djia_support_nikkei.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/sprkzoff/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:516: 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",
      "/home/sprkzoff/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:517: 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",
      "/home/sprkzoff/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:518: 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",
      "/home/sprkzoff/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:519: 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",
      "/home/sprkzoff/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:520: 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",
      "/home/sprkzoff/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:525: 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",
      "/home/sprkzoff/.local/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:541: 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",
      "/home/sprkzoff/.local/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:542: 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",
      "/home/sprkzoff/.local/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:543: 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",
      "/home/sprkzoff/.local/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:544: 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",
      "/home/sprkzoff/.local/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:545: 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",
      "/home/sprkzoff/.local/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:550: 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": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import sklearn.preprocessing\n",
    "import sklearn\n",
    "import os\n",
    "from datetime import datetime, timedelta\n",
    "from tqdm import tqdm\n",
    "\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.models import Sequential, Model\n",
    "from tensorflow.keras.optimizers import Adam\n",
    "from tensorflow.keras.layers import LSTM, GRU, Bidirectional, Dense, Dropout, Input, TimeDistributed, LeakyReLU, BatchNormalization\n",
    "from tensorflow.keras import backend as K\n",
    "from tensorflow.keras.callbacks import ModelCheckpoint, ReduceLROnPlateau,TensorBoard"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['DJI_ema.csv',\n",
       " 'DJI_STO_20period_sma10.csv',\n",
       " 'DJI_modified_sma.csv',\n",
       " '.DS_Store',\n",
       " 'DJI_modified_features - DJI.csv',\n",
       " 'DJI_fe.csv',\n",
       " 'N225withoutNULL.csv',\n",
       " 'dowjone1985_2019.csv',\n",
       " 'N225fillVOL.csv',\n",
       " 'test_djia.csv',\n",
       " 'DJI_denoise.csv',\n",
       " 'N225.csv',\n",
       " 'stocknews',\n",
       " 'preprocessed',\n",
       " 'downjone5y.csv']"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "os.listdir('./data/')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_nikkei = pd.read_csv('./data/N225withoutNULL.csv')\n",
    "df_djia = pd.read_csv('./data/DJI_modified_sma.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>Open</th>\n",
       "      <th>High</th>\n",
       "      <th>Low</th>\n",
       "      <th>Close</th>\n",
       "      <th>Adj Close</th>\n",
       "      <th>Volume</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1985-01-29</td>\n",
       "      <td>11843.070313</td>\n",
       "      <td>11843.070313</td>\n",
       "      <td>11843.070313</td>\n",
       "      <td>11843.070313</td>\n",
       "      <td>11843.070313</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1985-01-30</td>\n",
       "      <td>11960.629883</td>\n",
       "      <td>11960.629883</td>\n",
       "      <td>11960.629883</td>\n",
       "      <td>11960.629883</td>\n",
       "      <td>11960.629883</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1985-01-31</td>\n",
       "      <td>11992.309570</td>\n",
       "      <td>11992.309570</td>\n",
       "      <td>11992.309570</td>\n",
       "      <td>11992.309570</td>\n",
       "      <td>11992.309570</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1985-02-01</td>\n",
       "      <td>11946.959961</td>\n",
       "      <td>11946.959961</td>\n",
       "      <td>11946.959961</td>\n",
       "      <td>11946.959961</td>\n",
       "      <td>11946.959961</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1985-02-04</td>\n",
       "      <td>11891.080078</td>\n",
       "      <td>11891.080078</td>\n",
       "      <td>11891.080078</td>\n",
       "      <td>11891.080078</td>\n",
       "      <td>11891.080078</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         Date          Open          High           Low         Close  \\\n",
       "0  1985-01-29  11843.070313  11843.070313  11843.070313  11843.070313   \n",
       "1  1985-01-30  11960.629883  11960.629883  11960.629883  11960.629883   \n",
       "2  1985-01-31  11992.309570  11992.309570  11992.309570  11992.309570   \n",
       "3  1985-02-01  11946.959961  11946.959961  11946.959961  11946.959961   \n",
       "4  1985-02-04  11891.080078  11891.080078  11891.080078  11891.080078   \n",
       "\n",
       "      Adj Close  Volume  \n",
       "0  11843.070313       0  \n",
       "1  11960.629883       0  \n",
       "2  11992.309570       0  \n",
       "3  11946.959961       0  \n",
       "4  11891.080078       0  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_nikkei.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'1985-01-29'"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_djia['Date'][0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Preprocess"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|          | 0/8762 [00:00<?, ?it/s]/usr/lib/python3/dist-packages/ipykernel_launcher.py:7: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  import sys\n",
      "100%|██████████| 8762/8762 [00:02<00:00, 3886.43it/s]\n"
     ]
    }
   ],
   "source": [
    "# postpone 1 day and concatenate both table\n",
    "\n",
    "# decrement datetime of djia first\n",
    "df_djia_dc = df_djia.copy()\n",
    "# df_djia_dc['Date'] = df_djia['Date'] + timedelta(days=-1)\n",
    "for i in tqdm(range(len(df_djia))):\n",
    "    df_djia_dc['Date'][i] = str((datetime.strptime(df_djia.iloc[i]['Date'], '%Y-%m-%d') + timedelta(days=-1)).date())\n",
    "\n",
    "# df_djia_dc.head()\n",
    "df = pd.merge(df_nikkei, df_djia_dc, on='Date', how='inner')\n",
    "df = df.rename({'Open_x': 'Open(Nikkei)', 'High_x': 'High(Nikkei)', 'Low_x': 'Low(Nikkei)', 'Close_x': 'Close(Nikkei)',\n",
    "                'Open_y': 'Open(DJIA)', 'High_y': 'High(DJIA)', 'Low_y': 'Low(DJIA)', 'Close_y': 'Close(DJIA)'\n",
    "               }, axis=1)\n"
   ]
  },
  {
   "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>Date</th>\n",
       "      <th>Open(Nikkei)</th>\n",
       "      <th>High(Nikkei)</th>\n",
       "      <th>Low(Nikkei)</th>\n",
       "      <th>Close(Nikkei)</th>\n",
       "      <th>Adj Close_x</th>\n",
       "      <th>Volume_x</th>\n",
       "      <th>Open(DJIA)</th>\n",
       "      <th>High(DJIA)</th>\n",
       "      <th>Low(DJIA)</th>\n",
       "      <th>Close(DJIA)</th>\n",
       "      <th>Adj Close_y</th>\n",
       "      <th>Volume_y</th>\n",
       "      <th>sma5</th>\n",
       "      <th>sma10</th>\n",
       "      <th>sma25</th>\n",
       "      <th>sma75</th>\n",
       "      <th>sma125</th>\n",
       "      <th>sma200</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1985-01-29</td>\n",
       "      <td>11843.070313</td>\n",
       "      <td>11843.070313</td>\n",
       "      <td>11843.070313</td>\n",
       "      <td>11843.070313</td>\n",
       "      <td>11843.070313</td>\n",
       "      <td>0</td>\n",
       "      <td>1297.369995</td>\n",
       "      <td>1305.099976</td>\n",
       "      <td>1278.930054</td>\n",
       "      <td>1287.880005</td>\n",
       "      <td>1287.880005</td>\n",
       "      <td>16820000</td>\n",
       "      <td>1287.880005</td>\n",
       "      <td>1287.880005</td>\n",
       "      <td>1287.880005</td>\n",
       "      <td>1287.880005</td>\n",
       "      <td>1287.880005</td>\n",
       "      <td>1287.880005</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1985-01-30</td>\n",
       "      <td>11960.629883</td>\n",
       "      <td>11960.629883</td>\n",
       "      <td>11960.629883</td>\n",
       "      <td>11960.629883</td>\n",
       "      <td>11960.629883</td>\n",
       "      <td>0</td>\n",
       "      <td>1283.239990</td>\n",
       "      <td>1293.400024</td>\n",
       "      <td>1272.640015</td>\n",
       "      <td>1286.770020</td>\n",
       "      <td>1286.770020</td>\n",
       "      <td>14070000</td>\n",
       "      <td>1286.770020</td>\n",
       "      <td>1286.770020</td>\n",
       "      <td>1286.770020</td>\n",
       "      <td>1286.770020</td>\n",
       "      <td>1286.770020</td>\n",
       "      <td>1286.770020</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1985-01-31</td>\n",
       "      <td>11992.309570</td>\n",
       "      <td>11992.309570</td>\n",
       "      <td>11992.309570</td>\n",
       "      <td>11992.309570</td>\n",
       "      <td>11992.309570</td>\n",
       "      <td>0</td>\n",
       "      <td>1276.939941</td>\n",
       "      <td>1286.109985</td>\n",
       "      <td>1269.770020</td>\n",
       "      <td>1277.719971</td>\n",
       "      <td>1277.719971</td>\n",
       "      <td>10980000</td>\n",
       "      <td>1277.719971</td>\n",
       "      <td>1277.719971</td>\n",
       "      <td>1277.719971</td>\n",
       "      <td>1277.719971</td>\n",
       "      <td>1277.719971</td>\n",
       "      <td>1277.719971</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1985-02-04</td>\n",
       "      <td>11891.080078</td>\n",
       "      <td>11891.080078</td>\n",
       "      <td>11891.080078</td>\n",
       "      <td>11891.080078</td>\n",
       "      <td>11891.080078</td>\n",
       "      <td>0</td>\n",
       "      <td>1294.060059</td>\n",
       "      <td>1301.130005</td>\n",
       "      <td>1278.599976</td>\n",
       "      <td>1285.229980</td>\n",
       "      <td>1285.229980</td>\n",
       "      <td>13800000</td>\n",
       "      <td>1285.535986</td>\n",
       "      <td>1285.229980</td>\n",
       "      <td>1285.229980</td>\n",
       "      <td>1285.229980</td>\n",
       "      <td>1285.229980</td>\n",
       "      <td>1285.229980</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1985-02-05</td>\n",
       "      <td>11823.429688</td>\n",
       "      <td>11823.429688</td>\n",
       "      <td>11823.429688</td>\n",
       "      <td>11823.429688</td>\n",
       "      <td>11823.429688</td>\n",
       "      <td>0</td>\n",
       "      <td>1284.890015</td>\n",
       "      <td>1294.500000</td>\n",
       "      <td>1275.069946</td>\n",
       "      <td>1280.589966</td>\n",
       "      <td>1280.589966</td>\n",
       "      <td>14610000</td>\n",
       "      <td>1284.077979</td>\n",
       "      <td>1280.589966</td>\n",
       "      <td>1280.589966</td>\n",
       "      <td>1280.589966</td>\n",
       "      <td>1280.589966</td>\n",
       "      <td>1280.589966</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         Date  Open(Nikkei)  High(Nikkei)   Low(Nikkei)  Close(Nikkei)  \\\n",
       "0  1985-01-29  11843.070313  11843.070313  11843.070313   11843.070313   \n",
       "1  1985-01-30  11960.629883  11960.629883  11960.629883   11960.629883   \n",
       "2  1985-01-31  11992.309570  11992.309570  11992.309570   11992.309570   \n",
       "3  1985-02-04  11891.080078  11891.080078  11891.080078   11891.080078   \n",
       "4  1985-02-05  11823.429688  11823.429688  11823.429688   11823.429688   \n",
       "\n",
       "    Adj Close_x  Volume_x   Open(DJIA)   High(DJIA)    Low(DJIA)  Close(DJIA)  \\\n",
       "0  11843.070313         0  1297.369995  1305.099976  1278.930054  1287.880005   \n",
       "1  11960.629883         0  1283.239990  1293.400024  1272.640015  1286.770020   \n",
       "2  11992.309570         0  1276.939941  1286.109985  1269.770020  1277.719971   \n",
       "3  11891.080078         0  1294.060059  1301.130005  1278.599976  1285.229980   \n",
       "4  11823.429688         0  1284.890015  1294.500000  1275.069946  1280.589966   \n",
       "\n",
       "   Adj Close_y  Volume_y         sma5        sma10        sma25        sma75  \\\n",
       "0  1287.880005  16820000  1287.880005  1287.880005  1287.880005  1287.880005   \n",
       "1  1286.770020  14070000  1286.770020  1286.770020  1286.770020  1286.770020   \n",
       "2  1277.719971  10980000  1277.719971  1277.719971  1277.719971  1277.719971   \n",
       "3  1285.229980  13800000  1285.535986  1285.229980  1285.229980  1285.229980   \n",
       "4  1280.589966  14610000  1284.077979  1280.589966  1280.589966  1280.589966   \n",
       "\n",
       "        sma125       sma200  \n",
       "0  1287.880005  1287.880005  \n",
       "1  1286.770020  1286.770020  \n",
       "2  1277.719971  1277.719971  \n",
       "3  1285.229980  1285.229980  \n",
       "4  1280.589966  1280.589966  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# df_djia_dc.iloc[0]['Date']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "event_col = ['Open(Nikkei)', 'High(Nikkei)', 'Low(Nikkei)', 'Close(Nikkei)',\n",
    "            'Open(DJIA)', 'High(DJIA)', 'Low(DJIA)', 'Close(DJIA)'\n",
    "            ]\n",
    "total_col = df.columns\n",
    "for col in total_col:\n",
    "    if col not in event_col:\n",
    "        df = df.drop(col, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Open(Nikkei)</th>\n",
       "      <th>High(Nikkei)</th>\n",
       "      <th>Low(Nikkei)</th>\n",
       "      <th>Close(Nikkei)</th>\n",
       "      <th>Open(DJIA)</th>\n",
       "      <th>High(DJIA)</th>\n",
       "      <th>Low(DJIA)</th>\n",
       "      <th>Close(DJIA)</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>6906</th>\n",
       "      <td>22619.769531</td>\n",
       "      <td>22648.810547</td>\n",
       "      <td>22457.890625</td>\n",
       "      <td>22625.380859</td>\n",
       "      <td>26893.92969</td>\n",
       "      <td>26931.77930</td>\n",
       "      <td>26714.33984</td>\n",
       "      <td>26805.52930</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6907</th>\n",
       "      <td>22725.439453</td>\n",
       "      <td>22780.990234</td>\n",
       "      <td>22704.330078</td>\n",
       "      <td>22750.599609</td>\n",
       "      <td>26789.60938</td>\n",
       "      <td>27015.36914</td>\n",
       "      <td>26765.67969</td>\n",
       "      <td>26958.06055</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6908</th>\n",
       "      <td>22854.439453</td>\n",
       "      <td>22896.220703</td>\n",
       "      <td>22830.570313</td>\n",
       "      <td>22867.269531</td>\n",
       "      <td>27061.07031</td>\n",
       "      <td>27165.93945</td>\n",
       "      <td>27039.75977</td>\n",
       "      <td>27071.46094</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6909</th>\n",
       "      <td>22950.789063</td>\n",
       "      <td>23008.429688</td>\n",
       "      <td>22935.349609</td>\n",
       "      <td>22974.130859</td>\n",
       "      <td>27110.71094</td>\n",
       "      <td>27204.35938</td>\n",
       "      <td>26999.64063</td>\n",
       "      <td>27186.68945</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6910</th>\n",
       "      <td>22953.169922</td>\n",
       "      <td>22961.230469</td>\n",
       "      <td>22827.929688</td>\n",
       "      <td>22843.119141</td>\n",
       "      <td>27188.36914</td>\n",
       "      <td>27188.36914</td>\n",
       "      <td>26918.28906</td>\n",
       "      <td>27046.23047</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      Open(Nikkei)  High(Nikkei)   Low(Nikkei)  Close(Nikkei)   Open(DJIA)  \\\n",
       "6906  22619.769531  22648.810547  22457.890625   22625.380859  26893.92969   \n",
       "6907  22725.439453  22780.990234  22704.330078   22750.599609  26789.60938   \n",
       "6908  22854.439453  22896.220703  22830.570313   22867.269531  27061.07031   \n",
       "6909  22950.789063  23008.429688  22935.349609   22974.130859  27110.71094   \n",
       "6910  22953.169922  22961.230469  22827.929688   22843.119141  27188.36914   \n",
       "\n",
       "       High(DJIA)    Low(DJIA)  Close(DJIA)  \n",
       "6906  26931.77930  26714.33984  26805.52930  \n",
       "6907  27015.36914  26765.67969  26958.06055  \n",
       "6908  27165.93945  27039.75977  27071.46094  \n",
       "6909  27204.35938  26999.64063  27186.68945  \n",
       "6910  27188.36914  26918.28906  27046.23047  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def normalize_data(df, event_col):\n",
    "    min_max_scaler = sklearn.preprocessing.MinMaxScaler()\n",
    "    for col in event_col:\n",
    "        if col != 'Date':\n",
    "            df[col] = min_max_scaler.fit_transform(df[col].to_numpy().reshape(-1, 1))\n",
    "    return df\n",
    "\n",
    "def load_data(stock, seq_len):\n",
    "    data_raw = stock.as_matrix()\n",
    "    data = list()\n",
    "    for index in range(len(data_raw) - seq_len):\n",
    "        data.append(data_raw[index: index+seq_len])\n",
    "    data = np.array(data)\n",
    "    val_set_size = int(np.round(val_set_size_percentage/100*data.shape[0]))\n",
    "    test_set_size = int(np.round(test_set_size_percentage/100*data.shape[0]))\n",
    "    train_set_size = data.shape[0] - (val_set_size + test_set_size)\n",
    "    \n",
    "    x_train = data[:train_set_size, :-1, :]\n",
    "    y_train = data[:train_set_size, -1, :]\n",
    "    \n",
    "    x_val = data[train_set_size:train_set_size + val_set_size, :-1, :]\n",
    "    y_val = data[train_set_size:train_set_size + val_set_size, -1, :]\n",
    "    \n",
    "    x_test = data[train_set_size+val_set_size:,:-1,:]\n",
    "    y_test = data[train_set_size+val_set_size:,-1,:]\n",
    "    \n",
    "    return [x_train, y_train, x_val, y_val, x_test, y_test]\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "x_train.shape =  (5513, 19, 8)\n",
      "y_train.shape =  (5513, 8)\n",
      "x_val.shape =  (689, 19, 8)\n",
      "y_val.shape =  (689, 8)\n",
      "x_test.shape =  (689, 19, 8)\n",
      "y_test.shape =  (689, 8)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/lib/python3/dist-packages/ipykernel_launcher.py:9: FutureWarning: Method .as_matrix will be removed in a future version. Use .values instead.\n",
      "  if __name__ == '__main__':\n"
     ]
    }
   ],
   "source": [
    "# parameter for preprocessing #\n",
    "\n",
    "val_set_size_percentage = 10\n",
    "test_set_size_percentage = 10\n",
    "seq_len = 20\n",
    "\n",
    "#*****************************#\n",
    "\n",
    "# normalize index\n",
    "df_norm = normalize_data(df, event_col)\n",
    "\n",
    "# split train, val, test data\n",
    "\n",
    "x_train, y_train, x_val, y_val, x_test, y_test = load_data(df_norm, seq_len)\n",
    "\n",
    "print('x_train.shape = ', x_train.shape)\n",
    "print('y_train.shape = ', y_train.shape)\n",
    "print('x_val.shape = ', x_val.shape)\n",
    "print('y_val.shape = ', y_val.shape)\n",
    "print('x_test.shape = ', x_test.shape)\n",
    "print('y_test.shape = ', y_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot graph before train\n",
    "\n",
    "# !set color in graph\n",
    "# color_palette = ...\n",
    "plt.figure(figsize=(15, 10))\n",
    "plt.grid()\n",
    "color_palette = plt.get_cmap('Set1')\n",
    "for i, col in enumerate(event_col):\n",
    "    plt.plot(df_norm[col], color=color_palette(i), label=col)\n",
    "plt.title('Stock price')\n",
    "plt.xlabel('Time [days]')\n",
    "plt.ylabel('Normalized price / volume')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "# hyperparameter \n",
    "batch_size = 1024\n",
    "epochs = 300\n",
    "\n",
    "# set up callback_list\n",
    "weight_path_model = \"saved_model/model_djia_support_nikkei.h5\"\n",
    "tensorboard_path = 'Graph/djia_support_nikkei_1'\n",
    "callbacks_list_model = [\n",
    "    TensorBoard(log_dir=tensorboard_path),\n",
    "    ModelCheckpoint(\n",
    "            weight_path_model,\n",
    "            save_best_only=True,\n",
    "            save_weights_only=True,\n",
    "            monitor='val_loss',\n",
    "            mode='min',\n",
    "            verbose=2\n",
    "        ),\n",
    "    ReduceLROnPlateau(monitor='val_loss', factor=0.9, patience=3, min_lr=1e-9)\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /home/sprkzoff/.local/lib/python3.6/site-packages/tensorflow/python/ops/init_ops.py:97: calling Orthogonal.__init__ (from tensorflow.python.ops.init_ops) with dtype is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Call initializer instance with the dtype argument instead of passing it to the constructor\n",
      "WARNING:tensorflow:From /home/sprkzoff/.local/lib/python3.6/site-packages/tensorflow/python/ops/init_ops.py:97: calling Zeros.__init__ (from tensorflow.python.ops.init_ops) with dtype is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Call initializer instance with the dtype argument instead of passing it to the constructor\n",
      "Model: \"model\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_1 (InputLayer)         [(None, 19, 8)]           0         \n",
      "_________________________________________________________________\n",
      "bidirectional (Bidirectional (None, 19, 512)           542720    \n",
      "_________________________________________________________________\n",
      "time_distributed (TimeDistri (None, 19, 512)           0         \n",
      "_________________________________________________________________\n",
      "time_distributed_1 (TimeDist (None, 19, 512)           2048      \n",
      "_________________________________________________________________\n",
      "gru (GRU)                    (None, 128)               246144    \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_1 (LeakyReLU)    (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "batch_normalization_1 (Batch (None, 128)               512       \n",
      "_________________________________________________________________\n",
      "dropout (Dropout)            (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 8)                 1032      \n",
      "=================================================================\n",
      "Total params: 792,456\n",
      "Trainable params: 791,176\n",
      "Non-trainable params: 1,280\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "K.clear_session()\n",
    "\n",
    "def get_bi_gru():\n",
    "    \n",
    "    input1 = Input(shape=(19,8,))\n",
    "    x = Bidirectional(LSTM(256, return_sequences=True,  kernel_initializer='Orthogonal'))(input1)\n",
    "    x = TimeDistributed(LeakyReLU())(x)\n",
    "    x = TimeDistributed(BatchNormalization())(x)\n",
    "    x = GRU(128,kernel_initializer='Orthogonal')(x)\n",
    "    x =(LeakyReLU())(x)\n",
    "    x = (BatchNormalization())(x)\n",
    "    x = (Dropout(0.25))(x)\n",
    "    out = Dense(8,kernel_initializer='he_normal')(x)\n",
    "\n",
    "    \n",
    "    \n",
    "    model = Model(inputs=input1, outputs=out)\n",
    "    model.compile(optimizer=Adam(lr=1e-2, decay=1e-2),\n",
    "                loss='mse',\n",
    "                metrics=['mse'])   \n",
    "    \n",
    "    return model\n",
    "\n",
    "model = get_bi_gru()\n",
    "model.summary()\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Experiment and Result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 5513 samples, validate on 689 samples\n",
      "WARNING:tensorflow:From /home/sprkzoff/.local/lib/python3.6/site-packages/tensorflow/python/ops/math_grad.py:1250: add_dispatch_support.<locals>.wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.where in 2.0, which has the same broadcast rule as np.where\n",
      "Epoch 1/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 2.7933 - mean_squared_error: 2.7933\n",
      "Epoch 00001: val_loss improved from inf to 0.70757, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 6s 1ms/sample - loss: 2.6164 - mean_squared_error: 2.6164 - val_loss: 0.7076 - val_mean_squared_error: 0.7076\n",
      "Epoch 2/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.1221 - mean_squared_error: 0.1221\n",
      "Epoch 00002: val_loss did not improve from 0.70757\n",
      "5513/5513 [==============================] - 5s 844us/sample - loss: 0.1181 - mean_squared_error: 0.1181 - val_loss: 0.7946 - val_mean_squared_error: 0.7946\n",
      "Epoch 3/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0493 - mean_squared_error: 0.0493\n",
      "Epoch 00003: val_loss did not improve from 0.70757\n",
      "5513/5513 [==============================] - 4s 767us/sample - loss: 0.0478 - mean_squared_error: 0.0478 - val_loss: 0.7847 - val_mean_squared_error: 0.7847\n",
      "Epoch 4/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0202 - mean_squared_error: 0.0202\n",
      "Epoch 00004: val_loss did not improve from 0.70757\n",
      "5513/5513 [==============================] - 4s 786us/sample - loss: 0.0197 - mean_squared_error: 0.0197 - val_loss: 0.7586 - val_mean_squared_error: 0.7586\n",
      "Epoch 5/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0143 - mean_squared_error: 0.0143\n",
      "Epoch 00005: val_loss did not improve from 0.70757\n",
      "5513/5513 [==============================] - 5s 976us/sample - loss: 0.0142 - mean_squared_error: 0.0142 - val_loss: 0.7793 - val_mean_squared_error: 0.7793\n",
      "Epoch 6/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0116 - mean_squared_error: 0.0116\n",
      "Epoch 00006: val_loss did not improve from 0.70757\n",
      "5513/5513 [==============================] - 5s 961us/sample - loss: 0.0115 - mean_squared_error: 0.0115 - val_loss: 0.8096 - val_mean_squared_error: 0.8096\n",
      "Epoch 7/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0090 - mean_squared_error: 0.0090\n",
      "Epoch 00007: val_loss did not improve from 0.70757\n",
      "5513/5513 [==============================] - 5s 925us/sample - loss: 0.0090 - mean_squared_error: 0.0090 - val_loss: 0.8237 - val_mean_squared_error: 0.8237\n",
      "Epoch 8/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0083 - mean_squared_error: 0.0083\n",
      "Epoch 00008: val_loss did not improve from 0.70757\n",
      "5513/5513 [==============================] - 5s 918us/sample - loss: 0.0083 - mean_squared_error: 0.0083 - val_loss: 0.8296 - val_mean_squared_error: 0.8296\n",
      "Epoch 9/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0073 - mean_squared_error: 0.0073\n",
      "Epoch 00009: val_loss did not improve from 0.70757\n",
      "5513/5513 [==============================] - 5s 918us/sample - loss: 0.0073 - mean_squared_error: 0.0073 - val_loss: 0.8055 - val_mean_squared_error: 0.8055\n",
      "Epoch 10/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0071 - mean_squared_error: 0.0071\n",
      "Epoch 00010: val_loss did not improve from 0.70757\n",
      "5513/5513 [==============================] - 5s 926us/sample - loss: 0.0071 - mean_squared_error: 0.0071 - val_loss: 0.7489 - val_mean_squared_error: 0.7489\n",
      "Epoch 11/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0064 - mean_squared_error: 0.0064\n",
      "Epoch 00011: val_loss improved from 0.70757 to 0.68485, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 935us/sample - loss: 0.0064 - mean_squared_error: 0.0064 - val_loss: 0.6848 - val_mean_squared_error: 0.6848\n",
      "Epoch 12/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0061 - mean_squared_error: 0.0061\n",
      "Epoch 00012: val_loss improved from 0.68485 to 0.63305, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 955us/sample - loss: 0.0061 - mean_squared_error: 0.0061 - val_loss: 0.6331 - val_mean_squared_error: 0.6331\n",
      "Epoch 13/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0058 - mean_squared_error: 0.0058\n",
      "Epoch 00013: val_loss improved from 0.63305 to 0.59857, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 925us/sample - loss: 0.0057 - mean_squared_error: 0.0057 - val_loss: 0.5986 - val_mean_squared_error: 0.5986\n",
      "Epoch 14/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0056 - mean_squared_error: 0.0056\n",
      "Epoch 00014: val_loss improved from 0.59857 to 0.56147, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 929us/sample - loss: 0.0056 - mean_squared_error: 0.0056 - val_loss: 0.5615 - val_mean_squared_error: 0.5615\n",
      "Epoch 15/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0055 - mean_squared_error: 0.0055\n",
      "Epoch 00015: val_loss improved from 0.56147 to 0.51297, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 926us/sample - loss: 0.0054 - mean_squared_error: 0.0054 - val_loss: 0.5130 - val_mean_squared_error: 0.5130\n",
      "Epoch 16/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0051 - mean_squared_error: 0.0051\n",
      "Epoch 00016: val_loss improved from 0.51297 to 0.46404, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 931us/sample - loss: 0.0052 - mean_squared_error: 0.0052 - val_loss: 0.4640 - val_mean_squared_error: 0.4640\n",
      "Epoch 17/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0050 - mean_squared_error: 0.0050\n",
      "Epoch 00017: val_loss improved from 0.46404 to 0.40164, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 923us/sample - loss: 0.0050 - mean_squared_error: 0.0050 - val_loss: 0.4016 - val_mean_squared_error: 0.4016\n",
      "Epoch 18/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0048 - mean_squared_error: 0.0048\n",
      "Epoch 00018: val_loss improved from 0.40164 to 0.32947, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 944us/sample - loss: 0.0049 - mean_squared_error: 0.0049 - val_loss: 0.3295 - val_mean_squared_error: 0.3295\n",
      "Epoch 19/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0047 - mean_squared_error: 0.0047\n",
      "Epoch 00019: val_loss improved from 0.32947 to 0.26624, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 935us/sample - loss: 0.0048 - mean_squared_error: 0.0048 - val_loss: 0.2662 - val_mean_squared_error: 0.2662\n",
      "Epoch 20/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0045 - mean_squared_error: 0.0045\n",
      "Epoch 00020: val_loss improved from 0.26624 to 0.21113, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 928us/sample - loss: 0.0045 - mean_squared_error: 0.0045 - val_loss: 0.2111 - val_mean_squared_error: 0.2111\n",
      "Epoch 21/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0044 - mean_squared_error: 0.0044\n",
      "Epoch 00021: val_loss improved from 0.21113 to 0.16395, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 6s 1ms/sample - loss: 0.0044 - mean_squared_error: 0.0044 - val_loss: 0.1640 - val_mean_squared_error: 0.1640\n",
      "Epoch 22/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0043 - mean_squared_error: 0.0043\n",
      "Epoch 00022: val_loss improved from 0.16395 to 0.12658, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 6s 1000us/sample - loss: 0.0043 - mean_squared_error: 0.0043 - val_loss: 0.1266 - val_mean_squared_error: 0.1266\n",
      "Epoch 23/300\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0042 - mean_squared_error: 0.0042\n",
      "Epoch 00023: val_loss improved from 0.12658 to 0.10118, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 960us/sample - loss: 0.0042 - mean_squared_error: 0.0042 - val_loss: 0.1012 - val_mean_squared_error: 0.1012\n",
      "Epoch 24/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0040 - mean_squared_error: 0.0040\n",
      "Epoch 00024: val_loss improved from 0.10118 to 0.08710, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 945us/sample - loss: 0.0040 - mean_squared_error: 0.0040 - val_loss: 0.0871 - val_mean_squared_error: 0.0871\n",
      "Epoch 25/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0041 - mean_squared_error: 0.0041\n",
      "Epoch 00025: val_loss improved from 0.08710 to 0.07910, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 982us/sample - loss: 0.0041 - mean_squared_error: 0.0041 - val_loss: 0.0791 - val_mean_squared_error: 0.0791\n",
      "Epoch 26/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0040 - mean_squared_error: 0.0040\n",
      "Epoch 00026: val_loss improved from 0.07910 to 0.07367, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 970us/sample - loss: 0.0040 - mean_squared_error: 0.0040 - val_loss: 0.0737 - val_mean_squared_error: 0.0737\n",
      "Epoch 27/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0037 - mean_squared_error: 0.0037\n",
      "Epoch 00027: val_loss improved from 0.07367 to 0.06898, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 934us/sample - loss: 0.0038 - mean_squared_error: 0.0038 - val_loss: 0.0690 - val_mean_squared_error: 0.0690\n",
      "Epoch 28/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0037 - mean_squared_error: 0.0037\n",
      "Epoch 00028: val_loss improved from 0.06898 to 0.06558, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 983us/sample - loss: 0.0037 - mean_squared_error: 0.0037 - val_loss: 0.0656 - val_mean_squared_error: 0.0656\n",
      "Epoch 29/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0037 - mean_squared_error: 0.0037\n",
      "Epoch 00029: val_loss improved from 0.06558 to 0.06382, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 934us/sample - loss: 0.0037 - mean_squared_error: 0.0037 - val_loss: 0.0638 - val_mean_squared_error: 0.0638\n",
      "Epoch 30/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0036 - mean_squared_error: 0.0036\n",
      "Epoch 00030: val_loss did not improve from 0.06382\n",
      "5513/5513 [==============================] - 5s 952us/sample - loss: 0.0036 - mean_squared_error: 0.0036 - val_loss: 0.0638 - val_mean_squared_error: 0.0638\n",
      "Epoch 31/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0036 - mean_squared_error: 0.0036\n",
      "Epoch 00031: val_loss did not improve from 0.06382\n",
      "5513/5513 [==============================] - 5s 986us/sample - loss: 0.0036 - mean_squared_error: 0.0036 - val_loss: 0.0654 - val_mean_squared_error: 0.0654\n",
      "Epoch 32/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0035 - mean_squared_error: 0.0035\n",
      "Epoch 00032: val_loss did not improve from 0.06382\n",
      "5513/5513 [==============================] - 6s 1ms/sample - loss: 0.0036 - mean_squared_error: 0.0036 - val_loss: 0.0653 - val_mean_squared_error: 0.0653\n",
      "Epoch 33/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0034 - mean_squared_error: 0.0034\n",
      "Epoch 00033: val_loss did not improve from 0.06382\n",
      "5513/5513 [==============================] - 5s 945us/sample - loss: 0.0034 - mean_squared_error: 0.0034 - val_loss: 0.0655 - val_mean_squared_error: 0.0655\n",
      "Epoch 34/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0033 - mean_squared_error: 0.0033\n",
      "Epoch 00034: val_loss did not improve from 0.06382\n",
      "5513/5513 [==============================] - 5s 914us/sample - loss: 0.0034 - mean_squared_error: 0.0034 - val_loss: 0.0665 - val_mean_squared_error: 0.0665\n",
      "Epoch 35/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0034 - mean_squared_error: 0.0034\n",
      "Epoch 00035: val_loss did not improve from 0.06382\n",
      "5513/5513 [==============================] - 5s 946us/sample - loss: 0.0034 - mean_squared_error: 0.0034 - val_loss: 0.0662 - val_mean_squared_error: 0.0662\n",
      "Epoch 36/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0033 - mean_squared_error: 0.0033\n",
      "Epoch 00036: val_loss did not improve from 0.06382\n",
      "5513/5513 [==============================] - 5s 922us/sample - loss: 0.0033 - mean_squared_error: 0.0033 - val_loss: 0.0659 - val_mean_squared_error: 0.0659\n",
      "Epoch 37/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0032 - mean_squared_error: 0.0032\n",
      "Epoch 00037: val_loss did not improve from 0.06382\n",
      "5513/5513 [==============================] - 5s 923us/sample - loss: 0.0032 - mean_squared_error: 0.0032 - val_loss: 0.0668 - val_mean_squared_error: 0.0668\n",
      "Epoch 38/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0032 - mean_squared_error: 0.0032\n",
      "Epoch 00038: val_loss did not improve from 0.06382\n",
      "5513/5513 [==============================] - 5s 930us/sample - loss: 0.0032 - mean_squared_error: 0.0032 - val_loss: 0.0682 - val_mean_squared_error: 0.0682\n",
      "Epoch 39/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0031 - mean_squared_error: 0.0031\n",
      "Epoch 00039: val_loss did not improve from 0.06382\n",
      "5513/5513 [==============================] - 5s 976us/sample - loss: 0.0031 - mean_squared_error: 0.0031 - val_loss: 0.0676 - val_mean_squared_error: 0.0676\n",
      "Epoch 40/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0032 - mean_squared_error: 0.0032\n",
      "Epoch 00040: val_loss did not improve from 0.06382\n",
      "5513/5513 [==============================] - 5s 967us/sample - loss: 0.0032 - mean_squared_error: 0.0032 - val_loss: 0.0678 - val_mean_squared_error: 0.0678\n",
      "Epoch 41/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0030 - mean_squared_error: 0.0030\n",
      "Epoch 00041: val_loss did not improve from 0.06382\n",
      "5513/5513 [==============================] - 5s 974us/sample - loss: 0.0030 - mean_squared_error: 0.0030 - val_loss: 0.0687 - val_mean_squared_error: 0.0687\n",
      "Epoch 42/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0031 - mean_squared_error: 0.0031\n",
      "Epoch 00042: val_loss did not improve from 0.06382\n",
      "5513/5513 [==============================] - 5s 969us/sample - loss: 0.0031 - mean_squared_error: 0.0031 - val_loss: 0.0685 - val_mean_squared_error: 0.0685\n",
      "Epoch 43/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0031 - mean_squared_error: 0.0031\n",
      "Epoch 00043: val_loss did not improve from 0.06382\n",
      "5513/5513 [==============================] - 5s 947us/sample - loss: 0.0031 - mean_squared_error: 0.0031 - val_loss: 0.0679 - val_mean_squared_error: 0.0679\n",
      "Epoch 44/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0030 - mean_squared_error: 0.0030\n",
      "Epoch 00044: val_loss did not improve from 0.06382\n",
      "5513/5513 [==============================] - 5s 957us/sample - loss: 0.0030 - mean_squared_error: 0.0030 - val_loss: 0.0663 - val_mean_squared_error: 0.0663\n",
      "Epoch 45/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0030 - mean_squared_error: 0.0030\n",
      "Epoch 00045: val_loss did not improve from 0.06382\n",
      "5513/5513 [==============================] - 5s 936us/sample - loss: 0.0030 - mean_squared_error: 0.0030 - val_loss: 0.0670 - val_mean_squared_error: 0.0670\n",
      "Epoch 46/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0030 - mean_squared_error: 0.0030\n",
      "Epoch 00046: val_loss did not improve from 0.06382\n",
      "5513/5513 [==============================] - 5s 968us/sample - loss: 0.0030 - mean_squared_error: 0.0030 - val_loss: 0.0672 - val_mean_squared_error: 0.0672\n",
      "Epoch 47/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0030 - mean_squared_error: 0.0030\n",
      "Epoch 00047: val_loss did not improve from 0.06382\n",
      "5513/5513 [==============================] - 5s 989us/sample - loss: 0.0030 - mean_squared_error: 0.0030 - val_loss: 0.0647 - val_mean_squared_error: 0.0647\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 48/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0029 - mean_squared_error: 0.0029\n",
      "Epoch 00048: val_loss improved from 0.06382 to 0.06255, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 922us/sample - loss: 0.0029 - mean_squared_error: 0.0029 - val_loss: 0.0625 - val_mean_squared_error: 0.0625\n",
      "Epoch 49/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0030 - mean_squared_error: 0.0030\n",
      "Epoch 00049: val_loss improved from 0.06255 to 0.06172, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 915us/sample - loss: 0.0030 - mean_squared_error: 0.0030 - val_loss: 0.0617 - val_mean_squared_error: 0.0617\n",
      "Epoch 50/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0030 - mean_squared_error: 0.0030\n",
      "Epoch 00050: val_loss improved from 0.06172 to 0.06027, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 920us/sample - loss: 0.0030 - mean_squared_error: 0.0030 - val_loss: 0.0603 - val_mean_squared_error: 0.0603\n",
      "Epoch 51/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0029 - mean_squared_error: 0.0029\n",
      "Epoch 00051: val_loss did not improve from 0.06027\n",
      "5513/5513 [==============================] - 5s 905us/sample - loss: 0.0029 - mean_squared_error: 0.0029 - val_loss: 0.0603 - val_mean_squared_error: 0.0603\n",
      "Epoch 52/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0029 - mean_squared_error: 0.0029\n",
      "Epoch 00052: val_loss improved from 0.06027 to 0.05954, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 923us/sample - loss: 0.0029 - mean_squared_error: 0.0029 - val_loss: 0.0595 - val_mean_squared_error: 0.0595\n",
      "Epoch 53/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0029 - mean_squared_error: 0.0029\n",
      "Epoch 00053: val_loss improved from 0.05954 to 0.05909, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 934us/sample - loss: 0.0029 - mean_squared_error: 0.0029 - val_loss: 0.0591 - val_mean_squared_error: 0.0591\n",
      "Epoch 54/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0029 - mean_squared_error: 0.0029\n",
      "Epoch 00054: val_loss improved from 0.05909 to 0.05782, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 959us/sample - loss: 0.0029 - mean_squared_error: 0.0029 - val_loss: 0.0578 - val_mean_squared_error: 0.0578\n",
      "Epoch 55/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0029 - mean_squared_error: 0.0029\n",
      "Epoch 00055: val_loss improved from 0.05782 to 0.05688, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 965us/sample - loss: 0.0029 - mean_squared_error: 0.0029 - val_loss: 0.0569 - val_mean_squared_error: 0.0569\n",
      "Epoch 56/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0029 - mean_squared_error: 0.0029\n",
      "Epoch 00056: val_loss improved from 0.05688 to 0.05580, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 954us/sample - loss: 0.0029 - mean_squared_error: 0.0029 - val_loss: 0.0558 - val_mean_squared_error: 0.0558\n",
      "Epoch 57/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0029 - mean_squared_error: 0.0029\n",
      "Epoch 00057: val_loss improved from 0.05580 to 0.05522, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 985us/sample - loss: 0.0029 - mean_squared_error: 0.0029 - val_loss: 0.0552 - val_mean_squared_error: 0.0552\n",
      "Epoch 58/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0028 - mean_squared_error: 0.0028\n",
      "Epoch 00058: val_loss improved from 0.05522 to 0.05419, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 939us/sample - loss: 0.0028 - mean_squared_error: 0.0028 - val_loss: 0.0542 - val_mean_squared_error: 0.0542\n",
      "Epoch 59/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0028 - mean_squared_error: 0.0028\n",
      "Epoch 00059: val_loss improved from 0.05419 to 0.05299, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 931us/sample - loss: 0.0028 - mean_squared_error: 0.0028 - val_loss: 0.0530 - val_mean_squared_error: 0.0530\n",
      "Epoch 60/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0027 - mean_squared_error: 0.0027\n",
      "Epoch 00060: val_loss improved from 0.05299 to 0.05214, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 935us/sample - loss: 0.0028 - mean_squared_error: 0.0028 - val_loss: 0.0521 - val_mean_squared_error: 0.0521\n",
      "Epoch 61/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0028 - mean_squared_error: 0.0028\n",
      "Epoch 00061: val_loss did not improve from 0.05214\n",
      "5513/5513 [==============================] - 5s 926us/sample - loss: 0.0028 - mean_squared_error: 0.0028 - val_loss: 0.0523 - val_mean_squared_error: 0.0523\n",
      "Epoch 62/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0028 - mean_squared_error: 0.0028\n",
      "Epoch 00062: val_loss improved from 0.05214 to 0.05137, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 934us/sample - loss: 0.0028 - mean_squared_error: 0.0028 - val_loss: 0.0514 - val_mean_squared_error: 0.0514\n",
      "Epoch 63/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0028 - mean_squared_error: 0.0028\n",
      "Epoch 00063: val_loss improved from 0.05137 to 0.05004, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 918us/sample - loss: 0.0028 - mean_squared_error: 0.0028 - val_loss: 0.0500 - val_mean_squared_error: 0.0500\n",
      "Epoch 64/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0028 - mean_squared_error: 0.0028\n",
      "Epoch 00064: val_loss improved from 0.05004 to 0.04945, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 929us/sample - loss: 0.0028 - mean_squared_error: 0.0028 - val_loss: 0.0494 - val_mean_squared_error: 0.0494\n",
      "Epoch 65/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0027 - mean_squared_error: 0.0027\n",
      "Epoch 00065: val_loss improved from 0.04945 to 0.04944, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 925us/sample - loss: 0.0027 - mean_squared_error: 0.0027 - val_loss: 0.0494 - val_mean_squared_error: 0.0494\n",
      "Epoch 66/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0027 - mean_squared_error: 0.0027\n",
      "Epoch 00066: val_loss improved from 0.04944 to 0.04891, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 933us/sample - loss: 0.0027 - mean_squared_error: 0.0027 - val_loss: 0.0489 - val_mean_squared_error: 0.0489\n",
      "Epoch 67/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0027 - mean_squared_error: 0.0027\n",
      "Epoch 00067: val_loss improved from 0.04891 to 0.04801, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 926us/sample - loss: 0.0028 - mean_squared_error: 0.0028 - val_loss: 0.0480 - val_mean_squared_error: 0.0480\n",
      "Epoch 68/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0027 - mean_squared_error: 0.0027\n",
      "Epoch 00068: val_loss improved from 0.04801 to 0.04719, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 933us/sample - loss: 0.0027 - mean_squared_error: 0.0027 - val_loss: 0.0472 - val_mean_squared_error: 0.0472\n",
      "Epoch 69/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0027 - mean_squared_error: 0.0027\n",
      "Epoch 00069: val_loss did not improve from 0.04719\n",
      "5513/5513 [==============================] - 5s 947us/sample - loss: 0.0026 - mean_squared_error: 0.0026 - val_loss: 0.0474 - val_mean_squared_error: 0.0474\n",
      "Epoch 70/300\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0027 - mean_squared_error: 0.0027\n",
      "Epoch 00070: val_loss improved from 0.04719 to 0.04605, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 935us/sample - loss: 0.0027 - mean_squared_error: 0.0027 - val_loss: 0.0460 - val_mean_squared_error: 0.0460\n",
      "Epoch 71/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0027 - mean_squared_error: 0.0027\n",
      "Epoch 00071: val_loss improved from 0.04605 to 0.04505, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 925us/sample - loss: 0.0027 - mean_squared_error: 0.0027 - val_loss: 0.0451 - val_mean_squared_error: 0.0451\n",
      "Epoch 72/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0027 - mean_squared_error: 0.0027\n",
      "Epoch 00072: val_loss did not improve from 0.04505\n",
      "5513/5513 [==============================] - 5s 930us/sample - loss: 0.0027 - mean_squared_error: 0.0027 - val_loss: 0.0454 - val_mean_squared_error: 0.0454\n",
      "Epoch 73/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0026 - mean_squared_error: 0.0026\n",
      "Epoch 00073: val_loss did not improve from 0.04505\n",
      "5513/5513 [==============================] - 5s 946us/sample - loss: 0.0026 - mean_squared_error: 0.0026 - val_loss: 0.0458 - val_mean_squared_error: 0.0458\n",
      "Epoch 74/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0027 - mean_squared_error: 0.0027\n",
      "Epoch 00074: val_loss improved from 0.04505 to 0.04428, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 968us/sample - loss: 0.0027 - mean_squared_error: 0.0027 - val_loss: 0.0443 - val_mean_squared_error: 0.0443\n",
      "Epoch 75/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0026 - mean_squared_error: 0.0026\n",
      "Epoch 00075: val_loss did not improve from 0.04428\n",
      "5513/5513 [==============================] - 5s 955us/sample - loss: 0.0026 - mean_squared_error: 0.0026 - val_loss: 0.0449 - val_mean_squared_error: 0.0449\n",
      "Epoch 76/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0027 - mean_squared_error: 0.0027\n",
      "Epoch 00076: val_loss improved from 0.04428 to 0.04367, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 953us/sample - loss: 0.0027 - mean_squared_error: 0.0027 - val_loss: 0.0437 - val_mean_squared_error: 0.0437\n",
      "Epoch 77/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0026 - mean_squared_error: 0.0026\n",
      "Epoch 00077: val_loss improved from 0.04367 to 0.04224, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 941us/sample - loss: 0.0026 - mean_squared_error: 0.0026 - val_loss: 0.0422 - val_mean_squared_error: 0.0422\n",
      "Epoch 78/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0025 - mean_squared_error: 0.0025\n",
      "Epoch 00078: val_loss improved from 0.04224 to 0.04157, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 988us/sample - loss: 0.0025 - mean_squared_error: 0.0025 - val_loss: 0.0416 - val_mean_squared_error: 0.0416\n",
      "Epoch 79/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0026 - mean_squared_error: 0.0026\n",
      "Epoch 00079: val_loss improved from 0.04157 to 0.03993, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 960us/sample - loss: 0.0026 - mean_squared_error: 0.0026 - val_loss: 0.0399 - val_mean_squared_error: 0.0399\n",
      "Epoch 80/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0026 - mean_squared_error: 0.0026\n",
      "Epoch 00080: val_loss improved from 0.03993 to 0.03860, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 931us/sample - loss: 0.0026 - mean_squared_error: 0.0026 - val_loss: 0.0386 - val_mean_squared_error: 0.0386\n",
      "Epoch 81/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0026 - mean_squared_error: 0.0026\n",
      "Epoch 00081: val_loss did not improve from 0.03860\n",
      "5513/5513 [==============================] - 5s 945us/sample - loss: 0.0026 - mean_squared_error: 0.0026 - val_loss: 0.0393 - val_mean_squared_error: 0.0393\n",
      "Epoch 82/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0026 - mean_squared_error: 0.0026\n",
      "Epoch 00082: val_loss did not improve from 0.03860\n",
      "5513/5513 [==============================] - 5s 924us/sample - loss: 0.0026 - mean_squared_error: 0.0026 - val_loss: 0.0391 - val_mean_squared_error: 0.0391\n",
      "Epoch 83/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0026 - mean_squared_error: 0.0026\n",
      "Epoch 00083: val_loss improved from 0.03860 to 0.03859, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 925us/sample - loss: 0.0026 - mean_squared_error: 0.0026 - val_loss: 0.0386 - val_mean_squared_error: 0.0386\n",
      "Epoch 84/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0025 - mean_squared_error: 0.0025\n",
      "Epoch 00084: val_loss improved from 0.03859 to 0.03776, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 969us/sample - loss: 0.0026 - mean_squared_error: 0.0026 - val_loss: 0.0378 - val_mean_squared_error: 0.0378\n",
      "Epoch 85/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0026 - mean_squared_error: 0.0026\n",
      "Epoch 00085: val_loss did not improve from 0.03776\n",
      "5513/5513 [==============================] - 5s 938us/sample - loss: 0.0026 - mean_squared_error: 0.0026 - val_loss: 0.0382 - val_mean_squared_error: 0.0382\n",
      "Epoch 86/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0025 - mean_squared_error: 0.0025\n",
      "Epoch 00086: val_loss did not improve from 0.03776\n",
      "5513/5513 [==============================] - 5s 926us/sample - loss: 0.0026 - mean_squared_error: 0.0026 - val_loss: 0.0382 - val_mean_squared_error: 0.0382\n",
      "Epoch 87/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0025 - mean_squared_error: 0.0025\n",
      "Epoch 00087: val_loss improved from 0.03776 to 0.03568, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 938us/sample - loss: 0.0025 - mean_squared_error: 0.0025 - val_loss: 0.0357 - val_mean_squared_error: 0.0357\n",
      "Epoch 88/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0025 - mean_squared_error: 0.0025\n",
      "Epoch 00088: val_loss improved from 0.03568 to 0.03481, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 949us/sample - loss: 0.0026 - mean_squared_error: 0.0026 - val_loss: 0.0348 - val_mean_squared_error: 0.0348\n",
      "Epoch 89/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0025 - mean_squared_error: 0.0025\n",
      "Epoch 00089: val_loss improved from 0.03481 to 0.03384, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 949us/sample - loss: 0.0025 - mean_squared_error: 0.0025 - val_loss: 0.0338 - val_mean_squared_error: 0.0338\n",
      "Epoch 90/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0024 - mean_squared_error: 0.0024\n",
      "Epoch 00090: val_loss improved from 0.03384 to 0.03366, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 942us/sample - loss: 0.0024 - mean_squared_error: 0.0024 - val_loss: 0.0337 - val_mean_squared_error: 0.0337\n",
      "Epoch 91/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0025 - mean_squared_error: 0.0025\n",
      "Epoch 00091: val_loss improved from 0.03366 to 0.03238, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 926us/sample - loss: 0.0025 - mean_squared_error: 0.0025 - val_loss: 0.0324 - val_mean_squared_error: 0.0324\n",
      "Epoch 92/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0025 - mean_squared_error: 0.0025\n",
      "Epoch 00092: val_loss improved from 0.03238 to 0.03121, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 950us/sample - loss: 0.0025 - mean_squared_error: 0.0025 - val_loss: 0.0312 - val_mean_squared_error: 0.0312\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 93/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0024 - mean_squared_error: 0.0024\n",
      "Epoch 00093: val_loss improved from 0.03121 to 0.02979, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 956us/sample - loss: 0.0025 - mean_squared_error: 0.0025 - val_loss: 0.0298 - val_mean_squared_error: 0.0298\n",
      "Epoch 94/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0025 - mean_squared_error: 0.0025\n",
      "Epoch 00094: val_loss improved from 0.02979 to 0.02854, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 938us/sample - loss: 0.0025 - mean_squared_error: 0.0025 - val_loss: 0.0285 - val_mean_squared_error: 0.0285\n",
      "Epoch 95/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0025 - mean_squared_error: 0.0025\n",
      "Epoch 00095: val_loss improved from 0.02854 to 0.02766, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 932us/sample - loss: 0.0025 - mean_squared_error: 0.0025 - val_loss: 0.0277 - val_mean_squared_error: 0.0277\n",
      "Epoch 96/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0026 - mean_squared_error: 0.0026\n",
      "Epoch 00096: val_loss improved from 0.02766 to 0.02757, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 933us/sample - loss: 0.0025 - mean_squared_error: 0.0025 - val_loss: 0.0276 - val_mean_squared_error: 0.0276\n",
      "Epoch 97/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0025 - mean_squared_error: 0.0025\n",
      "Epoch 00097: val_loss improved from 0.02757 to 0.02654, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 936us/sample - loss: 0.0025 - mean_squared_error: 0.0025 - val_loss: 0.0265 - val_mean_squared_error: 0.0265\n",
      "Epoch 98/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0024 - mean_squared_error: 0.0024\n",
      "Epoch 00098: val_loss improved from 0.02654 to 0.02642, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 932us/sample - loss: 0.0024 - mean_squared_error: 0.0024 - val_loss: 0.0264 - val_mean_squared_error: 0.0264\n",
      "Epoch 99/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0025 - mean_squared_error: 0.0025\n",
      "Epoch 00099: val_loss improved from 0.02642 to 0.02469, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 946us/sample - loss: 0.0025 - mean_squared_error: 0.0025 - val_loss: 0.0247 - val_mean_squared_error: 0.0247\n",
      "Epoch 100/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0026 - mean_squared_error: 0.0026\n",
      "Epoch 00100: val_loss improved from 0.02469 to 0.02415, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 937us/sample - loss: 0.0025 - mean_squared_error: 0.0025 - val_loss: 0.0242 - val_mean_squared_error: 0.0242\n",
      "Epoch 101/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0025 - mean_squared_error: 0.0025\n",
      "Epoch 00101: val_loss improved from 0.02415 to 0.02324, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 927us/sample - loss: 0.0025 - mean_squared_error: 0.0025 - val_loss: 0.0232 - val_mean_squared_error: 0.0232\n",
      "Epoch 102/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0025 - mean_squared_error: 0.0025\n",
      "Epoch 00102: val_loss improved from 0.02324 to 0.02222, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 940us/sample - loss: 0.0025 - mean_squared_error: 0.0025 - val_loss: 0.0222 - val_mean_squared_error: 0.0222\n",
      "Epoch 103/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0024 - mean_squared_error: 0.0024\n",
      "Epoch 00103: val_loss did not improve from 0.02222\n",
      "5513/5513 [==============================] - 5s 931us/sample - loss: 0.0024 - mean_squared_error: 0.0024 - val_loss: 0.0224 - val_mean_squared_error: 0.0224\n",
      "Epoch 104/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0026 - mean_squared_error: 0.0026\n",
      "Epoch 00104: val_loss improved from 0.02222 to 0.01973, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 930us/sample - loss: 0.0026 - mean_squared_error: 0.0026 - val_loss: 0.0197 - val_mean_squared_error: 0.0197\n",
      "Epoch 105/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0025 - mean_squared_error: 0.0025\n",
      "Epoch 00105: val_loss did not improve from 0.01973\n",
      "5513/5513 [==============================] - 5s 945us/sample - loss: 0.0025 - mean_squared_error: 0.0025 - val_loss: 0.0216 - val_mean_squared_error: 0.0216\n",
      "Epoch 106/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0025 - mean_squared_error: 0.0025\n",
      "Epoch 00106: val_loss did not improve from 0.01973\n",
      "5513/5513 [==============================] - 5s 944us/sample - loss: 0.0025 - mean_squared_error: 0.0025 - val_loss: 0.0212 - val_mean_squared_error: 0.0212\n",
      "Epoch 107/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0024 - mean_squared_error: 0.0024\n",
      "Epoch 00107: val_loss did not improve from 0.01973\n",
      "5513/5513 [==============================] - 5s 933us/sample - loss: 0.0024 - mean_squared_error: 0.0024 - val_loss: 0.0202 - val_mean_squared_error: 0.0202\n",
      "Epoch 108/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0025 - mean_squared_error: 0.0025\n",
      "Epoch 00108: val_loss improved from 0.01973 to 0.01943, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 938us/sample - loss: 0.0025 - mean_squared_error: 0.0025 - val_loss: 0.0194 - val_mean_squared_error: 0.0194\n",
      "Epoch 109/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0024 - mean_squared_error: 0.0024\n",
      "Epoch 00109: val_loss improved from 0.01943 to 0.01844, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 944us/sample - loss: 0.0024 - mean_squared_error: 0.0024 - val_loss: 0.0184 - val_mean_squared_error: 0.0184\n",
      "Epoch 110/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0024 - mean_squared_error: 0.0024\n",
      "Epoch 00110: val_loss improved from 0.01844 to 0.01645, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 931us/sample - loss: 0.0024 - mean_squared_error: 0.0024 - val_loss: 0.0165 - val_mean_squared_error: 0.0165\n",
      "Epoch 111/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0024 - mean_squared_error: 0.0024\n",
      "Epoch 00111: val_loss improved from 0.01645 to 0.01583, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 923us/sample - loss: 0.0024 - mean_squared_error: 0.0024 - val_loss: 0.0158 - val_mean_squared_error: 0.0158\n",
      "Epoch 112/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0024 - mean_squared_error: 0.0024\n",
      "Epoch 00112: val_loss did not improve from 0.01583\n",
      "5513/5513 [==============================] - 5s 937us/sample - loss: 0.0024 - mean_squared_error: 0.0024 - val_loss: 0.0165 - val_mean_squared_error: 0.0165\n",
      "Epoch 113/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0024 - mean_squared_error: 0.0024\n",
      "Epoch 00113: val_loss improved from 0.01583 to 0.01493, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 6s 1ms/sample - loss: 0.0025 - mean_squared_error: 0.0025 - val_loss: 0.0149 - val_mean_squared_error: 0.0149\n",
      "Epoch 114/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00114: val_loss improved from 0.01493 to 0.01443, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 978us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0144 - val_mean_squared_error: 0.0144\n",
      "Epoch 115/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0024 - mean_squared_error: 0.0024\n",
      "Epoch 00115: val_loss improved from 0.01443 to 0.01433, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 6s 1ms/sample - loss: 0.0024 - mean_squared_error: 0.0024 - val_loss: 0.0143 - val_mean_squared_error: 0.0143\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 116/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0024 - mean_squared_error: 0.0024\n",
      "Epoch 00116: val_loss improved from 0.01433 to 0.01403, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 975us/sample - loss: 0.0024 - mean_squared_error: 0.0024 - val_loss: 0.0140 - val_mean_squared_error: 0.0140\n",
      "Epoch 117/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00117: val_loss improved from 0.01403 to 0.01273, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 921us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0127 - val_mean_squared_error: 0.0127\n",
      "Epoch 118/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0024 - mean_squared_error: 0.0024\n",
      "Epoch 00118: val_loss improved from 0.01273 to 0.01248, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 923us/sample - loss: 0.0024 - mean_squared_error: 0.0024 - val_loss: 0.0125 - val_mean_squared_error: 0.0125\n",
      "Epoch 119/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00119: val_loss improved from 0.01248 to 0.01154, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 928us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0115 - val_mean_squared_error: 0.0115\n",
      "Epoch 120/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00120: val_loss improved from 0.01154 to 0.01076, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 932us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0108 - val_mean_squared_error: 0.0108\n",
      "Epoch 121/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0024 - mean_squared_error: 0.0024\n",
      "Epoch 00121: val_loss improved from 0.01076 to 0.00995, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 924us/sample - loss: 0.0024 - mean_squared_error: 0.0024 - val_loss: 0.0100 - val_mean_squared_error: 0.0100\n",
      "Epoch 122/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00122: val_loss did not improve from 0.00995\n",
      "5513/5513 [==============================] - 5s 914us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0104 - val_mean_squared_error: 0.0104\n",
      "Epoch 123/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0024 - mean_squared_error: 0.0024\n",
      "Epoch 00123: val_loss improved from 0.00995 to 0.00943, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 925us/sample - loss: 0.0024 - mean_squared_error: 0.0024 - val_loss: 0.0094 - val_mean_squared_error: 0.0094\n",
      "Epoch 124/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0024 - mean_squared_error: 0.0024\n",
      "Epoch 00124: val_loss improved from 0.00943 to 0.00802, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 927us/sample - loss: 0.0024 - mean_squared_error: 0.0024 - val_loss: 0.0080 - val_mean_squared_error: 0.0080\n",
      "Epoch 125/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0024 - mean_squared_error: 0.0024\n",
      "Epoch 00125: val_loss did not improve from 0.00802\n",
      "5513/5513 [==============================] - 5s 907us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0088 - val_mean_squared_error: 0.0088\n",
      "Epoch 126/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0024 - mean_squared_error: 0.0024\n",
      "Epoch 00126: val_loss improved from 0.00802 to 0.00786, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 944us/sample - loss: 0.0024 - mean_squared_error: 0.0024 - val_loss: 0.0079 - val_mean_squared_error: 0.0079\n",
      "Epoch 127/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00127: val_loss improved from 0.00786 to 0.00756, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 923us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0076 - val_mean_squared_error: 0.0076\n",
      "Epoch 128/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00128: val_loss did not improve from 0.00756\n",
      "5513/5513 [==============================] - 5s 920us/sample - loss: 0.0024 - mean_squared_error: 0.0024 - val_loss: 0.0080 - val_mean_squared_error: 0.0080\n",
      "Epoch 129/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00129: val_loss improved from 0.00756 to 0.00672, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 920us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0067 - val_mean_squared_error: 0.0067\n",
      "Epoch 130/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0024 - mean_squared_error: 0.0024\n",
      "Epoch 00130: val_loss improved from 0.00672 to 0.00621, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 937us/sample - loss: 0.0024 - mean_squared_error: 0.0024 - val_loss: 0.0062 - val_mean_squared_error: 0.0062\n",
      "Epoch 131/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0024 - mean_squared_error: 0.0024\n",
      "Epoch 00131: val_loss did not improve from 0.00621\n",
      "5513/5513 [==============================] - 5s 918us/sample - loss: 0.0024 - mean_squared_error: 0.0024 - val_loss: 0.0068 - val_mean_squared_error: 0.0068\n",
      "Epoch 132/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0024 - mean_squared_error: 0.0024\n",
      "Epoch 00132: val_loss did not improve from 0.00621\n",
      "5513/5513 [==============================] - 5s 959us/sample - loss: 0.0024 - mean_squared_error: 0.0024 - val_loss: 0.0063 - val_mean_squared_error: 0.0063\n",
      "Epoch 133/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00133: val_loss did not improve from 0.00621\n",
      "5513/5513 [==============================] - 5s 927us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0062 - val_mean_squared_error: 0.0062\n",
      "Epoch 134/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00134: val_loss improved from 0.00621 to 0.00581, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 945us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0058 - val_mean_squared_error: 0.0058\n",
      "Epoch 135/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00135: val_loss improved from 0.00581 to 0.00551, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 941us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0055 - val_mean_squared_error: 0.0055\n",
      "Epoch 136/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00136: val_loss did not improve from 0.00551\n",
      "5513/5513 [==============================] - 5s 974us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0060 - val_mean_squared_error: 0.0060\n",
      "Epoch 137/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00137: val_loss improved from 0.00551 to 0.00503, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 962us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0050 - val_mean_squared_error: 0.0050\n",
      "Epoch 138/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0024 - mean_squared_error: 0.0024\n",
      "Epoch 00138: val_loss did not improve from 0.00503\n",
      "5513/5513 [==============================] - 5s 945us/sample - loss: 0.0024 - mean_squared_error: 0.0024 - val_loss: 0.0057 - val_mean_squared_error: 0.0057\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 139/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00139: val_loss did not improve from 0.00503\n",
      "5513/5513 [==============================] - 5s 937us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0056 - val_mean_squared_error: 0.0056\n",
      "Epoch 140/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00140: val_loss improved from 0.00503 to 0.00473, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 958us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0047 - val_mean_squared_error: 0.0047\n",
      "Epoch 141/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00141: val_loss did not improve from 0.00473\n",
      "5513/5513 [==============================] - 5s 941us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0053 - val_mean_squared_error: 0.0053\n",
      "Epoch 142/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00142: val_loss did not improve from 0.00473\n",
      "5513/5513 [==============================] - 5s 933us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0048 - val_mean_squared_error: 0.0048\n",
      "Epoch 143/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00143: val_loss did not improve from 0.00473\n",
      "5513/5513 [==============================] - 5s 929us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0050 - val_mean_squared_error: 0.0050\n",
      "Epoch 144/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00144: val_loss improved from 0.00473 to 0.00399, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 958us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0040 - val_mean_squared_error: 0.0040\n",
      "Epoch 145/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00145: val_loss did not improve from 0.00399\n",
      "5513/5513 [==============================] - 5s 965us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0045 - val_mean_squared_error: 0.0045\n",
      "Epoch 146/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00146: val_loss did not improve from 0.00399\n",
      "5513/5513 [==============================] - 5s 944us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0044 - val_mean_squared_error: 0.0044\n",
      "Epoch 147/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00147: val_loss did not improve from 0.00399\n",
      "5513/5513 [==============================] - 5s 959us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0040 - val_mean_squared_error: 0.0040\n",
      "Epoch 148/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00148: val_loss improved from 0.00399 to 0.00356, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 940us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0036 - val_mean_squared_error: 0.0036\n",
      "Epoch 149/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00149: val_loss did not improve from 0.00356\n",
      "5513/5513 [==============================] - 5s 946us/sample - loss: 0.0024 - mean_squared_error: 0.0024 - val_loss: 0.0037 - val_mean_squared_error: 0.0037\n",
      "Epoch 150/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00150: val_loss did not improve from 0.00356\n",
      "5513/5513 [==============================] - 5s 934us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0040 - val_mean_squared_error: 0.0040\n",
      "Epoch 151/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00151: val_loss did not improve from 0.00356\n",
      "5513/5513 [==============================] - 5s 958us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0041 - val_mean_squared_error: 0.0041\n",
      "Epoch 152/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00152: val_loss did not improve from 0.00356\n",
      "5513/5513 [==============================] - 5s 940us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0036 - val_mean_squared_error: 0.0036\n",
      "Epoch 153/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00153: val_loss did not improve from 0.00356\n",
      "5513/5513 [==============================] - 5s 950us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0036 - val_mean_squared_error: 0.0036\n",
      "Epoch 154/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00154: val_loss improved from 0.00356 to 0.00349, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 959us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0035 - val_mean_squared_error: 0.0035\n",
      "Epoch 155/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00155: val_loss did not improve from 0.00349\n",
      "5513/5513 [==============================] - 5s 957us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0037 - val_mean_squared_error: 0.0037\n",
      "Epoch 156/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00156: val_loss did not improve from 0.00349\n",
      "5513/5513 [==============================] - 5s 942us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0038 - val_mean_squared_error: 0.0038\n",
      "Epoch 157/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00157: val_loss improved from 0.00349 to 0.00318, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 952us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0032 - val_mean_squared_error: 0.0032\n",
      "Epoch 158/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00158: val_loss did not improve from 0.00318\n",
      "5513/5513 [==============================] - 5s 931us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0033 - val_mean_squared_error: 0.0033\n",
      "Epoch 159/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00159: val_loss did not improve from 0.00318\n",
      "5513/5513 [==============================] - 5s 923us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0033 - val_mean_squared_error: 0.0033\n",
      "Epoch 160/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00160: val_loss improved from 0.00318 to 0.00293, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 946us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0029 - val_mean_squared_error: 0.0029\n",
      "Epoch 161/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00161: val_loss did not improve from 0.00293\n",
      "5513/5513 [==============================] - 5s 939us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0032 - val_mean_squared_error: 0.0032\n",
      "Epoch 162/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00162: val_loss did not improve from 0.00293\n",
      "5513/5513 [==============================] - 5s 942us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0032 - val_mean_squared_error: 0.0032\n",
      "Epoch 163/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00163: val_loss improved from 0.00293 to 0.00286, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 954us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0029 - val_mean_squared_error: 0.0029\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 164/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00164: val_loss improved from 0.00286 to 0.00269, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 947us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0027 - val_mean_squared_error: 0.0027\n",
      "Epoch 165/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00165: val_loss improved from 0.00269 to 0.00255, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 943us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0025 - val_mean_squared_error: 0.0025\n",
      "Epoch 166/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00166: val_loss improved from 0.00255 to 0.00254, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 934us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0025 - val_mean_squared_error: 0.0025\n",
      "Epoch 167/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00167: val_loss did not improve from 0.00254\n",
      "5513/5513 [==============================] - 5s 914us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0025 - val_mean_squared_error: 0.0025\n",
      "Epoch 168/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00168: val_loss did not improve from 0.00254\n",
      "5513/5513 [==============================] - 5s 933us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0026 - val_mean_squared_error: 0.0026\n",
      "Epoch 169/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00169: val_loss did not improve from 0.00254\n",
      "5513/5513 [==============================] - 5s 947us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0027 - val_mean_squared_error: 0.0027\n",
      "Epoch 170/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00170: val_loss did not improve from 0.00254\n",
      "5513/5513 [==============================] - 5s 939us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0026 - val_mean_squared_error: 0.0026\n",
      "Epoch 171/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00171: val_loss did not improve from 0.00254\n",
      "5513/5513 [==============================] - 5s 931us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0027 - val_mean_squared_error: 0.0027\n",
      "Epoch 172/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00172: val_loss did not improve from 0.00254\n",
      "5513/5513 [==============================] - 5s 940us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0025 - val_mean_squared_error: 0.0025\n",
      "Epoch 173/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00173: val_loss did not improve from 0.00254\n",
      "5513/5513 [==============================] - 5s 932us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0026 - val_mean_squared_error: 0.0026\n",
      "Epoch 174/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00174: val_loss improved from 0.00254 to 0.00217, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 932us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0022 - val_mean_squared_error: 0.0022\n",
      "Epoch 175/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00175: val_loss improved from 0.00217 to 0.00211, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 934us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0021 - val_mean_squared_error: 0.0021\n",
      "Epoch 176/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00176: val_loss did not improve from 0.00211\n",
      "5513/5513 [==============================] - 5s 936us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0025 - val_mean_squared_error: 0.0025\n",
      "Epoch 177/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00177: val_loss did not improve from 0.00211\n",
      "5513/5513 [==============================] - 5s 936us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0023 - val_mean_squared_error: 0.0023\n",
      "Epoch 178/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00178: val_loss did not improve from 0.00211\n",
      "5513/5513 [==============================] - 5s 927us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0021 - val_mean_squared_error: 0.0021\n",
      "Epoch 179/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00179: val_loss did not improve from 0.00211\n",
      "5513/5513 [==============================] - 5s 929us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0021 - val_mean_squared_error: 0.0021\n",
      "Epoch 180/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00180: val_loss did not improve from 0.00211\n",
      "5513/5513 [==============================] - 5s 924us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0023 - val_mean_squared_error: 0.0023\n",
      "Epoch 181/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00181: val_loss did not improve from 0.00211\n",
      "5513/5513 [==============================] - 5s 937us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0024 - val_mean_squared_error: 0.0024\n",
      "Epoch 182/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00182: val_loss did not improve from 0.00211\n",
      "5513/5513 [==============================] - 5s 939us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0023 - val_mean_squared_error: 0.0023\n",
      "Epoch 183/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00183: val_loss improved from 0.00211 to 0.00195, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 939us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0020 - val_mean_squared_error: 0.0020\n",
      "Epoch 184/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00184: val_loss improved from 0.00195 to 0.00173, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 950us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0017 - val_mean_squared_error: 0.0017\n",
      "Epoch 185/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00185: val_loss improved from 0.00173 to 0.00172, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 938us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0017 - val_mean_squared_error: 0.0017\n",
      "Epoch 186/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00186: val_loss did not improve from 0.00172\n",
      "5513/5513 [==============================] - 5s 933us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0017 - val_mean_squared_error: 0.0017\n",
      "Epoch 187/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00187: val_loss did not improve from 0.00172\n",
      "5513/5513 [==============================] - 5s 938us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0017 - val_mean_squared_error: 0.0017\n",
      "Epoch 188/300\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00188: val_loss did not improve from 0.00172\n",
      "5513/5513 [==============================] - 5s 932us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0018 - val_mean_squared_error: 0.0018\n",
      "Epoch 189/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00189: val_loss improved from 0.00172 to 0.00167, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 940us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0017 - val_mean_squared_error: 0.0017\n",
      "Epoch 190/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00190: val_loss did not improve from 0.00167\n",
      "5513/5513 [==============================] - 5s 922us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0017 - val_mean_squared_error: 0.0017\n",
      "Epoch 191/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00191: val_loss did not improve from 0.00167\n",
      "5513/5513 [==============================] - 5s 942us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0018 - val_mean_squared_error: 0.0018\n",
      "Epoch 192/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00192: val_loss did not improve from 0.00167\n",
      "5513/5513 [==============================] - 5s 936us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0017 - val_mean_squared_error: 0.0017\n",
      "Epoch 193/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00193: val_loss improved from 0.00167 to 0.00160, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 941us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0016 - val_mean_squared_error: 0.0016\n",
      "Epoch 194/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00194: val_loss did not improve from 0.00160\n",
      "5513/5513 [==============================] - 5s 930us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0016 - val_mean_squared_error: 0.0016\n",
      "Epoch 195/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00195: val_loss did not improve from 0.00160\n",
      "5513/5513 [==============================] - 5s 942us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0018 - val_mean_squared_error: 0.0018\n",
      "Epoch 196/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00196: val_loss did not improve from 0.00160\n",
      "5513/5513 [==============================] - 5s 944us/sample - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0018 - val_mean_squared_error: 0.0018\n",
      "Epoch 197/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00197: val_loss did not improve from 0.00160\n",
      "5513/5513 [==============================] - 5s 924us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0017 - val_mean_squared_error: 0.0017\n",
      "Epoch 198/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00198: val_loss did not improve from 0.00160\n",
      "5513/5513 [==============================] - 5s 940us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0017 - val_mean_squared_error: 0.0017\n",
      "Epoch 199/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00199: val_loss did not improve from 0.00160\n",
      "5513/5513 [==============================] - 5s 934us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0019 - val_mean_squared_error: 0.0019\n",
      "Epoch 200/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00200: val_loss did not improve from 0.00160\n",
      "5513/5513 [==============================] - 5s 952us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0020 - val_mean_squared_error: 0.0020\n",
      "Epoch 201/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00201: val_loss did not improve from 0.00160\n",
      "5513/5513 [==============================] - 5s 954us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0019 - val_mean_squared_error: 0.0019\n",
      "Epoch 202/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00202: val_loss did not improve from 0.00160\n",
      "5513/5513 [==============================] - 5s 960us/sample - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0019 - val_mean_squared_error: 0.0019\n",
      "Epoch 203/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00203: val_loss did not improve from 0.00160\n",
      "5513/5513 [==============================] - 5s 942us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0018 - val_mean_squared_error: 0.0018\n",
      "Epoch 204/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00204: val_loss did not improve from 0.00160\n",
      "5513/5513 [==============================] - 5s 945us/sample - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0019 - val_mean_squared_error: 0.0019\n",
      "Epoch 205/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00205: val_loss did not improve from 0.00160\n",
      "5513/5513 [==============================] - 5s 929us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0018 - val_mean_squared_error: 0.0018\n",
      "Epoch 206/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00206: val_loss did not improve from 0.00160\n",
      "5513/5513 [==============================] - 5s 933us/sample - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0018 - val_mean_squared_error: 0.0018\n",
      "Epoch 207/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00207: val_loss did not improve from 0.00160\n",
      "5513/5513 [==============================] - 5s 947us/sample - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0018 - val_mean_squared_error: 0.0018\n",
      "Epoch 208/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00208: val_loss did not improve from 0.00160\n",
      "5513/5513 [==============================] - 5s 960us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0017 - val_mean_squared_error: 0.0017\n",
      "Epoch 209/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00209: val_loss did not improve from 0.00160\n",
      "5513/5513 [==============================] - 6s 998us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0017 - val_mean_squared_error: 0.0017\n",
      "Epoch 210/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00210: val_loss did not improve from 0.00160\n",
      "5513/5513 [==============================] - 5s 925us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0017 - val_mean_squared_error: 0.0017\n",
      "Epoch 211/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00211: val_loss did not improve from 0.00160\n",
      "5513/5513 [==============================] - 5s 926us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0017 - val_mean_squared_error: 0.0017\n",
      "Epoch 212/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00212: val_loss did not improve from 0.00160\n",
      "5513/5513 [==============================] - 5s 952us/sample - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0017 - val_mean_squared_error: 0.0017\n",
      "Epoch 213/300\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00213: val_loss did not improve from 0.00160\n",
      "5513/5513 [==============================] - 5s 984us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0017 - val_mean_squared_error: 0.0017\n",
      "Epoch 214/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00214: val_loss did not improve from 0.00160\n",
      "5513/5513 [==============================] - 6s 1ms/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0016 - val_mean_squared_error: 0.0016\n",
      "Epoch 215/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00215: val_loss improved from 0.00160 to 0.00158, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 938us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0016 - val_mean_squared_error: 0.0016\n",
      "Epoch 216/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00216: val_loss improved from 0.00158 to 0.00158, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 935us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0016 - val_mean_squared_error: 0.0016\n",
      "Epoch 217/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00217: val_loss improved from 0.00158 to 0.00158, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 931us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0016 - val_mean_squared_error: 0.0016\n",
      "Epoch 218/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00218: val_loss improved from 0.00158 to 0.00155, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 939us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 219/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00219: val_loss did not improve from 0.00155\n",
      "5513/5513 [==============================] - 5s 930us/sample - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0016 - val_mean_squared_error: 0.0016\n",
      "Epoch 220/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00220: val_loss did not improve from 0.00155\n",
      "5513/5513 [==============================] - 5s 928us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0016 - val_mean_squared_error: 0.0016\n",
      "Epoch 221/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00221: val_loss did not improve from 0.00155\n",
      "5513/5513 [==============================] - 5s 941us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0016 - val_mean_squared_error: 0.0016\n",
      "Epoch 222/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00222: val_loss did not improve from 0.00155\n",
      "5513/5513 [==============================] - 5s 932us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0016 - val_mean_squared_error: 0.0016\n",
      "Epoch 223/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00223: val_loss did not improve from 0.00155\n",
      "5513/5513 [==============================] - 5s 933us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0016 - val_mean_squared_error: 0.0016\n",
      "Epoch 224/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00224: val_loss did not improve from 0.00155\n",
      "5513/5513 [==============================] - 5s 928us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0016 - val_mean_squared_error: 0.0016\n",
      "Epoch 225/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00225: val_loss did not improve from 0.00155\n",
      "5513/5513 [==============================] - 5s 925us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0016 - val_mean_squared_error: 0.0016\n",
      "Epoch 226/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00226: val_loss did not improve from 0.00155\n",
      "5513/5513 [==============================] - 5s 929us/sample - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0016 - val_mean_squared_error: 0.0016\n",
      "Epoch 227/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00227: val_loss did not improve from 0.00155\n",
      "5513/5513 [==============================] - 5s 920us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0016 - val_mean_squared_error: 0.0016\n",
      "Epoch 228/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00228: val_loss did not improve from 0.00155\n",
      "5513/5513 [==============================] - 5s 927us/sample - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0016 - val_mean_squared_error: 0.0016\n",
      "Epoch 229/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00229: val_loss did not improve from 0.00155\n",
      "5513/5513 [==============================] - 5s 929us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0016 - val_mean_squared_error: 0.0016\n",
      "Epoch 230/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00230: val_loss improved from 0.00155 to 0.00154, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 930us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 231/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00231: val_loss improved from 0.00154 to 0.00152, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 958us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 232/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00232: val_loss improved from 0.00152 to 0.00152, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 984us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 233/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00233: val_loss improved from 0.00152 to 0.00151, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 968us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 234/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00234: val_loss improved from 0.00151 to 0.00150, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 969us/sample - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 235/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00235: val_loss improved from 0.00150 to 0.00147, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 956us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 236/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00236: val_loss improved from 0.00147 to 0.00146, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 963us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 237/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00237: val_loss improved from 0.00146 to 0.00145, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 968us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 238/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00238: val_loss improved from 0.00145 to 0.00144, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 984us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 239/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00239: val_loss did not improve from 0.00144\n",
      "5513/5513 [==============================] - 5s 952us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 240/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00240: val_loss improved from 0.00144 to 0.00144, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 947us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 241/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00241: val_loss did not improve from 0.00144\n",
      "5513/5513 [==============================] - 5s 950us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 242/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00242: val_loss did not improve from 0.00144\n",
      "5513/5513 [==============================] - 5s 936us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 243/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00243: val_loss did not improve from 0.00144\n",
      "5513/5513 [==============================] - 5s 939us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 244/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00244: val_loss did not improve from 0.00144\n",
      "5513/5513 [==============================] - 5s 931us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 245/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00245: val_loss did not improve from 0.00144\n",
      "5513/5513 [==============================] - 5s 944us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 246/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00246: val_loss did not improve from 0.00144\n",
      "5513/5513 [==============================] - 5s 932us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 247/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00247: val_loss did not improve from 0.00144\n",
      "5513/5513 [==============================] - 5s 962us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 248/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00248: val_loss did not improve from 0.00144\n",
      "5513/5513 [==============================] - 5s 918us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 249/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00249: val_loss did not improve from 0.00144\n",
      "5513/5513 [==============================] - 5s 912us/sample - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 250/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00250: val_loss did not improve from 0.00144\n",
      "5513/5513 [==============================] - 5s 910us/sample - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 251/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00251: val_loss did not improve from 0.00144\n",
      "5513/5513 [==============================] - 5s 911us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 252/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00252: val_loss did not improve from 0.00144\n",
      "5513/5513 [==============================] - 5s 909us/sample - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 253/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00253: val_loss did not improve from 0.00144\n",
      "5513/5513 [==============================] - 5s 912us/sample - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 254/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00254: val_loss did not improve from 0.00144\n",
      "5513/5513 [==============================] - 5s 919us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 255/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00255: val_loss did not improve from 0.00144\n",
      "5513/5513 [==============================] - 5s 911us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 256/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00256: val_loss improved from 0.00144 to 0.00143, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 915us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 257/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00257: val_loss improved from 0.00143 to 0.00142, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 921us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 258/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00258: val_loss improved from 0.00142 to 0.00141, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 918us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 259/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00259: val_loss improved from 0.00141 to 0.00140, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 920us/sample - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 260/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00260: val_loss improved from 0.00140 to 0.00140, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 914us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 261/300\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00261: val_loss did not improve from 0.00140\n",
      "5513/5513 [==============================] - 5s 903us/sample - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 262/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00262: val_loss did not improve from 0.00140\n",
      "5513/5513 [==============================] - 5s 918us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 263/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00263: val_loss did not improve from 0.00140\n",
      "5513/5513 [==============================] - 5s 906us/sample - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 264/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00264: val_loss did not improve from 0.00140\n",
      "5513/5513 [==============================] - 5s 967us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 265/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00265: val_loss did not improve from 0.00140\n",
      "5513/5513 [==============================] - 5s 966us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 266/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0023 - mean_squared_error: 0.0023\n",
      "Epoch 00266: val_loss did not improve from 0.00140\n",
      "5513/5513 [==============================] - 5s 982us/sample - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 267/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00267: val_loss did not improve from 0.00140\n",
      "5513/5513 [==============================] - 5s 981us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 268/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00268: val_loss did not improve from 0.00140\n",
      "5513/5513 [==============================] - 5s 957us/sample - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 269/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00269: val_loss did not improve from 0.00140\n",
      "5513/5513 [==============================] - 5s 978us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 270/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00270: val_loss did not improve from 0.00140\n",
      "5513/5513 [==============================] - 5s 985us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 271/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00271: val_loss did not improve from 0.00140\n",
      "5513/5513 [==============================] - 5s 942us/sample - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 272/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00272: val_loss did not improve from 0.00140\n",
      "5513/5513 [==============================] - 5s 945us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 273/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00273: val_loss did not improve from 0.00140\n",
      "5513/5513 [==============================] - 5s 944us/sample - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 274/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00274: val_loss did not improve from 0.00140\n",
      "5513/5513 [==============================] - 5s 950us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 275/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00275: val_loss did not improve from 0.00140\n",
      "5513/5513 [==============================] - 5s 976us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 276/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00276: val_loss did not improve from 0.00140\n",
      "5513/5513 [==============================] - 5s 943us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 277/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00277: val_loss did not improve from 0.00140\n",
      "5513/5513 [==============================] - 5s 961us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 278/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00278: val_loss did not improve from 0.00140\n",
      "5513/5513 [==============================] - 5s 977us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 279/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00279: val_loss improved from 0.00140 to 0.00140, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 950us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 280/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00280: val_loss improved from 0.00140 to 0.00139, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 937us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 281/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00281: val_loss improved from 0.00139 to 0.00139, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 931us/sample - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 282/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00282: val_loss improved from 0.00139 to 0.00138, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 949us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 283/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00283: val_loss improved from 0.00138 to 0.00138, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 966us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 284/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00284: val_loss did not improve from 0.00138\n",
      "5513/5513 [==============================] - 5s 953us/sample - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 285/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00285: val_loss did not improve from 0.00138\n",
      "5513/5513 [==============================] - 5s 982us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 286/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00286: val_loss did not improve from 0.00138\n",
      "5513/5513 [==============================] - 5s 937us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 287/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00287: val_loss improved from 0.00138 to 0.00138, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 974us/sample - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 288/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00288: val_loss improved from 0.00138 to 0.00138, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 933us/sample - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 289/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00289: val_loss improved from 0.00138 to 0.00138, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 939us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 290/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00290: val_loss improved from 0.00138 to 0.00138, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 933us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 291/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00291: val_loss improved from 0.00138 to 0.00138, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 925us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 292/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00292: val_loss improved from 0.00138 to 0.00137, saving model to saved_model/model_djia_support_nikkei.h5\n",
      "5513/5513 [==============================] - 5s 952us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 293/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00293: val_loss did not improve from 0.00137\n",
      "5513/5513 [==============================] - 5s 941us/sample - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 294/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00294: val_loss did not improve from 0.00137\n",
      "5513/5513 [==============================] - 5s 932us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 295/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00295: val_loss did not improve from 0.00137\n",
      "5513/5513 [==============================] - 5s 934us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 296/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00296: val_loss did not improve from 0.00137\n",
      "5513/5513 [==============================] - 5s 926us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 297/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00297: val_loss did not improve from 0.00137\n",
      "5513/5513 [==============================] - 5s 947us/sample - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 298/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0021 - mean_squared_error: 0.0021\n",
      "Epoch 00298: val_loss did not improve from 0.00137\n",
      "5513/5513 [==============================] - 5s 958us/sample - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 299/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00299: val_loss did not improve from 0.00137\n",
      "5513/5513 [==============================] - 5s 953us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 300/300\n",
      "5120/5513 [==========================>...] - ETA: 0s - loss: 0.0022 - mean_squared_error: 0.0022\n",
      "Epoch 00300: val_loss did not improve from 0.00137\n",
      "5513/5513 [==============================] - 5s 940us/sample - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(x_train, y_train, epochs=epochs , callbacks=callbacks_list_model,\n",
    "          validation_data=(x_val, y_val),batch_size=batch_size) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "dict_keys(['loss', 'mean_squared_error', 'val_loss', 'val_mean_squared_error', 'lr'])\n"
     ]
    }
   ],
   "source": [
    "print(history.history.keys())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.title(\"Loss (MSE)\")\n",
    "plt.plot(history.history['loss'],color = 'blue',label=\"train_loss\")\n",
    "plt.plot(history.history['val_loss'],color = 'red',label=\"val_loss\")\n",
    "plt.xlabel(\"epochs\")\n",
    "plt.ylabel(\"loss\")\n",
    "plt.legend()\n",
    "plt.savefig('loss_support.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train_pred = model.predict(x_train)\n",
    "y_val_pred = model.predict(x_val)\n",
    "y_test_pred = model.predict(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "#fix!\n",
    "def plot_compare(x, event_col):\n",
    "#     Col = {'Open':0, 'High':1, 'Low':2, 'Close':3}\n",
    "    Col = dict()\n",
    "    count = 0\n",
    "    for i, col in enumerate(event_col):\n",
    "        Col[col] = i\n",
    "    ft = Col[x]\n",
    "    plt.figure(figsize=(15, 5))\n",
    "    plt.subplot(1, 2, 1)\n",
    "\n",
    "    plt.plot(np.arange(y_train.shape[0]), y_train[:, ft], color='blue', label='train target')\n",
    "\n",
    "    plt.plot(np.arange(y_train.shape[0], y_train.shape[0] + y_val.shape[0]), y_val[:, ft], \n",
    "             color='gray', label='val target')\n",
    "    plt.plot(np.arange(y_train.shape[0] + y_val.shape[0], y_train.shape[0] + y_val.shape[0] + y_test.shape[0]),\n",
    "            y_test[:, ft], color='black', label='test target')\n",
    "    plt.plot(np.arange(y_train_pred.shape[0]), y_train_pred[:, ft],\n",
    "            color='red', label='train prediction') \n",
    "    plt.plot(np.arange(y_train_pred.shape[0], y_train_pred.shape[0] + y_val_pred.shape[0]),\n",
    "            y_val_pred[:, ft], color='orange', label='val prediction')\n",
    "    plt.plot(np.arange(y_train_pred.shape[0] + y_val_pred.shape[0],\n",
    "            y_train_pred.shape[0] + y_val_pred.shape[0] + y_test_pred.shape[0]),\n",
    "            y_test_pred[:, ft], color ='green', label='test prediction')\n",
    "    plt.title('past and future stock price :' + x)\n",
    "    plt.xlabel('Time [Days]')\n",
    "    plt.ylabel('Normalized price')\n",
    "    plt.legend(loc='best')\n",
    "\n",
    "    plt.subplot(1, 2, 2)\n",
    "    plt.plot(np.arange(y_train.shape[0], y_train.shape[0] + y_test.shape[0]),\n",
    "            y_test[:, ft], color='black', label='test target')\n",
    "    plt.plot(np.arange(y_train.shape[0], y_train.shape[0] + y_test_pred.shape[0]),\n",
    "            y_test_pred[:, ft], color='green', label='test prediction')\n",
    "    plt.title('future stock prices :' + x)\n",
    "    plt.xlabel('Time [Days]')\n",
    "    plt.ylabel('Normalized price')\n",
    "    plt.legend(loc='best')\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_compare('Close(DJIA)', event_col)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "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": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}