{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## LSTM 모델\n",
    "\n",
    "일반신경망에 시계열 개념이 추가되어 은닉계층에 이전 정보를 기억시킬 수 있는 순환신경망이 시계열 데이터인 주가 예측 모델로 적합<br>\n",
    "1. 단방향 LSTM 순환신경망\n",
    "    - 입력 순서를 시간 순서대로 처리하기 때문에 결과물은 주로 직전 패턴을 기반으로 하는 경향 (한계점)\n",
    "2. 양방향 LSTM 순환신경망\n",
    "    - 데이터의 흐름에 역방향에 은닉계층이 추가되는 양방향 LSTM 순환신경망\n",
    "3. 성능 측정 : 실제 주가와 예측된 주가 간의 평균 제곱근 오차(RMSE)\n",
    "---\n",
    "\n",
    "**2주차 목표 : 단방향 LSTM 과 양방향 LSTM의 성능(RMSE)과 예측률 비교**\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 단방향(Unidirectional) LSTM\n",
    "----"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## (1) 데이터 불러오기\n",
    "### case1. Simple Price Data (시가, 종가, 고가, 저가, 거래량)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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/scaled/test_scaled.csv')\n",
    "validation = pd.read_csv('/Users/yangyoonji/Documents/2020_2학기/캡스톤디자인/data/scaled/validation_scaled.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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": 6,
   "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": 7,
   "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",
       "      <th>open</th>\n",
       "      <th>high</th>\n",
       "      <th>low</th>\n",
       "      <th>vol</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.011719</td>\n",
       "      <td>0.011665</td>\n",
       "      <td>0.011228</td>\n",
       "      <td>0.012979</td>\n",
       "      <td>0.290444</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      close      open      high       low       vol\n",
       "0  0.011719  0.011665  0.011228  0.012979  0.290444"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.head(1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "#### Keras RNN 계열의 모델을 트레이닝할 때 요구되는 데이터의 형식 : 3차원\n",
    "3차원 데이터 : **(size, timestep, feature)** <br>\n",
    "일반적인 MLP모델에서는 size, feature만 있는 2차원 모델 <br>\n",
    "RNN계열은 '시간'이라는 개념이 있기 때문에 한 차원 늘어난다. -> **timestep**\n",
    "\n",
    "---\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "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": 69,
   "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",
    "train_feature, train_label = make_dataset(train_feature,train_label,20)\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,20)\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,20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, y_train, X_val, y_val = train_feature, train_label, validation_feature, validation_label"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test, y_test = test_feature, test_label"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(167, 20, 4)\n"
     ]
    }
   ],
   "source": [
    "print(X_val.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(167, 1)\n"
     ]
    }
   ],
   "source": [
    "print(y_val.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## (2) 모델 만들기"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "lstm_13 (LSTM)               (None, 20, 64)            17664     \n",
      "_________________________________________________________________\n",
      "lstm_14 (LSTM)               (None, 50)                23000     \n",
      "_________________________________________________________________\n",
      "dense_7 (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": [
      "Train on 3444 samples, validate on 167 samples\n",
      "Epoch 1/20\n",
      "3444/3444 [==============================] - 30s 9ms/step - loss: 0.0061 - val_loss: 0.0096\n",
      "Epoch 2/20\n",
      "3444/3444 [==============================] - 27s 8ms/step - loss: 7.1951e-04 - val_loss: 0.0086\n",
      "Epoch 3/20\n",
      "3444/3444 [==============================] - 25s 7ms/step - loss: 6.2036e-04 - val_loss: 0.0069\n",
      "Epoch 4/20\n",
      "3444/3444 [==============================] - 26s 8ms/step - loss: 4.9895e-04 - val_loss: 0.0053\n",
      "Epoch 5/20\n",
      "3444/3444 [==============================] - 25s 7ms/step - loss: 4.0447e-04 - val_loss: 0.0050\n",
      "Epoch 6/20\n",
      "3444/3444 [==============================] - 26s 8ms/step - loss: 3.6695e-04 - val_loss: 0.0040\n",
      "Epoch 7/20\n",
      "3444/3444 [==============================] - 28s 8ms/step - loss: 3.1896e-04 - val_loss: 0.0033\n",
      "Epoch 8/20\n",
      "3444/3444 [==============================] - 27s 8ms/step - loss: 2.9615e-04 - val_loss: 0.0030\n",
      "Epoch 9/20\n",
      "3444/3444 [==============================] - 24s 7ms/step - loss: 2.6400e-04 - val_loss: 0.0030\n",
      "Epoch 10/20\n",
      "3444/3444 [==============================] - 24s 7ms/step - loss: 2.3818e-04 - val_loss: 0.0026\n",
      "Epoch 11/20\n",
      "3444/3444 [==============================] - 24s 7ms/step - loss: 2.3996e-04 - val_loss: 0.0026\n",
      "Epoch 12/20\n",
      "3444/3444 [==============================] - 31s 9ms/step - loss: 2.1847e-04 - val_loss: 0.0023\n",
      "Epoch 13/20\n",
      "3444/3444 [==============================] - 27s 8ms/step - loss: 2.1816e-04 - val_loss: 0.0024\n",
      "Epoch 14/20\n",
      "3444/3444 [==============================] - 28s 8ms/step - loss: 2.4145e-04 - val_loss: 0.0024\n",
      "Epoch 15/20\n",
      "3444/3444 [==============================] - 25s 7ms/step - loss: 2.1806e-04 - val_loss: 0.0028\n",
      "Epoch 16/20\n",
      "3444/3444 [==============================] - 25s 7ms/step - loss: 2.0134e-04 - val_loss: 0.0022\n",
      "Epoch 17/20\n",
      "3444/3444 [==============================] - 28s 8ms/step - loss: 2.1454e-04 - val_loss: 0.0022\n",
      "Epoch 18/20\n",
      "3444/3444 [==============================] - 29s 8ms/step - loss: 2.2625e-04 - val_loss: 0.0025\n",
      "Epoch 19/20\n",
      "3444/3444 [==============================] - 29s 8ms/step - loss: 2.1269e-04 - val_loss: 0.0021\n",
      "Epoch 20/20\n",
      "2330/3444 [===================>..........] - ETA: 8s - loss: 2.0376e-04"
     ]
    }
   ],
   "source": [
    "model.fit(X_train, y_train,\n",
    "    validation_data=(X_val, y_val),\n",
    "    batch_size=10,\n",
    "    epochs=3) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "226/226 [==============================] - 0s 1ms/step\n"
     ]
    }
   ],
   "source": [
    "score = model.evaluate(X_test, y_test, batch_size = 32)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "print(score)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [],
   "source": [
    "predict = model.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x648 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12, 9))\n",
    "plt.plot(y_test, label='actual')\n",
    "plt.plot(predict, label='prediction')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "196/196 [==============================] - 0s 2ms/step\n",
      "정확도 : 0.0102\n"
     ]
    }
   ],
   "source": [
    "print(\"정확도 : %.4f\" %(model.evaluate(X_test,y_test)[1]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "## GridSearch\n",
    "하이퍼파라미터를 최적화하면 모델 성능을 향상시키는데 큰 도움이 된다.<br>\n",
    "GridSearch는 리스트로 지정된 여러 하이퍼파라미터 값을 받아 모든 조합에 대해 모델 성능을 평가하여<br>\n",
    "최적의 하이퍼파라미터 조합을 찾는다.\n",
    "\n",
    "---"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import GridSearchCV\n",
    "grid_param_LSTM = {\n",
    "    'batch_size' : [10,20,30,40,50,75,100,150],\n",
    "    'epochs' : [5,10,50,100],\n",
    "    'learning_rate' : [0.0001,0.001,0.01,0.1],\n",
    "    'optimizer' : ['Adam','RMSProp'],\n",
    "    'activation' : ['relu','linear','tanh','sigmoid'],\n",
    "    'dropout_rate' : [0.05,0.1,0.2,0.3,0.4,0.5,0.6],\n",
    "}\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}