{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n",
      "/Users/yangyoonji/anaconda3/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:526: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint8 = np.dtype([(\"qint8\", np.int8, 1)])\n",
      "/Users/yangyoonji/anaconda3/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:527: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint8 = np.dtype([(\"quint8\", np.uint8, 1)])\n",
      "/Users/yangyoonji/anaconda3/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:528: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint16 = np.dtype([(\"qint16\", np.int16, 1)])\n",
      "/Users/yangyoonji/anaconda3/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:529: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint16 = np.dtype([(\"quint16\", np.uint16, 1)])\n",
      "/Users/yangyoonji/anaconda3/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:530: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint32 = np.dtype([(\"qint32\", np.int32, 1)])\n",
      "/Users/yangyoonji/anaconda3/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:535: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  np_resource = np.dtype([(\"resource\", np.ubyte, 1)])\n"
     ]
    }
   ],
   "source": [
    "from keras.models import Model, Sequential\n",
    "from keras.layers import Input, Dense, LSTM, Bidirectional\n",
    "from keras.callbacks import EarlyStopping, ModelCheckpoint\n",
    "from keras import backend as K\n",
    "import matplotlib.pyplot as plt\n",
    "from keras.layers.core import Dense, Activation, Dropout\n",
    "import pandas as pd\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "train = pd.read_csv('/Users/yangyoonji/Documents/2020_2학기/캡스톤디자인/data/scaled/train_scaled.csv')\n",
    "test = pd.read_csv('/Users/yangyoonji/Documents/2020_2학기/캡스톤디자인/data/recent_test/recent_data_scaled.csv')\n",
    "validation = pd.read_csv('/Users/yangyoonji/Documents/2020_2학기/캡스톤디자인/data/scaled/validation_scaled.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 불필요한 열 삭제\n",
    "del train['Unnamed: 0']\n",
    "del test['Unnamed: 0']\n",
    "del validation['Unnamed: 0']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "test.columns = ['close','open','high','low','vol']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "checking if any null values are present in train\n",
      " close    0\n",
      "open     0\n",
      "high     0\n",
      "low      0\n",
      "vol      0\n",
      "dtype: int64\n",
      "checking if any null values are present in test\n",
      " close    0\n",
      "open     0\n",
      "high     0\n",
      "low      0\n",
      "vol      0\n",
      "dtype: int64\n",
      "checking if any null values are present in validation\n",
      " close    0\n",
      "open     0\n",
      "high     0\n",
      "low      0\n",
      "vol      0\n",
      "dtype: int64\n"
     ]
    }
   ],
   "source": [
    "# NULL value 없나 확인해보기\n",
    "print(\"checking if any null values are present in train\\n\", train.isna().sum())\n",
    "print(\"checking if any null values are present in test\\n\", test.isna().sum())\n",
    "print(\"checking if any null values are present in validation\\n\", validation.isna().sum())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_dataset(data, label,window_size = 20):\n",
    "    feature_list = []\n",
    "    label_list = []\n",
    "    for i in range(len(data)-window_size):\n",
    "        feature_list.append(np.array(data.iloc[i:i+window_size]))\n",
    "        label_list.append(np.array(label.iloc[i+window_size]))\n",
    "    return np.array(feature_list), np.array(label_list)    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "# feature , label 분리\n",
    "feature_cols = ['open','high','low','vol']\n",
    "label_cols = ['close']\n",
    "\n",
    "##----train_data----##\n",
    "train_feature = train[feature_cols]\n",
    "train_label = train[label_cols]\n",
    "\n",
    "train_feature, train_label = make_dataset(train_feature,train_label,3)\n",
    "\n",
    "##----test_data----##\n",
    "test_feature = test[feature_cols]\n",
    "test_label = test[label_cols]\n",
    "test_feature, test_label = make_dataset(test_feature,test_label,3)\n",
    "\n",
    "##----validation_data----##\n",
    "validation_feature = validation[feature_cols]\n",
    "validation_label = validation[label_cols]\n",
    "validation_feature, validation_label = make_dataset(validation_feature,validation_label,3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "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>close</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>3458</th>\n",
       "      <td>0.672652</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3459</th>\n",
       "      <td>0.673757</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3460</th>\n",
       "      <td>0.671060</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3461</th>\n",
       "      <td>0.655714</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3462</th>\n",
       "      <td>0.653901</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3463</th>\n",
       "      <td>0.660136</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         close\n",
       "3458  0.672652\n",
       "3459  0.673757\n",
       "3460  0.671060\n",
       "3461  0.655714\n",
       "3462  0.653901\n",
       "3463  0.660136"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train[label_cols].tail(6)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_label_pd = pd.DataFrame(train_label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "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>0</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>3455</th>\n",
       "      <td>0.672652</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3456</th>\n",
       "      <td>0.673757</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3457</th>\n",
       "      <td>0.671060</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3458</th>\n",
       "      <td>0.655714</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3459</th>\n",
       "      <td>0.653901</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3460</th>\n",
       "      <td>0.660136</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             0\n",
       "3455  0.672652\n",
       "3456  0.673757\n",
       "3457  0.671060\n",
       "3458  0.655714\n",
       "3459  0.653901\n",
       "3460  0.660136"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_label_pd.tail(6)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3461"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(train_label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, y_train, X_val, y_val = train_feature, train_label, validation_feature, validation_label"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test, y_test = test_feature, test_label"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "len(y_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "lstm_7 (LSTM)                (None, 20, 64)            17664     \n",
      "_________________________________________________________________\n",
      "lstm_8 (LSTM)                (None, 50)                23000     \n",
      "_________________________________________________________________\n",
      "dense_4 (Dense)              (None, 1)                 51        \n",
      "=================================================================\n",
      "Total params: 40,715\n",
      "Trainable params: 40,715\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "#keras 활용한 LSTM 모델 생성\n",
    "\n",
    "model = Sequential() \n",
    "model.add(LSTM(64, return_sequences=True, input_shape=(20, 4))) #첫번째 LSTM 은 유닛수 50개\n",
    "model.add(LSTM(50, return_sequences=False)) #두번째 LSTM 유닛수는 64개\n",
    "model.add(Dense(1, activation='linear')) # 아웃풋으로 나오는 값은 1개 (다음날 하루 예측)\n",
    "model.compile(loss='mse', optimizer='adam') #손실 함수 ,optimizer= rmsprop\n",
    "model.summary() #모델의 개요"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /Users/yangyoonji/anaconda3/lib/python3.7/site-packages/tensorflow/python/ops/math_ops.py:3066: to_int32 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.cast instead.\n",
      "Train on 3444 samples, validate on 167 samples\n",
      "Epoch 1/100\n",
      "3444/3444 [==============================] - 53s 15ms/step - loss: 0.0041 - val_loss: 0.0097\n",
      "Epoch 2/100\n",
      "3444/3444 [==============================] - 48s 14ms/step - loss: 7.3472e-04 - val_loss: 0.0073\n",
      "Epoch 3/100\n",
      "2510/3444 [====================>.........] - ETA: 13s - loss: 5.2751e-04"
     ]
    }
   ],
   "source": [
    "early_stopping = EarlyStopping() \n",
    "hist = model.fit(X_train, y_train, validation_data=(X_val, y_val),batch_size=5, epochs=100,callbacks=[early_stopping])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "score = model.evaluate(X_test, y_test, batch_size = 5)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}