"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0aa9200a",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "cfbe4df3",
   "metadata": {},
   "source": [
    "# PRO2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "57da9744",
   "metadata": {},
   "outputs": [],
   "source": [
    "pro2 = pro[pro['SUBJECT_ID'].isin(subid)].reset_index(drop=True)\n",
    "pro2 = pro2[['SUBJECT_ID','STARTTIME','ENDTIME','ITEMID']]\n",
    "pro2['STARTTIME'], pro2['ENDTIME'] = only_date(pro2['STARTTIME']), only_date(pro2['ENDTIME'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "ea99b672",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SUBJECT_ID    0\n",
       "STARTTIME     0\n",
       "ENDTIME       0\n",
       "ITEMID        0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pro2.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "e68ca95d",
   "metadata": {},
   "outputs": [],
   "source": [
    "# index에 맞춰서 itemid 넣어주려고 dic 생성, key: index, value: 기간\n",
    "pro2_dic = {}\n",
    "for i in range(len(pro2)):\n",
    "    d = date_range(pro2['STARTTIME'][i], pro2['ENDTIME'][i])\n",
    "    pro2_dic[i] = d\n",
    "    \n",
    "lst_time = []\n",
    "for i in range(len(pro2)):\n",
    "    lst_time.extend(pro2_dic[i][:])\n",
    "    \n",
    "# pro2 NDC열에 들어갈 lst\n",
    "lst_itemid = []\n",
    "for i in range(len(pro2)):\n",
    "    for j in pro2_dic[i]:\n",
    "        lst_itemid.append(pro2['ITEMID'][i])\n",
    "        \n",
    "# pro2 subid열에 들어갈 lst\n",
    "lst_subid = []\n",
    "for i in range(len(pro2)):\n",
    "    for j in pro2_dic[i]:\n",
    "        lst_subid.append(pro2['SUBJECT_ID'][i])\n",
    "\n",
    "pro2sub = pd.DataFrame(lst_subid)\n",
    "pro2time = pd.DataFrame(lst_time)\n",
    "pro2item = pd.DataFrame(lst_itemid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "ac9cae52",
   "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>SUBJECT_ID</th>\n",
       "      <th>DATE</th>\n",
       "      <th>ITEM</th>\n",
       "      <th>Type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>31608</td>\n",
       "      <td>2133-01-10</td>\n",
       "      <td>224385</td>\n",
       "      <td>PRO</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>31608</td>\n",
       "      <td>2133-01-10</td>\n",
       "      <td>225792</td>\n",
       "      <td>PRO</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>31608</td>\n",
       "      <td>2133-01-11</td>\n",
       "      <td>225792</td>\n",
       "      <td>PRO</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>31608</td>\n",
       "      <td>2133-01-12</td>\n",
       "      <td>225792</td>\n",
       "      <td>PRO</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>31608</td>\n",
       "      <td>2133-01-13</td>\n",
       "      <td>225792</td>\n",
       "      <td>PRO</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>164937</th>\n",
       "      <td>41958</td>\n",
       "      <td>2155-08-10</td>\n",
       "      <td>225204</td>\n",
       "      <td>PRO</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>164938</th>\n",
       "      <td>41958</td>\n",
       "      <td>2155-08-09</td>\n",
       "      <td>225402</td>\n",
       "      <td>PRO</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>164939</th>\n",
       "      <td>41958</td>\n",
       "      <td>2155-08-09</td>\n",
       "      <td>225459</td>\n",
       "      <td>PRO</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>164940</th>\n",
       "      <td>41958</td>\n",
       "      <td>2155-08-09</td>\n",
       "      <td>225966</td>\n",
       "      <td>PRO</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>164941</th>\n",
       "      <td>41958</td>\n",
       "      <td>2155-08-09</td>\n",
       "      <td>225401</td>\n",
       "      <td>PRO</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>164942 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        SUBJECT_ID        DATE    ITEM Type\n",
       "0            31608  2133-01-10  224385  PRO\n",
       "1            31608  2133-01-10  225792  PRO\n",
       "2            31608  2133-01-11  225792  PRO\n",
       "3            31608  2133-01-12  225792  PRO\n",
       "4            31608  2133-01-13  225792  PRO\n",
       "...            ...         ...     ...  ...\n",
       "164937       41958  2155-08-10  225204  PRO\n",
       "164938       41958  2155-08-09  225402  PRO\n",
       "164939       41958  2155-08-09  225459  PRO\n",
       "164940       41958  2155-08-09  225966  PRO\n",
       "164941       41958  2155-08-09  225401  PRO\n",
       "\n",
       "[164942 rows x 4 columns]"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pro2_completed = pd.concat([pro2sub, pro2time, pro2item], axis = 1)\n",
    "pro2_completed.columns = ['SUBJECT_ID', 'DATE', 'ITEM']\n",
    "pro2_completed['Type'] = 'PRO'\n",
    "pro2_completed"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d36f79d2",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "07d8c63e",
   "metadata": {},
   "source": [
    "# Merge"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "d2ef489d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(Index(['SUBJECT_ID', 'DATE', 'ITEM', 'Type'], dtype='object'),\n",
       " Index(['SUBJECT_ID', 'DATE', 'ITEM', 'Type'], dtype='object'),\n",
       " Index(['SUBJECT_ID', 'DATE', 'ITEM', 'Type'], dtype='object'))"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lab2.columns, pre2_completed.columns, pro2_completed.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "a2d90d6b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "4813536"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(lab2)+len(pro2)+len(pre2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "1195f8ab",
   "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>SUBJECT_ID</th>\n",
       "      <th>DATE</th>\n",
       "      <th>ITEM</th>\n",
       "      <th>Type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>9</td>\n",
       "      <td>2149-11-09</td>\n",
       "      <td>50882.0</td>\n",
       "      <td>LAB</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>9</td>\n",
       "      <td>2149-11-09</td>\n",
       "      <td>50931.0</td>\n",
       "      <td>LAB</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>9</td>\n",
       "      <td>2149-11-09</td>\n",
       "      <td>50971.0</td>\n",
       "      <td>LAB</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>9</td>\n",
       "      <td>2149-11-09</td>\n",
       "      <td>51275.0</td>\n",
       "      <td>LAB</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>9</td>\n",
       "      <td>2149-11-09</td>\n",
       "      <td>51288.0</td>\n",
       "      <td>LAB</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9221915</th>\n",
       "      <td>99985</td>\n",
       "      <td>2181-03-06</td>\n",
       "      <td>51288.0</td>\n",
       "      <td>LAB</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9221916</th>\n",
       "      <td>99985</td>\n",
       "      <td>2182-03-14</td>\n",
       "      <td>51279.0</td>\n",
       "      <td>LAB</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9221917</th>\n",
       "      <td>99985</td>\n",
       "      <td>2182-03-14</td>\n",
       "      <td>51006.0</td>\n",
       "      <td>LAB</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9221918</th>\n",
       "      <td>99985</td>\n",
       "      <td>2182-03-14</td>\n",
       "      <td>51221.0</td>\n",
       "      <td>LAB</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9221919</th>\n",
       "      <td>99985</td>\n",
       "      <td>2182-03-14</td>\n",
       "      <td>51222.0</td>\n",
       "      <td>LAB</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>9221920 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         SUBJECT_ID        DATE     ITEM Type\n",
       "0                 9  2149-11-09  50882.0  LAB\n",
       "1                 9  2149-11-09  50931.0  LAB\n",
       "2                 9  2149-11-09  50971.0  LAB\n",
       "3                 9  2149-11-09  51275.0  LAB\n",
       "4                 9  2149-11-09  51288.0  LAB\n",
       "...             ...         ...      ...  ...\n",
       "9221915       99985  2181-03-06  51288.0  LAB\n",
       "9221916       99985  2182-03-14  51279.0  LAB\n",
       "9221917       99985  2182-03-14  51006.0  LAB\n",
       "9221918       99985  2182-03-14  51221.0  LAB\n",
       "9221919       99985  2182-03-14  51222.0  LAB\n",
       "\n",
       "[9221920 rows x 4 columns]"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "m1 = pd.merge(lab2, pre2_completed, on=['SUBJECT_ID','DATE','ITEM','Type'], how='outer').sort_values(['SUBJECT_ID','DATE'])\n",
    "total_data = pd.merge(m1, pro2_completed, on=['SUBJECT_ID','DATE','ITEM','Type'], how='outer').sort_values(['SUBJECT_ID','DATE']).reset_index(drop=True)\n",
    "total_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "072531b1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "4069"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "total_data['ITEM'].nunique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "edc787d1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "7799"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "total_data['SUBJECT_ID'].nunique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "f3d99519",
   "metadata": {},
   "outputs": [],
   "source": [
    "total_data['ITEM'] = total_data['ITEM'].astype(int)\n",
    "total_data.to_csv('./data/merge_data.csv',index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6814a250",
   "metadata": {},
   "source": [
    "# Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "5a3ff37a",
   "metadata": {},
   "outputs": [],
   "source": [
    "total_data = pd.read_csv('../data/merge_data.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fccd996d",
   "metadata": {},
   "source": [
    "### Label"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "767a3066",
   "metadata": {},
   "outputs": [],
   "source": [
    "폐렴환자1 = pd.read_csv('/project/data/폐렴환자.csv')\n",
    "label = 폐렴환자1[폐렴환자1['SUBJECT_ID'].isin(subid)][['SUBJECT_ID','EXPIRE_FLAG']].sort_values('SUBJECT_ID').reset_index()['EXPIRE_FLAG']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "24722fef",
   "metadata": {},
   "outputs": [],
   "source": [
    "np.save(\"./data/2030days_label.npy\",label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "78e2fa03",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(7799, 4069)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "655fcf08",
   "metadata": {},
   "source": [
    "## 시계열"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "ec4e5bab",
   "metadata": {},
   "outputs": [],
   "source": [
    "data = pd.read_csv('../data/merge_data.csv')\n",
    "data = data.set_index('SUBJECT_ID')\n",
    "label = pd.read_csv(\"../data/2030days_label.npy\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "9a0a6703",
   "metadata": {},
   "outputs": [],
   "source": [
    "admission = pd.read_csv('/data/MIMIC_III/ADMISSIONS.csv')\n",
    "# admission.groupby('SUBJECT_ID')['DISCHTIME'].max()\n",
    "admission = admission[admission['SUBJECT_ID'].isin(data.index.unique())]\n",
    "\n",
    "import datetime as dt\n",
    "admission['DISCHTIME'] = pd.to_datetime(admission['DISCHTIME'])\n",
    "admission['DISCHTIME'] = admission['DISCHTIME'].dt.date\n",
    "\n",
    "#final_date\n",
    "dic_sub2final_date = dict(admission.groupby('SUBJECT_ID')['DISCHTIME'].max())\n",
    "\n",
    "#sub2idx\n",
    "dic_sub2idx = {}\n",
    "for i,j in enumerate(data.index.unique()):\n",
    "    dic_sub2idx[j] = i\n",
    "\n",
    "#item2idx\n",
    "dic_item2idx = {}\n",
    "for i,j in enumerate(data['ITEM'].sort_values().unique()):\n",
    "    dic_item2idx[j] = i"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8862d570",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "1b8eed58",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "dic_sub2idx length :  7799\n",
      "dic_item2idx length :  4069\n",
      "dic_sub2final_date length :  7799\n"
     ]
    }
   ],
   "source": [
    "print('dic_sub2idx length : ', len(dic_sub2idx))\n",
    "print('dic_item2idx length : ', len(dic_item2idx))\n",
    "print('dic_sub2final_date length : ', len(dic_sub2final_date))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "f314757d",
   "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>DATE</th>\n",
       "      <th>ITEM</th>\n",
       "      <th>Type</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SUBJECT_ID</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>2149-11-09</td>\n",
       "      <td>50882</td>\n",
       "      <td>LAB</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  DATE   ITEM Type\n",
       "SUBJECT_ID                        \n",
       "9           2149-11-09  50882  LAB"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "82a6eb87",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "35"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dic_item2idx[50883]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5616acbf",
   "metadata": {},
   "source": [
    "### 10일치"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "id": "36271945",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "time : 0.2058424433072408 분\n",
      "2042722.0\n"
     ]
    }
   ],
   "source": [
    "from datetime import timedelta\n",
    "import time\n",
    "from tqdm.auto import tqdm\n",
    "import datetime\n",
    "\n",
    "start = time.time()  # 시작 시간 저장\n",
    "\n",
    "\n",
    "x = np.zeros((len(dic_sub2idx),10,len(dic_item2idx)))\n",
    "x.shape\n",
    "\n",
    "IF=open(\"../data/merge_data.csv\",'r')\n",
    "line=IF.readline()\n",
    "for line in IF:\n",
    "    ss=line.strip(\"\\n\").split(',')\n",
    "    sub, date, item = int(ss[0]), datetime.date.fromisoformat(ss[1]), int(ss[2])\n",
    "    subidx = dic_sub2idx[sub]\n",
    "    itemidx = dic_item2idx[item]\n",
    "    finaldate = dic_sub2final_date[sub]\n",
    "    dateidx = -(finaldate - date).days + 10\n",
    "#     print(finaldate, date, dateidx)\n",
    "    if (dateidx < 0) or (dateidx > 9) :\n",
    "        continue\n",
    "    else:\n",
    "#         print(finaldate, date, dateidx)\n",
    "        x[subidx, dateidx, itemidx] = 1\n",
    "IF.close()\n",
    "        \n",
    "        \n",
    "print(\"time :\", (time.time() - start)/60, '분') \n",
    "print(x.sum())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "78a0ea32",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "34fa70dd",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "ac048b69",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "time : 0.41984949906667074 분\n",
      "2042722.0\n"
     ]
    }
   ],
   "source": [
    "from datetime import timedelta\n",
    "import time\n",
    "from tqdm.auto import tqdm\n",
    "import datetime\n",
    "\n",
    "start = time.time()  # 시작 시간 저장\n",
    "\n",
    "\n",
    "x = np.zeros((len(dic_sub2idx),10,len(dic_item2idx)))\n",
    "x.shape\n",
    "\n",
    "IF=open(\"./data/merge_data.csv\",'r')\n",
    "line=IF.readline()\n",
    "for line in IF:\n",
    "    ss=line.strip(\"\\n\").split(',')\n",
    "    sub, date, item = int(ss[0]), datetime.date.fromisoformat(ss[1]), int(ss[2])\n",
    "\n",
    "    subidx = dic_sub2idx[sub]\n",
    "    itemidx = dic_item2idx[item]\n",
    "    finaldate = dic_sub2final_date[sub]\n",
    "    dateidx = -(finaldate - date).days + 10\n",
    "    \n",
    "    if (dateidx < 0) or (dateidx > 9) :\n",
    "        continue\n",
    "    else:\n",
    "#         print(finaldate, date, dateidx)\n",
    "        x[subidx, dateidx, itemidx] = 1\n",
    "    \n",
    "IF.close()\n",
    "        \n",
    "        \n",
    "print(\"time :\", (time.time() - start)/60, '분') \n",
    "print(x.sum())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "52df4816",
   "metadata": {},
   "outputs": [],
   "source": [
    "# np.save(\"./data/10days_data.npy\",x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "4f25a82e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(7799, 10, 4069)"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "65901a62",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5459.299999999999"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x.shape[0]*0.7"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "id": "62772bd6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((5460, 10, 4069), (1, 5), (2339, 10, 4069), (0, 5))"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 5460 기준으로 split\n",
    "i=5460\n",
    "X_train = x[:i,:,:]\n",
    "y_train = label[:i]\n",
    "\n",
    "\n",
    "X_test = x[i:,:,:]\n",
    "y_test = label[i:]\n",
    "\n",
    "\n",
    "X_train.shape, y_train.shape, X_test.shape, y_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "id": "0ac3c2fc",
   "metadata": {},
   "outputs": [],
   "source": [
    "import math\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, LSTM, Dropout\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from sklearn.metrics import mean_squared_error\n",
    "import pickle\n",
    "import gzip"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "844a9fb4",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# from keras.regularizers import l2\n",
    "\n",
    "# seed_num = 42\n",
    "# np.random.seed(seed_num)\n",
    "# tf.random.set_seed(seed_num)\n",
    "\n",
    "# lstm = Sequential()\n",
    "# lstm.add(LSTM(units=12, activation='tanh', input_shape=(x.shape[1], x.shape[2]),return_sequences=True,\n",
    "#          bias_regularizer=l2(0.001), kernel_regularizer=l2(0.001),recurrent_regularizer=l2(0.001)))\n",
    "# lstm.add(LSTM(units=12, activation='tanh', input_shape=(12, x.shape[2]),\n",
    "#               bias_regularizer=l2(0.001), kernel_regularizer=l2(0.001),recurrent_regularizer=l2(0.001)))\n",
    "# lstm.add(Dense(units=1, activation='sigmoid'))\n",
    "# print(lstm.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "15be216b",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# from keras.callbacks import EarlyStopping\n",
    "\n",
    "# lstm.compile(loss='binary_crossentropy', optimizer='adam',metrics=['acc']) #loss='categorical_crossentropy'\n",
    "# early_stop = EarlyStopping(monitor='val_loss', patience=10, verbose=1)\n",
    "# history = lstm.fit(X_train, y_train, validation_split=0.25, batch_size=128, epochs=2000, callbacks=[early_stop])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "efe5525c",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "lstm (LSTM)                  (None, 8)                 130496    \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 1)                 9         \n",
      "=================================================================\n",
      "Total params: 130,505\n",
      "Trainable params: 130,505\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "None\n",
      "Epoch 1/200\n",
      "32/32 [==============================] - 13s 29ms/step - loss: 0.6220 - acc: 0.6932 - val_loss: 0.6912 - val_acc: 0.5861\n",
      "Epoch 2/200\n",
      "32/32 [==============================] - 0s 12ms/step - loss: 0.5818 - acc: 0.6932 - val_loss: 0.6531 - val_acc: 0.5861\n",
      "Epoch 3/200\n",
      "32/32 [==============================] - 0s 13ms/step - loss: 0.5550 - acc: 0.6932 - val_loss: 0.6247 - val_acc: 0.5861\n",
      "Epoch 4/200\n",
      "32/32 [==============================] - 0s 12ms/step - loss: 0.5343 - acc: 0.6932 - val_loss: 0.6032 - val_acc: 0.5897\n",
      "Epoch 5/200\n",
      "32/32 [==============================] - 0s 12ms/step - loss: 0.5166 - acc: 0.6955 - val_loss: 0.5850 - val_acc: 0.5993\n",
      "Epoch 6/200\n",
      "32/32 [==============================] - 0s 13ms/step - loss: 0.5003 - acc: 0.7039 - val_loss: 0.5696 - val_acc: 0.6205\n",
      "Epoch 7/200\n",
      "32/32 [==============================] - 0s 12ms/step - loss: 0.4854 - acc: 0.7206 - val_loss: 0.5573 - val_acc: 0.6403\n",
      "Epoch 8/200\n",
      "32/32 [==============================] - 0s 13ms/step - loss: 0.4718 - acc: 0.7316 - val_loss: 0.5469 - val_acc: 0.6586\n",
      "Epoch 9/200\n",
      "32/32 [==============================] - 0s 12ms/step - loss: 0.4593 - acc: 0.7468 - val_loss: 0.5380 - val_acc: 0.6894\n",
      "Epoch 10/200\n",
      "32/32 [==============================] - 0s 13ms/step - loss: 0.4473 - acc: 0.7617 - val_loss: 0.5307 - val_acc: 0.7018\n",
      "Epoch 11/200\n",
      "32/32 [==============================] - 0s 13ms/step - loss: 0.4359 - acc: 0.7800 - val_loss: 0.5247 - val_acc: 0.7121\n",
      "Epoch 12/200\n",
      "32/32 [==============================] - 0s 12ms/step - loss: 0.4249 - acc: 0.7970 - val_loss: 0.5201 - val_acc: 0.7231\n",
      "Epoch 13/200\n",
      "32/32 [==============================] - 0s 13ms/step - loss: 0.4143 - acc: 0.8055 - val_loss: 0.5166 - val_acc: 0.7319\n",
      "Epoch 14/200\n",
      "32/32 [==============================] - 0s 13ms/step - loss: 0.4040 - acc: 0.8184 - val_loss: 0.5146 - val_acc: 0.7370\n",
      "Epoch 15/200\n",
      "32/32 [==============================] - 0s 13ms/step - loss: 0.3940 - acc: 0.8267 - val_loss: 0.5139 - val_acc: 0.7392\n",
      "Epoch 16/200\n",
      "32/32 [==============================] - 0s 12ms/step - loss: 0.3843 - acc: 0.8336 - val_loss: 0.5142 - val_acc: 0.7407\n",
      "Epoch 17/200\n",
      "32/32 [==============================] - 0s 13ms/step - loss: 0.3747 - acc: 0.8413 - val_loss: 0.5155 - val_acc: 0.7392\n",
      "Epoch 18/200\n",
      "32/32 [==============================] - 0s 13ms/step - loss: 0.3654 - acc: 0.8440 - val_loss: 0.5178 - val_acc: 0.7407\n",
      "Epoch 19/200\n",
      "32/32 [==============================] - 0s 13ms/step - loss: 0.3561 - acc: 0.8497 - val_loss: 0.5213 - val_acc: 0.7377\n",
      "Epoch 20/200\n",
      "32/32 [==============================] - 0s 13ms/step - loss: 0.3470 - acc: 0.8579 - val_loss: 0.5257 - val_acc: 0.7399\n",
      "Epoch 00020: early stopping\n"
     ]
    }
   ],
   "source": [
    "lstm2 = Sequential()\n",
    "lstm2.add(LSTM(units=8, activation='sigmoid', recurrent_activation='hard_sigmoid', input_shape=(x.shape[1], x.shape[2])))\n",
    "lstm2.add(Dense(units=1, activation='sigmoid'))\n",
    "print(lstm2.summary())\n",
    "\n",
    "from keras.callbacks import EarlyStopping\n",
    "\n",
    "lstm2.compile(loss='binary_crossentropy', optimizer='adam',metrics=['acc']) #loss='categorical_crossentropy'\n",
    "early_stop = EarlyStopping(monitor='val_loss', patience=5, verbose=1)\n",
    "history = lstm2.fit(X_train, y_train, validation_split=0.25, batch_size=128, epochs=200,  callbacks=[early_stop], shuffle=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "68f79b39",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEWCAYAAAB8LwAVAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8rg+JYAAAACXBIWXMAAAsTAAALEwEAmpwYAAA6ZElEQVR4nO3deXhU5fn/8fednSUQskCAEMK+byGAggqIsolQBRW0FaSKWm2rrbZq+1Vr66+2tVZtba27tlRUVERBQVDcFcJO2AlLEiBAQghLErLcvz/OCYxhEgLJZLLcr+uaK3O2OXcmyXxynuec54iqYowxxpQV4O8CjDHG1E4WEMYYY7yygDDGGOOVBYQxxhivLCCMMcZ4ZQFhjDHGKwsIY6pARBJEREUkqBLrzhCRL6v6OsbUFAsI02CIyC4ROSki0WXmr3Y/nBP8VJoxtZIFhGlodgLTSidEpA/Q2H/lGFN7WUCYhuY/wI0e09OB1zxXEJHmIvKaiBwUkd0i8lsRCXCXBYrI4yJySERSgSu8bPuiiOwTkQwR+YOIBJ5rkSLSRkTmi0i2iGwXkVs8lg0WkWQRyRWRTBF5wp0fJiL/FZEsEckRkRUi0upc921MKQsI09B8CzQTkR7uB/dU4L9l1vk70BzoCAzHCZSb3GW3ABOAAUASMKXMtq8ARUBnd53RwM3nUeccIB1o4+7j/4nIpe6yp4CnVLUZ0Al4050/3a27HRAF3Abknce+jQEsIEzDVHoUcTmwCcgoXeARGver6lFV3QX8FfiRu8q1wJOqmqaq2cAfPbZtBYwH7lLV46p6APib+3qVJiLtgGHAr1U1X1XXAC9w+sinEOgsItGqekxVv/WYHwV0VtViVV2pqrnnsm9jPFlAmIboP8D1wAzKNC8B0UAwsNtj3m6grfu8DZBWZlmp9u62+9wmnhzg30DLc6yvDZCtqkfLqeHHQFdgs9uMNMHj+1oEzBGRvSLyZxEJPsd9G3OKBYRpcFR1N05n9XjgnTKLD+H8J97eY148p48y9uE04XguK5UGFADRqhrhPpqpaq9zLHEvECki4d5qUNVtqjoNJ3j+BMwVkSaqWqiqv1PVnsBQnKawGzHmPFlAmIbqx8Clqnrcc6aqFuO06T8qIuEi0h74Baf7Kd4EfiYicSLSArjPY9t9wGLgryLSTEQCRKSTiAw/l8JUNQ34Gvij2/Hc1633vwAi8kMRiVHVEiDH3axEREaKSB+3mSwXJ+hKzmXfxniygDANkqruUNXkchb/FDgOpAJfAv8DXnKXPY/TjLMWWMWZRyA3AiHARuAwMBdofR4lTgMScI4m3gUeUtUl7rKxQIqIHMPpsJ6qqnlArLu/XJy+lc9wmp2MOS9iNwwyxhjjjR1BGGOM8coCwhhjjFcWEMYYY7yygDDGGONVvRlaODo6WhMSEvxdhjHG1CkrV648pKox3pbVm4BISEggObm8sxaNMcZ4IyK7y1vm0yYmERkrIlvc0Sjv87L8byKyxn1sdYcmKF02XUS2uY/pvqzTGGPMmXx2BOFezfkMzoBo6cAKEZmvqhtL11HVuz3W/ynO6JeISCTwEM5omQqsdLc97Kt6jTHGfJ8vjyAGA9tVNVVVT+IMXzypgvWnAa+7z8cAH6tqthsKH+NcPWqMMaaG+LIPoi3fH/UyHRjibUV3vJsOwCcVbNvWy3azgFkA8fHxZRdTWFhIeno6+fn551G+8SYsLIy4uDiCg22QUGPqu9rSST0VmOsOlFZpqvoc8BxAUlLSGWOGpKenEx4eTkJCAiJSPZU2YKpKVlYW6enpdOjQwd/lGGN8zJdNTBl8f1jkODxuzFLGVE43L53rtuXKz88nKirKwqGaiAhRUVF2RGZMA+HLgFgBdBGRDiISghMC88uuJCLdgRbANx6zFwGjRaSFO6TyaHfeObNwqF72fhrTcPgsIFS1CLgT54N9E/CmqqaIyCMiMtFj1anAHPUYVta9lePvcUJmBfCIO88HhZZAbgYUFfjk5Y0xpq7yaR+Eqi4EFpaZ92CZ6YfL2fYlTo/B7zvFhXA8CwqOQXQXkOrLzKysLEaNGgXA/v37CQwMJCbGuWBx+fLlhISElLttcnIyr732Gk8//XS11WOMMeeitnRS+09QKETEw+GdkLsXmsdV20tHRUWxZs0aAB5++GGaNm3KPffcc2p5UVERQUHefwRJSUkkJSVVWy3GGHOubLA+gEYR0CQGjh+EPN9eizdjxgxuu+02hgwZwq9+9SuWL1/OhRdeyIABAxg6dChbtmwBYNmyZUyY4NyL/uGHH2bmzJmMGDGCjh072lGFMaZGNJgjiN+9n8LGvbkVr1R4AvQQBDeuVFNTzzbNeOjKc70fvXP67ddff01gYCC5ubl88cUXBAUFsWTJEh544AHefvvtM7bZvHkzn376KUePHqVbt27cfvvtdi2CMcanGkxAVEpQIyckivKdkPCRa665hsDAQACOHDnC9OnT2bZtGyJCYWGh122uuOIKQkNDCQ0NpWXLlmRmZhIXV33NYcYYU1aDCYhK/6efnwvZO6BxJES090ktTZo0OfX8//7v/xg5ciTvvvsuu3btYsSIEV63CQ0NPfU8MDCQoqIin9RmjDGlrA+irLBm0LQVnMiGE1k+392RI0do29YZReSVV17x+f6MMaayLCC8CW8NIU0hJx0K83y6q1/96lfcf//9DBgwwI4KjDG1inhcn1anJSUladkbBm3atIkePXqc3wsWF8LBzSCBENMNAgKrocr6oUrvqzGmVhGRlarq9Zx6O4IoT2AwtEiA4gLI2QP1JEiNMaayLCAqEhruNDfl58CJQ/6uxhhjapQFxNk0bQWhzeBIBpw87u9qjDGmxlhAnI2Ic7prYDAc3gUl1pFsjGkYLCAqIzDI7Y8ohMO7rT/CGNMgWEBUVkgTaNYGCnLh+AF/V2OMMT5nAXEumsRAWIQz6mvBsUptMnLkSBYt+v69jp588kluv/12r+uPGDGC0tN1x48fT05OzhnrPPzwwzz++OMV7nfevHls3Ljx1PSDDz7IkiVLKlWzMcaABcS5EYGIdhAY6vRHFHsfN8nTtGnTmDNnzvfmzZkzh2nTpp1124ULFxIREXFepZYNiEceeYTLLrvsvF7LGNMwWUCcq4AgiExwOqsP7zprf8SUKVNYsGABJ0+eBGDXrl3s3buX119/naSkJHr16sVDDz3kdduEhAQOHXJOr3300Ufp2rUrF1100akhwQGef/55Bg0aRL9+/Zg8eTInTpzg66+/Zv78+dx7773079+fHTt2MGPGDObOnQvA0qVLGTBgAH369GHmzJkUFBSc2t9DDz1EYmIiffr0YfPmzVV8s4wxdVmDGayPD++D/eur7/VKCp2bC41+FJq1Lne1yMhIBg8ezIcffsikSZOYM2cO1157LQ888ACRkZEUFxczatQo1q1bR9++fb2+xsqVK5kzZw5r1qyhqKiIxMREBg4cCMDVV1/NLbfcAsBvf/tbXnzxRX76058yceJEJkyYwJQpU773Wvn5+cyYMYOlS5fStWtXbrzxRv71r39x1113ARAdHc2qVav45z//yeOPP84LL7xQDW+WMaYusiOI8xUQBEFhcGy/MwJsBTybmUqbl958800SExMZMGAAKSkp32sOKuuLL77gqquuonHjxjRr1oyJE0/f0nvDhg1cfPHF9OnTh9mzZ5OSklJhLVu2bKFDhw507doVgOnTp/P555+fWn711VcDMHDgQHbt2lXhaxlj6jefHkGIyFjgKSAQeEFVH/OyzrXAw4ACa1X1end+MVD6L/8eVZ1YdttzMu6MXVddSTEc2go5uyG6GwR5v8f0pEmTuPvuu1m1ahUnTpwgMjKSxx9/nBUrVtCiRQtmzJhBfn7+eZUwY8YM5s2bR79+/XjllVdYtmxZFb6h08OK25DixhifHUGISCDwDDAO6AlME5GeZdbpAtwPDFPVXsBdHovzVLW/+6haOPhKQCC06ABaAlnboOik19WaNm3KyJEjmTlzJtOmTSM3N5cmTZrQvHlzMjMz+fDDDyvczSWXXMK8efPIy8vj6NGjvP/++6eWHT16lNatW1NYWMjs2bNPzQ8PD+fo0aNnvFa3bt3YtWsX27dvB+A///kPw4cPP5/v3hhTz/myiWkwsF1VU1X1JDAHmFRmnVuAZ1T1MICq1r0LDILDIKqzczSRtQ2KCryuNm3aNNauXcu0adPo168fAwYMoHv37lx//fUMGzaswl0kJiZy3XXX0a9fP8aNG8egQYNOLfv973/PkCFDGDZsGN27dz81f+rUqfzlL39hwIAB7Nix49T8sLAwXn75Za655hr69OlDQEAAt912WxXfBGNMfeSz4b5FZAowVlVvdqd/BAxR1Ts91pkHbAWG4TRDPayqH7nLioA1QBHwmKrO87KPWcAsgPj4+IG7d+/+3vIaHZb65HHI2uEcVUR1hqDQs29TR9lw38bUH7V5uO8goAswApgGPC8iEe6y9m7R1wNPikinshur6nOqmqSqSTExMTVUcjlCmpw+kji0zbmvtTHG1GG+DIgMoJ3HdJw7z1M6MF9VC1V1J87RRBcAVc1wv6YCy4ABPqy1eoQ0hujOgMKh7VBoIWGMqbt8GRArgC4i0kFEQoCpwPwy68zDOXpARKKBrkCqiLQQkVCP+cOA8s8DrUCN3zEvuLFzJIE6fRI+vmVpTasvdyA0xpydzwJCVYuAO4FFwCbgTVVNEZFHRKT0rKRFQJaIbAQ+Be5V1SygB5AsImvd+Y+p6jkHRFhYGFlZWX4IiUYQ1QUQyNpeb0JCVcnKyiIsLMzfpRhjakC9vid1YWEh6enp532NQZUVF8Lxg85wHE1jIND7dRJ1SVhYGHFxcQQHB/u7FGNMNaiok7peD7URHBxMhw4d/FtEdiq8OtEZJvxH86Bton/rMcaYSvL3WUz1X2RHmLEAwprDaz+A9OSzbmKMMbWBBURNaNEeZiyExpFOSOz51t8VGWPMWVlA1JSIdnDTQghvBf+5GnZ95e+KjDGmQhYQNalZG6e5qXkc/HcypH7m74qMMaZcFhA1LTwWZnwAkR3gf9fC9qX+rsgYY7yygPCHpi1h+vvOtRKvT4Oti/1dkTHGnMECwl+aRMP0+dCyO7xxA6x7098VGWPM91hA+FPjSLjxPWibBO/cAu/dCSdP+LsqY4wBLCD8r1ELp7np4l/C6v/CC6Pg4BZ/V2WMMRYQtUJgEIx6EH74Nhw7AM+NgDX/83dVxpgGzgKiNuk8Cm77EtoOhHm3w7u3OTciMsYYP7CAqG2atXb6JYbfB2vnOEcTmSn+rsoY0wBZQNRGAYEw8n4nKPKPwPOXwspXnVFhjTGmhlhA1GYdhztNTvEXwPs/g7dvhoKj/q7KGNNAWEDUdk1bwg/fhUt/CynvwL+Hw751/q7KGNMAWEDUBQEBcMm9MP0DKDwBL1wGK16wJidjjE81+IDILyzmN++uZ+ehOnC2UMIwp8mpw8Ww4Jfw1gynj8IYY3ygwQfEoWMFLFy/j1mvJXOsoMjf5Zxdk2i4/i247Hew6X349yWQscrfVRlj6iGfBoSIjBWRLSKyXUTuK2eda0Vko4ikiMj/POZPF5Ft7mO6r2qMa9GYZ65PJPXQcX7xxhpKSupAs01AAFx0F9z0IRQXwYuj4fPHoeikvyszxtQjPgsIEQkEngHGAT2BaSLSs8w6XYD7gWGq2gu4y50fCTwEDAEGAw+JSAtf1Tq0czS/Gd+DxRszefqTbb7aTfWLHwK3fQHdx8Mnv4fnhsOe7/xdlTGmnvDlEcRgYLuqpqrqSWAOMKnMOrcAz6jqYQBVPeDOHwN8rKrZ7rKPgbE+rJWbhiUwOTGOJ5dsY3HKfl/uqno1joRrX4NpcyA/F14aDR/cDXk5/q7MGFPH+TIg2gJpHtPp7jxPXYGuIvKViHwrImPPYVtEZJaIJItI8sGDB6tUrIjw6FW96RfXnLvfWMO2zDp2vUG3cXDHd3DBHbDyFfjHINjwtp3pZIw5b/7upA4CugAjgGnA8yISUdmNVfU5VU1S1aSYmJgqFxMWHMizPxpIo5AgbnktmSN5hVV+zRoV2hTG/j+45VPn9qZzZ8Lsa+DwLn9XZoypg3wZEBlAO4/pOHeep3RgvqoWqupOYCtOYFRmW59o3bwRz/4wkYycPH72+mqK60KndVlt+sMtn8DYx2DPN/DMBfDlk1BcxwLPGONXvgyIFUAXEekgIiHAVGB+mXXm4Rw9ICLROE1OqcAiYLSItHA7p0e782pEUkIkv5vYm8+2HuQvi+rovRkCAuGC251mp04jYclDzsB/6cn+rswYU0f4LCBUtQi4E+eDfRPwpqqmiMgjIjLRXW0RkCUiG4FPgXtVNUtVs4Hf44TMCuARd16NuX5IPNcPiefZz3bw/tq9Nbnr6tU8Dqa9DtfNhhPZzlXYC+6xC+yMMWclWk86MZOSkjQ5uXr/Oz5ZVMINL3zL+owjvH37UHq1aV6tr1/j8nPh00fhu39DeCyM+xP0mAgi/q7MGOMnIrJSVZO8LfN3J3WtFhIUwD9vGEhEoxBmvbaS7ON1/EK0sGZOKNyy1Lki+80b4fVpkJN29m2NMQ2OBcRZxISH8tyNAzl4rIA7Zq+isLjE3yVVXduBcMsyGP0H2PkZPDMEPv8LFBzzd2XGmFrEAqIS+sZF8Mer+vBNahaPLtjk73KqR2AQDP3p6U7sT/4ATw+A5c/bkB3GGMACotImD4xj5rAOvPL1Lt5KrkdNMhHxMHU2/PhjiO4CC++BfyTB2jegpNjf1Rlj/MgC4hw8ML47wzpH8Zt5G1iTluPvcqpXu8EwYwHc8DaENYd3Z8GzF8OWj+xqbGMaKAuIcxAUGMA/piXSMjyUW/+TzIGj+f4uqXqJQJfLYNZnMOUlKMqD16+Dl8bC7q/9XZ0xpoZZQJyjFk1CeP7GJHLzirj9v6soKKqHzTABAdB7MtyxHCb8zRmq4+VxzrAd+9f7uzpjTA2xgDgPPVo34/Fr+rFy92Eenp9CfbmW5AyBwZA0E3622rlBUdp38OxF8PbNkJ3q7+qMMT5mAXGerujbmjtGduL15WnM/m6Pv8vxrZDGzg2Kfr4WLvoFbPrAGS32g1/A0To0NLox5pxYQFTBLy7vxshuMTw8P4XPt1ZtuPE6oVELuOwh+PkaSJwOq16Fp/rDkt85w3gYY+oVC4gqCAwQnpo2gM4tmzLzlRW8uzrd3yXVjPBYmPAE3LkCekyAL5+AJ3rAe3daH4Ux9YgFRBU1CwvmjVsvJCmhBXe/sZZ/LdtRf/skyorsCJNfgJ98C/2mwvq5Th/Fy+MhZZ5zv2xjTJ1lg/VVk4KiYu55ax3vr93LjRe256ErexEY0MAGwTuRDav/Cyueh5w90CwOBs2ExBnQJMrf1RljvKhosD4LiGpUUqI89tFmnvs8lTG9WvHU1AGEBQf6tSa/KCmGrR85o8bu/AwCQ6HPNTBkFrTu5+/qjDEeLCBq2Etf7uT3CzaSGN+CF25MokWTEH+X5D8HNsHy52DtHCg8AfEXwpBbofsE5zRaY4xfWUD4wYJ1+7j7jTXERTbi1ZsG0y6ysb9L8q+8w7B6thMWObshvI3T/DTwJmfocWOMX1hA+Mm3qVnMei2Z0OBAXrlpUN2/4VB1KCmGbYvhu2chdZnT/NR7Mgyc4YwHZTcvMqZGWUD40dbMo8x4aTlH8gp59kcDubhLjL9Lqj0ObPZofjoOEe2h77XQ51qI6erv6oxpEPx2RzkRGSsiW0Rku4jc52X5DBE5KCJr3MfNHsuKPebP92WdvtS1VTjv/GQY7SIbc9PLK3hnVQO5VqIyWnZ3rqe4Zwv84FnntNkv/grPDIJ/XwJf/8Ou1DbGj3x2BCEigcBW4HIgHVgBTFPVjR7rzACSVPVOL9sfU9Wmld1fbT2CKJWbX8itr63km9Qs7h3TjZ+M6IRYc8qZju6HDW/Dujdh3xqQAOhwCfS9zunYDmvm7wqNqVf8dQQxGNiuqqmqehKYA0zy4f5qtWZhwbwycxAT+7XhL4u28OB7KRSX1I/mvWoVHgsX3gG3fgZ3rICL73FGk513OzzeBd6aAVs+tLveGVMDgnz42m0Bz1uvpQNDvKw3WUQuwTnauFtVS7cJE5FkoAh4TFXnld1QRGYBswDi4+OrsXTfCA0K5Mnr+tO6eRj//jyVzNx8np7WQK+VqIyYrnDpb2DkA5C+wjmqSHkHUt51xoXqdZVzZNFuiHVuG+MDvmximgKMVdWb3ekfAUM8m5NEJAo4pqoFInIrcJ2qXuoua6uqGSLSEfgEGKWqO8rbX21vYirr5a928sgHGxnQLoIXpw9q2NdKnIviQtjxiRMWmxc4NzWKiIeek6DbFc6ZUAEWuMZUll/OYhKRC4GHVXWMO30/gKr+sZz1A4FsVT3jXFAReQX4QFXnlre/uhYQAB+u38fP31hDXEQjXpwxiA7RTfxdUt1ScNQJifVvQepnUFIIjaOg61joNg46XQoh9p4aUxF/BUQQTrPRKCADp5P6elVN8Vintaruc59fBfxaVS8QkRbACffIIhr4Bpjk2cFdVl0MCIDlO7O55bVkCotLeHBCT64b1M46r89Hfi5sX+L0T2xbBPlHnGssOo6A7uOd0AiP9XeVxtQ6frsOQkTGA08CgcBLqvqoiDwCJKvqfBH5IzARp58hG7hdVTeLyFDg30AJTkf6k6r6YkX7qqsBAbDvSB6/fHMtX+/I4rIerfjT5D5ENQ31d1l1V3Eh7PnGCYvNC5wrtwHaJjlHFt2vgJju1m9hDHahXJ1QUqK89NVO/rxoC83CgvnLlL6M7N7S32XVfarOeFBbFjiBkbHSmd8iwemz6DbOGR8q0JfnaxhTe1lA1CGb9+dy15w1bN5/lBuGxPObK3rQOMQ+vKpN7j5npNktHzpDfRQXQFiE0xTV4WLoMByiOtvRhWkwLCDqmIKiYv66eCvPf5FKh6gm/O26/vRrF+HvsuqfgmOQ+unpsMjNcOY3jXXCIuFi52uLDhYYpt6ygKijvt5xiHveXMuBowX8fFQXbh/RiaBAuwmgT6hCdirs+gJ2fuF8PZbpLGve7nRYJFwMEe38W6sx1cgCog47klfIg+9t4L01e0mMj+Bv1/WnfZSduulzqnBoK+z83AmLXV/CiSxnWYsENzCGO6FhZ0eZOswCoh54b00Gv523gZIS5aEre3FNUpydDluTSkrg4CYnMHZ+Abu/dE6lBYjqAvFDoE0itE2Elr0gyC58NHWDBUQ9kZGTxy/fXMO3qdmM7tmKxyb3JdKuwPaPkmLYv/70EUZ6MuRlO8sCQyG29+nAaJMI0V3sCm9TK1lA1CMlJcqLX+7kL4u20LxxMH+e0peR3ex0WL9Tda63yFgFe1dBxmpnNNqTx5zlIU2d+3G3GXA6NFokWOe3qR7Fhed9C18LiHpo0z7ndNgtmUf50QXtuX98dzsdtrYpKYZD22Dvajc0VjlHHcUFzvJGkacDo1VviO4KUZ0gyC6SNF6UFDv/hBzaDlnbnN+tQ9uc59FdYcYH5/WyFhD1VH5hMY8v2sILX+4krkUjHpnUi0u7t/J3WaYiRSfhwMbTgbF3tXMhnxY7yyXQObKI6eY0S0V3O/08zG5Z2yCcyIas7ac//A9tc6azU6HYY5j7Ri2c/q/oLs4/GYNuLv81K1DlgBCRnwMvA0eBF4ABwH2quvi8KvKBhhgQpb5LzeK38zaw7cAxxvRqxUNX9qJNRCN/l2Uq6+QJ54Pg4FY4tMU5e+rgVudDoaTw9HpNY50h0KO7Of8xlj4Pj7WmqrqkpNi5MdaRNMhJc75mp54OhROHTq8bEOTcaTGqC0R3Ph0IUV2gSVS1lFMdAbFWVfuJyBjgVuD/gP+oamK1VFgNGnJAAJwsKuGFL1N5euk2AkS4+7KuzBiWQLBdN1F3FRc5N0s65AbHwa3u861QkHt6vdBmzlFHRLxzzUZEO4+v8dA40gKkJhXmw5F0OLLH+VoaAjlpzrzcvVBS9P1tmsR4D4EW7c+7b6GyqiMg1qlqXxF5Climqu+KyGpVHVDdxZ6vhh4QpdKyT/DQ/BQ+2XyA7rHhPHpVHwa2b+Hvskx1UnX+Ay0Ni4NbnCAp/RAqPP799YObQPO4M4OjdDo81s6wOpuik85ZaieyPB7Z7iMLju51AyAdjh/4/rYSAOFt3Pc77syfQbO2EFrpuytXu+oIiJdx7hDXAeiHMzrrMlUdWJ2FVoUFxGmqyqKUTH73fgr7juQzbXA8vx7bjYjGdkpsvacKeYchZ8/3mzA8p0tPxy0VEOz8B9sowmnXDnO/NopwHt+b9lge1rzuBEvRSSc4Tx53mvROHoPCE6ef5x/x+NDPKhMG2d8/YisrtBk0beklfN0waNbG50cBVVEdAREA9AdSVTVHRCKBOFVdV62VVoEFxJmOFxTx5JKtvPTVLiIaBfPA+B5cndjWLrBr6AqOuU0gHsFx/CDk5TiP/BwnZPJyzjwa+R6BsGZOYIQ0cT4EA0OcR0DQ6eeBQWefjzgd9VritNFriTNdUuyEntdlJaefFxU4H/6F7gf+yRPff+7Zl1ORkKZOk1zjKOcss8ZRHg8v040i6/xFkdUREMOANap6XER+CCQCT6nq7uot9fxZQJRv495cfjNvPav35HBBx0j+8IPedG4Z7u+yTF1QVHBmaOQdPnO68IRzLn7xSad9vfik+/B4Xt58ynwGSYBzNpcEOEcop54HeFnmzgsKdUIquInzNaTxmdOnnruP4MZOIIQ0do4CGkdBcFhNv8N+Vy19EDhNS32BV3DOZLpWVYdXY51VYgFRsZIS5Y3kNB77cDMnThYx65KO3DmyC41C6kgTgam/it0O29IQMDWqooCo7E+jSJ0kmQT8Q1WfAexf0DokIECYNjiepb8czpX92vDMpzsY/eRnfLr5wNk3NsaXAoOch4VDrVPZn8hREbkf+BGwwO2TqL29LqZc0U1DeeLa/rx+ywWEBAZw0ysruP2/K9mbk+fv0owxtUxlA+I6oACYqar7gTjgLz6ryvjchZ2i+PDnl3DvmG58svkAl/51GU98vJUTJ4vOvrExpkGoVEC4oTAbaC4iE4B8VX3tbNuJyFgR2SIi20XkPi/LZ4jIQRFZ4z5u9lg2XUS2uY/p5/A9mUoKCQrgjpGdWfrL4VzeM5anl25j5OPLmLsynZKS+jEEizHm/FW2k/panCOGZYAAFwP3qurcCrYJBLYClwPpwApgmqpu9FhnBpCkqneW2TYSSAaScE5xWAkMVNXD5e3POqmrbuXubB75YBNr03LoG9ec/5vQk0EJkf4uyxjjQ9XRSf0bYJCqTlfVG4HBOMNtVGQwsF1VU1X1JDAHp5O7MsYAH6tqthsKHwNjK7mtOU8D20fy7u1D+dt1/TiQW8A1z37DHbNXkZZ9wt+lGWP8oLIBEaCqnqe7ZFVi27ZAmsd0ujuvrMkisk5E5opI6c1+K7WtiMwSkWQRST548OBZvwlzdgEBwlUD4vj0nhHcfVlXPtl8gFFPfMafPtrM0fxKXmxkjKkXKhsQH4nIIrfPYAawAFhYDft/H0hQ1b44RwmvnsvGqvqcqiapalJMTEw1lGNKNQoJ5OeXdeHTe0YwoW9r/rVsByMf/4w5y/dQbP0TxjQIle2kvhd4DudCub7Ac6r667NslgG085iOc+d5vm6Wqrp3T+EFYGBltzU1I7Z5GE9c25/37hhGQlRj7ntnPRP+/iVf7zh09o2NMXWaz24YJCJBOJ3Uo3A+3FcA16tqisc6rVV1n/v8KuDXqnqB20m9EmdID4BVOJ3UZUYZO806qX1PVVmwfh9/XLiZjJw8Lu/ZigfG96BDdBN/l2aMOU8VdVJXeI9KETnKGQOlOIsAVdVm5W2rqkUiciewCGf015dUNUVEHgGSVXU+8DMRmQgUAdnADHfbbBH5PU6oADxSUTiYmiEiTOjbhst6tOKlr3byzCfbGf23z5h+YQI/vbQLzRvbtZPG1Cd2y1Fz3g4czeeJxVt5IzmN8NAgZl3SkZuGdaBJqN0b25i6wu5JbXxq075c/rp4K0s2ZRLVJISfjOzMDUPiCQu2gQCNqe0sIEyNWLXnMH9dvIWvtmfRunkYPxvVhSkD4+y2p8bUYtVxoZwxZ5UY34LZN1/A/24ZQuvmYdz/znoue+Iz3luTYUN3GFMHWUCYaje0UzRv3z6UF6cn0TgkiJ/PWcO4p75gUcp+6ssRqzENgQWE8QkRYVSPViz46UX8fdoACotLuPU/K/nBM1/xxbaDFhTG1AEWEManAgKEK/u1YfHdl/DnKX05dOwkP3pxOVOf+5bkXXbmsjG1mXVSmxpVUFTMnOVp/P2T7Rw6VsDIbjH8cnQ3erdt7u/SjGmQ7CwmU+ucOFnEq1/v5tnPdnAkr5DRPVtxx8jO9GsX4e/SjGlQLCBMrXUkr5CXvtzJy1/tJDe/iIs6R/OTkZ24sGMUIuLv8oyp9ywgTK13rKCI2d/u5oUvd3LwaAH920Vwx8jOjOrekoAACwpjfMUCwtQZ+YXFzF2ZzrOf7SD9cB7dWoVz+4hOTOjbmiC74M6YamcBYeqcouISPli3j38u287WzGPERzbm1uEdmZwYZ0N4GFONLCBMnVVSoizZlMkzy3awNi2HluGh3HxxB64f0p6mNiigMVVmAWHqPFXlmx1ZPLNsO19tz6J5o2BmDE1gxtAEWjQJ8Xd5xtRZFhCmXlmTlsM/P93O4o2ZNA4J5PrB8dx8cUdim4f5uzRj6hwLCFMvbc08yr+W7WD+2r0IcGW/Nvz4og520Z0x58ACwtRradknePmrXbyxYg/HTxZzQcdIbr6oI5faKbLGnJUFhGkQcvMLeWN5Gi9/tZO9R/LpGN2EmRd1YHJiHI1C7MwnY7zx2/0gRGSsiGwRke0icl8F600WERWRJHc6QUTyRGSN+3jWl3Wa+qFZWDC3XNKRz341kqenDaBpWBC/nbeBoY8t5a+Lt3DgaL6/SzSmTvHZEYSIBAJbgcuBdGAFME1VN5ZZLxxYAIQAd6pqsogkAB+oau/K7s+OIExZqkry7sM8/3kqH2/KJDgggIn923DzxR3oHtvM3+UZUytUdAThyxPJBwPbVTXVLWIOMAnYWGa93wN/Au71YS2mARIRBiVEMighkl2HjvPyVzt5MzmduSvTubhLND++qAPDu8bYmE/GlMOXTUxtgTSP6XR33ikikgi0U9UFXrbvICKrReQzEbnYh3WaBiAhugm/m9Sbb+6/lF+N7cbWzKPMeHkFo//2OXOW7yG/sNjfJRpT6/htcBsRCQCeAH7pZfE+IF5VBwC/AP4nIme0CYjILBFJFpHkgwcP+rZgUy9ENA7hJyM688WvLuWJa/sRHBjAfe+sZ+hjn/DnjzaTkZPn7xKNqTV82QdxIfCwqo5xp+8HUNU/utPNgR3AMXeTWCAbmKiqyWVeaxlwT9n5nqwPwpwPVeWb1Cxe/XoXH2/MBGB0z1imD03ggo6R1vxk6j1/9UGsALqISAcgA5gKXF+6UFWPANEeRS7DDQERiQGyVbVYRDoCXYBUH9ZqGigRYWinaIZ2iib98An+++0e5qzYw0cp++keG86NFybwgwFtaBxi4z6ZhsdnTUyqWgTcCSwCNgFvqmqKiDwiIhPPsvklwDoRWQPMBW5TVbuBsfGpuBaNuW9cd769fxR/ntyXABEeeHc9F/y/pTy6YCN7sk74u0RjapRdKGdMOUpPk33l6118tGE/Japc2q0l04cmcHGXaGt+MvWCv5qYjKnTPE+T3X8kn/99t5v/Ld/DjS8tp2NME6ZfmMDkgXE27Lipt+wIwphzUFBUzML1+3jl692sTcuhaWgQUwbG8cML2tO5ZVN/l2fMObOxmIzxgTVpObz69S4+WLeXwmLlgo6R3DCkPWN6xRISZLdHNXWDBYQxPnToWAFvJafzv+W7ScvOI7ppCNcmtWPa4HjaRTb2d3nGVMgCwpgaUFKifLH9ELO/3c2STZkoMLxrDDcMac/IbjEEBdpRhal9LCCMqWH7juQxZ3kac1bsITO3gNbNw5g6KJ7rBrWzO9+ZWsUCwhg/KSouYenmA8z+bg+fbz1IYIBwWY+W3DCkPRd1jrYbGhm/s9NcjfGToMAAxvSKZUyvWHZnHef15Wm8lZzGopRM4iMbc/2QeK4ZGEdU01B/l2rMGewIwpgaVlBUzKKUTGZ/u5vvdmYTEhjAmN6xTB3Ujgs7RtlRhalR1sRkTC21/cBRZn+3h3dWZXAkr5B2kY24LqkdUwZaX4WpGRYQxtRy+YXFLErZzxsr0vh6RxYBAiO7teS6Qe0Y2b0lwXYGlPER64MwppYLCw5kUv+2TOrflt1Zx3kzOY23ktNZuvkAMeGhTBkYx3VJ7UiIbuLvUk0DYkcQxtRSRcUlfLrlIG+s2MMnmw9QonBhxyimDm7HmF6xhAUH+rtEUw9YE5Mxddz+I/m8vSqdN1aksSf7BM0bBXPVgLZcN6gdPVqfcbNFYyrNAsKYeqKkRPk2NYs5K9L4aMN+ThaX0C+uOdcktePKfm1o3ijY3yWaOsYCwph66PDxk7y7OoM3VqSxJfMoIUHONRfXDIxjWOdoAu10WVMJFhDG1GOqyvqMI7yVnM57azLIzS+idfMwrk5sy5SB7ehgHdumAhYQxjQQ+YXFLNmUydyV6Xy+9SAlCoMSWjBlYBxX9G1jNzcyZ7CAMKYB2n8kn3dWpzN3ZTqpB4/TKDiQcX1imTIwjgs62BXbxuG3gBCRscBTQCDwgqo+Vs56k4G5wCBVTXbn3Q/8GCgGfqaqiyralwWEMd6pKqv25DB3ZTofrN3L0YIi4lo0YnJiHFMGxtk9Kxo4vwSEiAQCW4HLgXRgBTBNVTeWWS8cWACEAHeqarKI9AReBwYDbYAlQFdVLS5vfxYQxpxd3sliFm/cz1vJ6Xy14xDqXlsxeWAcY3vHWhNUA1RRQPjy+v3BwHZVTVXVk8AcYJKX9X4P/AnI95g3CZijqgWquhPY7r6eMaYKGoU4V2z/9+YhfPnrS/nl5V3ZeySPe95ay6A/LOGuOav5fOtBikvqR9OzqRpf/rvQFkjzmE4HhniuICKJQDtVXSAi95bZ9tsy27YtuwMRmQXMAoiPj6+mso1pGNpGNOKno7pw56WdWbn7MO+szuCDtXuZt2YvLcND+cGAtlyd2JbusXYhXkPlt+NJEQkAngBmnO9rqOpzwHPgNDFVT2XGNCwiQlJCJEkJkTw4oSefbj7A26syeOnLnTz3eSo9WjdjcmJbJvZvQ8twG2G2IfFlQGQA7Tym49x5pcKB3sAyEQGIBeaLyMRKbGuM8YGw4EDG9WnNuD6tyTpWwAfr9vHO6gz+sGAT/2/hJi7uEsPViW0Z3TOWRiE2FlR958tO6iCcTupROB/uK4DrVTWlnPWXAfe4ndS9gP9xupN6KdDFOqmN8Y/tB44xb3UG767OICMnj6ahQYztHcvViW3tlNk6zi/DfatqkYjcCSzCOc31JVVNEZFHgGRVnV/Btiki8iawESgC7qgoHIwxvtW5ZVPuGdONX1zeleW7snlnVToL1+9n7sp02jQPY9KAtvygf1u6xYb7u1RTjexCOWPMeck7WczHmzJ5Z1U6X2w7RHGJ0j02nEn9nf6KthGN/F2iqQS7ktoY41OHjhWwcP0+5q3OYNWeHAAGJ0QysX8brujTmhZNQvxboCmXBYQxpsbsyTrB/LUZzFuzl+0HjhEcKAzvGsPE/m25vEcr69yuZSwgjDE1TlXZuC+X99bsZf6avezPzadxSCBjesUyqX8bLuocTZDda9vvLCCMMX5VUqJ8tzOb+WszWLBuH7n5RUQ1CWFC39ZM7N+WxPgI3NPdTQ2zgDDG1BoFRcV8tuUg763Zy5JNmRQUldAushET+rbhyr5t6NE63MKiBllAGGNqpaP5hSxKyeT9tXv5crtzJlSnmCZc2a8NE/q2oXPLpv4usd6zgDDG1HrZx0/y4YZ9vL92L9/tzEYVerRuxpX9WnNl3zY2LLmPWEAYY+qUzNx8Fq53wqL0tNn+7SK4sp9z2mxscxsTqrpYQBhj6qy07BMscMMiZW8uIjAoIZIr+7VhXO9YopuG+rvEOs0CwhhTL+w4eIwP1u7j/XXONRaBAcLQTlFc2bcNo3u1IqKxXZB3riwgjDH1iqqyJfMo76/dy/tr97En+wRBAcKwztFc0ae1hcU5sIAwxtRbqsqGjFw+WL+Xhev3kZadR1CAcFGXaMb3ac2YnrE0bxzs7zJrLQsIY0yDoKqszzjCgnX7WLB+H+mH8wgO9DiysLA4gwWEMabBUVXWpR9h4fp9fLBuHxk5Tlhc1DmaK/q24fKerWjeyMLCAsIY06CpKmvdsFjgERYXd4lhfJ/WDTosLCCMMcalqqxJyzkVFnuP5J9qhhrXO5bLe8YS2YCGJ7eAMMYYL1SV1Wk5fLRhPwvdPovAAOHCjlGM7R3LmF6xxITX7+ssLCCMMeYsVJWUvbksXL+PDzfsZ+eh46cuyhvfO5axvevnFdwWEMYYcw5Kr7P4cP1+Ptywj62ZxwBIjI9gXO/WjO0dW2/GhvJbQIjIWOApIBB4QVUfK7P8NuAOoBg4BsxS1Y0ikgBsAra4q36rqrdVtC8LCGOMr+w4eOxUM1TK3lwA+rRtzrg+sYzr3ZoO0U38XOH580tAiEggsBW4HEgHVgDTVHWjxzrNVDXXfT4R+ImqjnUD4gNV7V3Z/VlAGGNqwp6sE3y4wWmGWpOWA0D32HDG9HL6LOra/SwqCoggH+53MLBdVVPdIuYAk4BTAVEaDq4mQP1o7zLG1FvxUY25dXgnbh3eib05eXy0YT8fpezn6U+28dTSbcRHNnY7uFsxoF0LAgLqTliU5cuAaAukeUynA0PKriQidwC/AEKASz0WdRCR1UAu8FtV/cLLtrOAWQDx8fHVV7kxxlRCm4hGzLyoAzMv6sDBowUs2ZTJopT9vPzVTp77PJWW4aFc3rMVY3vHckHHKILr2D24fdnENAUYq6o3u9M/Aoao6p3lrH89MEZVp4tIKNBUVbNEZCAwD+hV5ojje6yJyRhTW+TmF/Lp5gMsStnPp5sPkldYTLOwIC7r0YoxvWO5pEsMjUIC/V0m4L8mpgygncd0nDuvPHOAfwGoagFQ4D5fKSI7gK6AJYAxptZrFhbMpP5tmdS/LfmFxXy+9SCLUjJZsimTd1Zn0Cg4kOFdYxjbO5aR3VvW2qu4fRkQK4AuItIBJximAtd7riAiXVR1mzt5BbDNnR8DZKtqsYh0BLoAqT6s1RhjfCIsOJDRvWIZ3SuWwuISvkvNZlHKfhalOH0XQQHChZ2iGN0rlst7tKpV11r4+jTX8cCTOKe5vqSqj4rII0Cyqs4XkaeAy4BC4DBwp6qmiMhk4BF3fgnwkKq+X9G+rInJGFOXlJQ4V3GXhsXurBMA9GsXweierRjTqxWdYpr6/Iwou1DOGGNqMVVl24FjLE7Zz8cbM1mbfgSAjtFNuLxXK0b3jGVAuwifnBFlAWGMMXXIviN5LNmYyeKNmXyzI4uiEiUmPJTLerRidK9WDO0URWhQ9XRyW0AYY0wddSSvkGVbDrA4JZNlWw5w/GQxTUICGdG9JaN7tmJk95Y0Czv/Tm4LCGOMqQfyC4v5JjWLxSmZfLwxk0PHCggOFMb0iuUf1yee12v66zRXY4wx1SgsOJCR3VoysltLHv1Bb1an5bB4o3MmlC9YQBhjTB0UECAMbN+Cge1b+G4fPntlY4wxdZoFhDHGGK8sIIwxxnhlAWGMMcYrCwhjjDFeWUAYY4zxygLCGGOMVxYQxhhjvKo3Q22IyEFgdxVeIho4VE3l+ILVVzVWX9VYfVVTm+trr6ox3hbUm4CoKhFJLm88ktrA6qsaq69qrL6qqe31lceamIwxxnhlAWGMMcYrC4jTnvN3AWdh9VWN1Vc1Vl/V1Pb6vLI+CGOMMV7ZEYQxxhivLCCMMcZ41aACQkTGisgWEdkuIvd5WR4qIm+4y78TkYQarK2diHwqIhtFJEVEfu5lnREickRE1riPB2uqPo8adonIenf/Z9zjVRxPu+/hOhE5v/sgnl9t3TzemzUikisid5VZp0bfQxF5SUQOiMgGj3mRIvKxiGxzv3q944uITHfX2SYi02uwvr+IyGb35/euiESUs22Fvws+rO9hEcnw+BmOL2fbCv/efVjfGx617RKRNeVs6/P3r8pUtUE8gEBgB9ARCAHWAj3LrPMT4Fn3+VTgjRqsrzWQ6D4PB7Z6qW8E8IGf38ddQHQFy8cDHwICXAB858ef936ci4D89h4ClwCJwAaPeX8G7nOf3wf8yct2kUCq+7WF+7xFDdU3Gghyn//JW32V+V3wYX0PA/dU4udf4d+7r+ors/yvwIP+ev+q+mhIRxCDge2qmqqqJ4E5wKQy60wCXnWfzwVGiYhvbvZahqruU9VV7vOjwCagbU3su5pNAl5Tx7dAhIi09kMdo4AdqlqVq+urTFU/B7LLzPb8PXsV+IGXTccAH6tqtqoeBj4GxtZEfaq6WFWL3Mlvgbjq3m9llfP+VUZl/t6rrKL63M+Oa4HXq3u/NaUhBURbIM1jOp0zP4BPreP+gRwBomqkOg9u09YA4Dsviy8UkbUi8qGI9KrZygBQYLGIrBSRWV6WV+Z9rglTKf8P09/vYStV3ec+3w+08rJObXkfZ+IcEXpztt8FX7rTbQJ7qZwmutrw/l0MZKrqtnKW+/P9q5SGFBB1gog0Bd4G7lLV3DKLV+E0mfQD/g7Mq+HyAC5S1URgHHCHiFzihxoqJCIhwETgLS+La8N7eIo6bQ218lxzEfkNUATMLmcVf/0u/AvoBPQH9uE049RG06j46KHW/y01pIDIANp5TMe587yuIyJBQHMgq0aqc/YZjBMOs1X1nbLLVTVXVY+5zxcCwSISXVP1ufvNcL8eAN7FOZT3VJn32dfGAatUNbPsgtrwHgKZpc1u7tcDXtbx6/soIjOACcANboidoRK/Cz6hqpmqWqyqJcDz5ezX3+9fEHA18EZ56/jr/TsXDSkgVgBdRKSD+x/mVGB+mXXmA6Vni0wBPinvj6O6ue2VLwKbVPWJctaJLe0TEZHBOD+/mgywJiISXvocpzNzQ5nV5gM3umczXQAc8WhOqSnl/ufm7/fQ5fl7Nh14z8s6i4DRItLCbUIZ7c7zOREZC/wKmKiqJ8pZpzK/C76qz7NP66py9luZv3dfugzYrKrp3hb68/07J/7uJa/JB84ZNltxzm74jTvvEZw/BIAwnGaJ7cByoGMN1nYRTlPDOmCN+xgP3Abc5q5zJ5CCc0bGt8DQGn7/Orr7XuvWUfoeetYowDPue7weSKrhGpvgfOA395jnt/cQJ6j2AYU47eA/xunXWgpsA5YAke66ScALHtvOdH8XtwM31WB923Ha70t/D0vP7GsDLKzod6GG6vuP+7u1DudDv3XZ+tzpM/7ea6I+d/4rpb9zHuvW+PtX1YcNtWGMMcarhtTEZIwx5hxYQBhjjPHKAsIYY4xXFhDGGGO8soAwxhjjlQWEMedARIrl+yPGVtsooSKS4DkqqDH+FuTvAoypY/JUtb+/izCmJtgRhDHVwB3b/8/u+P7LRaSzOz9BRD5xB5ZbKiLx7vxW7r0W1rqPoe5LBYrI8+LcE2SxiDTy2zdlGjwLCGPOTaMyTUzXeSw7oqp9gH8AT7rz/g68qqp9cQa9e9qd/zTwmTqDBibiXE0L0AV4RlV7ATnAZJ9+N8ZUwK6kNuYciMgxVW3qZf4u4FJVTXUHXdyvqlEicghnKIhCd/4+VY0WkYNAnKoWeLxGAs49ILq4078GglX1DzXwrRlzBjuCMKb6aDnPz0WBx/NirJ/Q+JEFhDHV5zqPr9+4z7/GGUkU4AbgC/f5UuB2ABEJFJHmNVWkMZVl/50Yc24albkJ/UeqWnqqawsRWYdzFDDNnfdT4GURuRc4CNzkzv858JyI/BjnSOF2nFFBjak1rA/CmGrg9kEkqeohf9diTHWxJiZjjDFe2RGEMcYYr+wIwhhjjFcWEMYYY7yygDDGGOOVBYQxxhivLCCMMcZ49f8BvfSIO2ZrUBQAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.title('Model loss')\n",
    "plt.ylabel('loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Validation'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "7f5b51ca",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['acc'])\n",
    "plt.plot(history.history['val_acc'])\n",
    "plt.title('Model acc')\n",
    "plt.ylabel('acc')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Validation'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "0e49af67",
   "metadata": {},
   "outputs": [],
   "source": [
    "pred = lstm2.predict(X_test)\n",
    "pred[pred>0.5]=1\n",
    "pred[pred<=0.5]=0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "895d2f0a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "정확도 : 0.7079948696023942\n"
     ]
    }
   ],
   "source": [
    "from sklearn import metrics \n",
    "print('정확도 :', metrics.accuracy_score(y_test, pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ae37a0f8",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6b905188",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "85bced58",
   "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": "3.8.5"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {
    "height": "calc(100% - 180px)",
    "left": "10px",
    "top": "150px",
    "width": "164.97px"
   },
   "toc_section_display": true,
   "toc_window_display": true
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}