Project-G6.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import tensorflow as tf\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "import os\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from scipy import stats\n",
    "from statsmodels.tsa.stattools import adfuller\n",
    "from sklearn import preprocessing\n",
    "from tensorflow.keras.preprocessing.sequence import TimeseriesGenerator\n",
    "import warnings\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import mean_squared_error\n",
    "from sklearn.metrics import mean_absolute_error\n",
    "import math\n",
    "warnings.filterwarnings('ignore')\n",
    "from statsmodels.tsa.stattools import pacf\n",
    "import datetime\n",
    "%matplotlib inline\n",
    "from tensorflow import keras\n",
    "from tensorflow.keras.optimizers import RMSprop, Adam,SGD"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_series(df,format='-',start = 0, end = None):\n",
    "    plt.plot(df['Umsatz'],format)\n",
    "    plt.xlabel(\"Time\")\n",
    "    plt.ylabel(\"sales\")\n",
    "    plt.grid(True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "dataframe = pd.read_csv('meteolytix_Corona_bakery_turnover_updated.csv')\n",
    "\n",
    "dataframe['Datum']=pd.to_datetime(dataframe['Datum'])\n",
    "\n",
    "dataframe['Quarter'] = dataframe['Datum'].apply(lambda x: x.quarter)\n",
    "dataframe['DayofTheWeek'] = dataframe['Datum'].apply(lambda x: x.weekday()+1)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "cdf = pd.read_csv('covid-data.csv')\n",
    "cdf['date'] = pd.to_datetime(cdf['date'])\n",
    "\n",
    "cdf.rename(columns={'date':'Datum'},inplace=True)\n",
    "\n",
    "cdf = cdf[['Datum','total_cases','new_cases','total_cases_per_million','new_cases_per_million']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataframe = dataframe.join(cdf.set_index('Datum'),on='Datum')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataframe[['total_cases','new_cases','total_cases_per_million','new_cases_per_million']]= dataframe[['total_cases','new_cases','total_cases_per_million','new_cases_per_million']].fillna(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "prodG6 = dataframe[dataframe['Warengruppe'] == 6]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "prodG6.drop('Warengruppe',axis=1,inplace=True) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "prodG6.set_index('Datum',inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x1918dd9aa88>]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1224x1152 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure(figsize=(17,16))\n",
    "fig.subplots_adjust(hspace=.4)\n",
    "ax1 = fig.add_subplot(5,1,1)\n",
    "ax1.plot(prodG6['Umsatz'].rolling(window=1).mean(),linewidth=1)\n",
    "ax1.set_title('Mean Global sales resampled over day')\n",
    "ax1.tick_params(axis='both', which='major')\n",
    "ax1.plot(prodG6['new_cases_per_million'])\n",
    "\n",
    "ax2 = fig.add_subplot(5,1,2, sharex=ax1)\n",
    "ax2.plot(prodG6['Umsatz'].rolling(window=7).mean(),linewidth=1)\n",
    "ax2.set_title('Mean Global Sales resampled over week')\n",
    "ax2.tick_params(axis='both', which='major')\n",
    "ax2.plot(prodG6['new_cases_per_million'].resample('W').mean())\n",
    "\n",
    "ax3 = fig.add_subplot(5,1,3, sharex=ax1)\n",
    "ax3.plot(prodG6['Umsatz'].rolling(window=30).mean(),linewidth=1)\n",
    "ax3.set_title('Mean Global Sales resampled over Month')\n",
    "ax3.tick_params(axis='both', which='major')\n",
    "ax3.plot(prodG6['new_cases_per_million'].resample('M').mean())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "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>Datum</th>\n",
       "      <th>Umsatz</th>\n",
       "      <th>Quarter</th>\n",
       "      <th>DayofTheWeek</th>\n",
       "      <th>total_cases</th>\n",
       "      <th>new_cases</th>\n",
       "      <th>total_cases_per_million</th>\n",
       "      <th>new_cases_per_million</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2015-11-02</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2015-11-03</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2015-11-04</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2015-11-05</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2015-11-06</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       Datum  Umsatz  Quarter  DayofTheWeek  total_cases  new_cases  \\\n",
       "0 2015-11-02     0.0        4             1          0.0        0.0   \n",
       "1 2015-11-03     0.0        4             2          0.0        0.0   \n",
       "2 2015-11-04     0.0        4             3          0.0        0.0   \n",
       "3 2015-11-05     0.0        4             4          0.0        0.0   \n",
       "4 2015-11-06     0.0        4             5          0.0        0.0   \n",
       "\n",
       "   total_cases_per_million  new_cases_per_million  \n",
       "0                      0.0                    0.0  \n",
       "1                      0.0                    0.0  \n",
       "2                      0.0                    0.0  \n",
       "3                      0.0                    0.0  \n",
       "4                      0.0                    0.0  "
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "meanprodG6 = prodG6\n",
    "meanprodG6['Umsatz'] = meanprodG6['Umsatz'].rolling(window=14).mean()\n",
    "meanprodG6['Umsatz'] = meanprodG6['Umsatz'].fillna(0)\n",
    "meanprodG6.reset_index(inplace=True)\n",
    "meanprodG6.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "    dataset = meanprodG6.Umsatz.values \n",
    "    dataset = dataset.astype('float32')\n",
    "    dataset = np.reshape(dataset, (-1, 1))\n",
    "    scaler = MinMaxScaler(feature_range=(0, 1))\n",
    "    dataset = scaler.fit_transform(dataset)\n",
    "    train_size =150\n",
    "    test_size = len(dataset) - train_size\n",
    "    train, test = dataset[0:train_size,:], dataset[train_size:len(dataset),:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "def create_dataset(dataset, window_size):\n",
    "    X, Y = [], []\n",
    "    for i in range(len(dataset)-window_size-1):\n",
    "        a = dataset[i:(i+window_size), 0]\n",
    "        X.append(a)\n",
    "        Y.append(dataset[i + window_size, 0])\n",
    "    return np.array(X), np.array(Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "window_size = 14\n",
    "X_train, Y_train = create_dataset(train, window_size)\n",
    "X_test, Y_test = create_dataset(test, window_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(135, 14)"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(73,)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Y_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = np.reshape(X_train, (X_train.shape[0], 1, X_train.shape[1]))\n",
    "X_test = np.reshape(X_test, (X_test.shape[0], 1, X_test.shape[1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_2\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv1d_2 (Conv1D)            (None, 1, 32)             2272      \n",
      "_________________________________________________________________\n",
      "lstm_4 (LSTM)                (None, 1, 50)             16600     \n",
      "_________________________________________________________________\n",
      "lstm_5 (LSTM)                (None, 50)                20200     \n",
      "_________________________________________________________________\n",
      "dense_4 (Dense)              (None, 128)               6528      \n",
      "_________________________________________________________________\n",
      "dropout_2 (Dropout)          (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "dense_5 (Dense)              (None, 1)                 129       \n",
      "_________________________________________________________________\n",
      "lambda_2 (Lambda)            (None, 1)                 0         \n",
      "=================================================================\n",
      "Total params: 45,729\n",
      "Trainable params: 45,729\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "Train on 135 samples\n",
      "Epoch 1/100\n",
      "135/135 [==============================] - 4s 27ms/sample - loss: 0.2653 - mae: 0.4325\n",
      "Epoch 2/100\n",
      "135/135 [==============================] - 0s 252us/sample - loss: 0.2736 - mae: 0.4523\n",
      "Epoch 3/100\n",
      "135/135 [==============================] - 0s 363us/sample - loss: 0.2529 - mae: 0.4184\n",
      "Epoch 4/100\n",
      "135/135 [==============================] - 0s 304us/sample - loss: 0.2930 - mae: 0.4436\n",
      "Epoch 5/100\n",
      "135/135 [==============================] - 0s 311us/sample - loss: 0.2871 - mae: 0.4395\n",
      "Epoch 6/100\n",
      "135/135 [==============================] - 0s 400us/sample - loss: 0.2888 - mae: 0.4381\n",
      "Epoch 7/100\n",
      "135/135 [==============================] - 0s 355us/sample - loss: 0.2890 - mae: 0.4521\n",
      "Epoch 8/100\n",
      "135/135 [==============================] - 0s 296us/sample - loss: 0.2329 - mae: 0.4066\n",
      "Epoch 9/100\n",
      "135/135 [==============================] - 0s 333us/sample - loss: 0.2596 - mae: 0.4273\n",
      "Epoch 10/100\n",
      "135/135 [==============================] - 0s 304us/sample - loss: 0.2901 - mae: 0.4519\n",
      "Epoch 11/100\n",
      "135/135 [==============================] - 0s 304us/sample - loss: 0.2546 - mae: 0.4152\n",
      "Epoch 12/100\n",
      "135/135 [==============================] - 0s 341us/sample - loss: 0.2594 - mae: 0.4170\n",
      "Epoch 13/100\n",
      "135/135 [==============================] - 0s 274us/sample - loss: 0.2174 - mae: 0.3920\n",
      "Epoch 14/100\n",
      "135/135 [==============================] - 0s 341us/sample - loss: 0.2696 - mae: 0.4268\n",
      "Epoch 15/100\n",
      "135/135 [==============================] - 0s 274us/sample - loss: 0.3005 - mae: 0.4511\n",
      "Epoch 16/100\n",
      "135/135 [==============================] - 0s 252us/sample - loss: 0.2628 - mae: 0.4163\n",
      "Epoch 17/100\n",
      "135/135 [==============================] - 0s 318us/sample - loss: 0.2486 - mae: 0.4203\n",
      "Epoch 18/100\n",
      "135/135 [==============================] - 0s 252us/sample - loss: 0.3258 - mae: 0.4553\n",
      "Epoch 19/100\n",
      "135/135 [==============================] - 0s 267us/sample - loss: 0.2494 - mae: 0.4076\n",
      "Epoch 20/100\n",
      "135/135 [==============================] - 0s 244us/sample - loss: 0.2911 - mae: 0.4321\n",
      "Epoch 21/100\n",
      "135/135 [==============================] - 0s 267us/sample - loss: 0.1890 - mae: 0.3593\n",
      "Epoch 22/100\n",
      "135/135 [==============================] - 0s 259us/sample - loss: 0.2195 - mae: 0.3785\n",
      "Epoch 23/100\n",
      "135/135 [==============================] - 0s 267us/sample - loss: 0.2766 - mae: 0.4275\n",
      "Epoch 24/100\n",
      "135/135 [==============================] - 0s 252us/sample - loss: 0.2445 - mae: 0.3909\n",
      "Epoch 25/100\n",
      "135/135 [==============================] - 0s 274us/sample - loss: 0.2994 - mae: 0.4286\n",
      "Epoch 26/100\n",
      "135/135 [==============================] - 0s 281us/sample - loss: 0.2300 - mae: 0.3815\n",
      "Epoch 27/100\n",
      "135/135 [==============================] - 0s 267us/sample - loss: 0.2907 - mae: 0.4380\n",
      "Epoch 28/100\n",
      "135/135 [==============================] - 0s 252us/sample - loss: 0.2673 - mae: 0.4477\n",
      "Epoch 29/100\n",
      "135/135 [==============================] - 0s 252us/sample - loss: 0.2672 - mae: 0.4232\n",
      "Epoch 30/100\n",
      "135/135 [==============================] - 0s 281us/sample - loss: 0.3054 - mae: 0.4546\n",
      "Epoch 31/100\n",
      "135/135 [==============================] - 0s 252us/sample - loss: 0.2332 - mae: 0.3792\n",
      "Epoch 32/100\n",
      "135/135 [==============================] - 0s 281us/sample - loss: 0.2623 - mae: 0.4121\n",
      "Epoch 33/100\n",
      "135/135 [==============================] - 0s 274us/sample - loss: 0.2004 - mae: 0.3627\n",
      "Epoch 34/100\n",
      "135/135 [==============================] - 0s 252us/sample - loss: 0.2075 - mae: 0.3706\n",
      "Epoch 35/100\n",
      "135/135 [==============================] - 0s 252us/sample - loss: 0.2548 - mae: 0.3999\n",
      "Epoch 36/100\n",
      "135/135 [==============================] - 0s 267us/sample - loss: 0.1838 - mae: 0.3303\n",
      "Epoch 37/100\n",
      "135/135 [==============================] - 0s 244us/sample - loss: 0.1939 - mae: 0.3625\n",
      "Epoch 38/100\n",
      "135/135 [==============================] - 0s 252us/sample - loss: 0.1740 - mae: 0.3393\n",
      "Epoch 39/100\n",
      "135/135 [==============================] - 0s 259us/sample - loss: 0.1887 - mae: 0.3458\n",
      "Epoch 40/100\n",
      "135/135 [==============================] - 0s 237us/sample - loss: 0.1850 - mae: 0.3352\n",
      "Epoch 41/100\n",
      "135/135 [==============================] - 0s 259us/sample - loss: 0.1891 - mae: 0.3292\n",
      "Epoch 42/100\n",
      "135/135 [==============================] - 0s 304us/sample - loss: 0.1594 - mae: 0.3007\n",
      "Epoch 43/100\n",
      "135/135 [==============================] - 0s 281us/sample - loss: 0.1507 - mae: 0.3004\n",
      "Epoch 44/100\n",
      "135/135 [==============================] - 0s 289us/sample - loss: 0.1720 - mae: 0.3231\n",
      "Epoch 45/100\n",
      "135/135 [==============================] - 0s 281us/sample - loss: 0.1686 - mae: 0.3087\n",
      "Epoch 46/100\n",
      "135/135 [==============================] - 0s 274us/sample - loss: 0.1385 - mae: 0.2937\n",
      "Epoch 47/100\n",
      "135/135 [==============================] - 0s 259us/sample - loss: 0.1320 - mae: 0.2831\n",
      "Epoch 48/100\n",
      "135/135 [==============================] - 0s 237us/sample - loss: 0.1258 - mae: 0.2833\n",
      "Epoch 49/100\n",
      "135/135 [==============================] - 0s 244us/sample - loss: 0.0986 - mae: 0.2451\n",
      "Epoch 50/100\n",
      "135/135 [==============================] - 0s 281us/sample - loss: 0.1296 - mae: 0.2903\n",
      "Epoch 51/100\n",
      "135/135 [==============================] - 0s 274us/sample - loss: 0.1449 - mae: 0.2974\n",
      "Epoch 52/100\n",
      "135/135 [==============================] - 0s 244us/sample - loss: 0.1294 - mae: 0.2844\n",
      "Epoch 53/100\n",
      "135/135 [==============================] - 0s 237us/sample - loss: 0.1351 - mae: 0.2877\n",
      "Epoch 54/100\n",
      "135/135 [==============================] - 0s 267us/sample - loss: 0.1408 - mae: 0.2936\n",
      "Epoch 55/100\n",
      "135/135 [==============================] - 0s 274us/sample - loss: 0.1391 - mae: 0.2839\n",
      "Epoch 56/100\n",
      "135/135 [==============================] - 0s 281us/sample - loss: 0.1298 - mae: 0.2872\n",
      "Epoch 57/100\n",
      "135/135 [==============================] - 0s 267us/sample - loss: 0.1079 - mae: 0.2632\n",
      "Epoch 58/100\n",
      "135/135 [==============================] - 0s 281us/sample - loss: 0.1448 - mae: 0.2904\n",
      "Epoch 59/100\n",
      "135/135 [==============================] - 0s 252us/sample - loss: 0.1565 - mae: 0.3133\n",
      "Epoch 60/100\n",
      "135/135 [==============================] - 0s 296us/sample - loss: 0.1462 - mae: 0.2874\n",
      "Epoch 61/100\n",
      "135/135 [==============================] - 0s 252us/sample - loss: 0.1354 - mae: 0.2980\n",
      "Epoch 62/100\n",
      "135/135 [==============================] - 0s 267us/sample - loss: 0.1372 - mae: 0.2936\n",
      "Epoch 63/100\n",
      "135/135 [==============================] - 0s 289us/sample - loss: 0.1007 - mae: 0.2499\n",
      "Epoch 64/100\n",
      "135/135 [==============================] - 0s 385us/sample - loss: 0.1209 - mae: 0.2649\n",
      "Epoch 65/100\n",
      "135/135 [==============================] - 0s 341us/sample - loss: 0.1108 - mae: 0.2568\n",
      "Epoch 66/100\n",
      "135/135 [==============================] - 0s 400us/sample - loss: 0.1271 - mae: 0.2604\n",
      "Epoch 67/100\n",
      "135/135 [==============================] - 0s 355us/sample - loss: 0.1278 - mae: 0.2892\n",
      "Epoch 68/100\n",
      "135/135 [==============================] - 0s 296us/sample - loss: 0.1203 - mae: 0.2695\n",
      "Epoch 69/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "135/135 [==============================] - 0s 333us/sample - loss: 0.1110 - mae: 0.2611\n",
      "Epoch 70/100\n",
      "135/135 [==============================] - 0s 370us/sample - loss: 0.1127 - mae: 0.2657\n",
      "Epoch 71/100\n",
      "135/135 [==============================] - 0s 355us/sample - loss: 0.1041 - mae: 0.2559\n",
      "Epoch 72/100\n",
      "135/135 [==============================] - 0s 355us/sample - loss: 0.0969 - mae: 0.2473\n",
      "Epoch 73/100\n",
      "135/135 [==============================] - 0s 355us/sample - loss: 0.1000 - mae: 0.2400\n",
      "Epoch 74/100\n",
      "135/135 [==============================] - 0s 318us/sample - loss: 0.0929 - mae: 0.2408\n",
      "Epoch 75/100\n",
      "135/135 [==============================] - 0s 341us/sample - loss: 0.0962 - mae: 0.2415\n",
      "Epoch 76/100\n",
      "135/135 [==============================] - 0s 370us/sample - loss: 0.0871 - mae: 0.2359\n",
      "Epoch 77/100\n",
      "135/135 [==============================] - 0s 363us/sample - loss: 0.0814 - mae: 0.2225\n",
      "Epoch 78/100\n",
      "135/135 [==============================] - 0s 333us/sample - loss: 0.0856 - mae: 0.2377\n",
      "Epoch 79/100\n",
      "135/135 [==============================] - 0s 355us/sample - loss: 0.0835 - mae: 0.2206\n",
      "Epoch 80/100\n",
      "135/135 [==============================] - 0s 341us/sample - loss: 0.0530 - mae: 0.1779\n",
      "Epoch 81/100\n",
      "135/135 [==============================] - 0s 318us/sample - loss: 0.0526 - mae: 0.1829\n",
      "Epoch 82/100\n",
      "135/135 [==============================] - 0s 378us/sample - loss: 0.0860 - mae: 0.2280\n",
      "Epoch 83/100\n",
      "135/135 [==============================] - 0s 318us/sample - loss: 0.0713 - mae: 0.2080\n",
      "Epoch 84/100\n",
      "135/135 [==============================] - 0s 363us/sample - loss: 0.0520 - mae: 0.1790\n",
      "Epoch 85/100\n",
      "135/135 [==============================] - 0s 444us/sample - loss: 0.0422 - mae: 0.1568\n",
      "Epoch 86/100\n",
      "135/135 [==============================] - 0s 370us/sample - loss: 0.0483 - mae: 0.1727\n",
      "Epoch 87/100\n",
      "135/135 [==============================] - 0s 333us/sample - loss: 0.0336 - mae: 0.1385\n",
      "Epoch 88/100\n",
      "135/135 [==============================] - 0s 333us/sample - loss: 0.0294 - mae: 0.1349\n",
      "Epoch 89/100\n",
      "135/135 [==============================] - 0s 348us/sample - loss: 0.0305 - mae: 0.1373\n",
      "Epoch 90/100\n",
      "135/135 [==============================] - 0s 333us/sample - loss: 0.0208 - mae: 0.1151\n",
      "Epoch 91/100\n",
      "135/135 [==============================] - 0s 333us/sample - loss: 0.0142 - mae: 0.0950\n",
      "Epoch 92/100\n",
      "135/135 [==============================] - 0s 355us/sample - loss: 0.0119 - mae: 0.0829\n",
      "Epoch 93/100\n",
      "135/135 [==============================] - 0s 370us/sample - loss: 0.0116 - mae: 0.0870\n",
      "Epoch 94/100\n",
      "135/135 [==============================] - 0s 311us/sample - loss: 0.0155 - mae: 0.0971\n",
      "Epoch 95/100\n",
      "135/135 [==============================] - 0s 400us/sample - loss: 0.0160 - mae: 0.1074\n",
      "Epoch 96/100\n",
      "135/135 [==============================] - 0s 281us/sample - loss: 0.0164 - mae: 0.1056\n",
      "Epoch 97/100\n",
      "135/135 [==============================] - 0s 341us/sample - loss: 0.0111 - mae: 0.0823\n",
      "Epoch 98/100\n",
      "135/135 [==============================] - 0s 370us/sample - loss: 0.0122 - mae: 0.0875\n",
      "Epoch 99/100\n",
      "135/135 [==============================] - 0s 333us/sample - loss: 0.0066 - mae: 0.0632\n",
      "Epoch 100/100\n",
      "135/135 [==============================] - 0s 311us/sample - loss: 0.0056 - mae: 0.0592\n"
     ]
    }
   ],
   "source": [
    "model = tf.keras.models.Sequential([\n",
    "  tf.keras.layers.Conv1D(filters=32, \n",
    "                         kernel_size=5,\n",
    "                         strides=1,\n",
    "                         padding=\"causal\",\n",
    "                         activation='relu',\n",
    "                         input_shape=(X_train.shape[1], X_train.shape[2])),\n",
    "  tf.keras.layers.LSTM(50,return_sequences=True),\n",
    "  tf.keras.layers.LSTM(50),\n",
    "  tf.keras.layers.Dense(128,activation='relu'),\n",
    "  tf.keras.layers.Dropout(0.1),\n",
    "  tf.keras.layers.Dense(1),\n",
    "  tf.keras.layers.Lambda(lambda x: x * 400)\n",
    "])\n",
    "\n",
    "model.summary()\n",
    "\n",
    "lr_schedule = tf.keras.callbacks.LearningRateScheduler(\n",
    "    lambda epoch: 1e-8 * 10**(epoch / 20))\n",
    "optimizer = tf.keras.optimizers.Adam(lr=1e-8)\n",
    "model.compile(loss='mse',\n",
    "              optimizer=optimizer,\n",
    "              metrics=[\"mae\"])\n",
    "\n",
    "history = model.fit(X_train, Y_train, epochs=100,callbacks=[lr_schedule])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1e-08, 0.1, 0.0001, 0.02)"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.semilogx(history.history[\"lr\"], history.history[\"loss\"])\n",
    "plt.axis([1e-8, 1e-1, 0.0001, 0.02])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 135 samples, validate on 73 samples\n",
      "Epoch 1/300\n",
      "135/135 - 3s - loss: 5.7364 - mae: 1.9179 - val_loss: 4.3896 - val_mae: 2.0949\n",
      "Epoch 2/300\n",
      "135/135 - 0s - loss: 1.5273 - mae: 1.0729 - val_loss: 1.6492 - val_mae: 1.2812\n",
      "Epoch 3/300\n",
      "135/135 - 0s - loss: 0.6045 - mae: 0.6607 - val_loss: 0.6590 - val_mae: 0.8092\n",
      "Epoch 4/300\n",
      "135/135 - 0s - loss: 0.4423 - mae: 0.6074 - val_loss: 0.0661 - val_mae: 0.2446\n",
      "Epoch 5/300\n",
      "135/135 - 0s - loss: 0.2053 - mae: 0.3980 - val_loss: 0.0112 - val_mae: 0.0857\n",
      "Epoch 6/300\n",
      "135/135 - 0s - loss: 0.0454 - mae: 0.1720 - val_loss: 0.0608 - val_mae: 0.2403\n",
      "Epoch 7/300\n",
      "135/135 - 0s - loss: 0.0434 - mae: 0.1622 - val_loss: 0.0079 - val_mae: 0.0757\n",
      "Epoch 8/300\n",
      "135/135 - 0s - loss: 0.0397 - mae: 0.1605 - val_loss: 0.0027 - val_mae: 0.0381\n",
      "Epoch 9/300\n",
      "135/135 - 0s - loss: 0.0289 - mae: 0.1318 - val_loss: 0.0086 - val_mae: 0.0827\n",
      "Epoch 10/300\n",
      "135/135 - 0s - loss: 0.0165 - mae: 0.0981 - val_loss: 0.0017 - val_mae: 0.0313\n",
      "Epoch 11/300\n",
      "135/135 - 0s - loss: 0.0185 - mae: 0.1117 - val_loss: 0.0013 - val_mae: 0.0311\n",
      "Epoch 12/300\n",
      "135/135 - 0s - loss: 0.0106 - mae: 0.0768 - val_loss: 0.0051 - val_mae: 0.0638\n",
      "Epoch 13/300\n",
      "135/135 - 0s - loss: 0.0109 - mae: 0.0823 - val_loss: 0.0012 - val_mae: 0.0274\n",
      "Epoch 14/300\n",
      "135/135 - 0s - loss: 0.0122 - mae: 0.0860 - val_loss: 0.0017 - val_mae: 0.0354\n",
      "Epoch 15/300\n",
      "135/135 - 0s - loss: 0.0148 - mae: 0.0917 - val_loss: 0.0015 - val_mae: 0.0304\n",
      "Epoch 16/300\n",
      "135/135 - 0s - loss: 0.0117 - mae: 0.0879 - val_loss: 0.0012 - val_mae: 0.0269\n",
      "Epoch 17/300\n",
      "135/135 - 0s - loss: 0.0110 - mae: 0.0828 - val_loss: 0.0032 - val_mae: 0.0486\n",
      "Epoch 18/300\n",
      "135/135 - 0s - loss: 0.0085 - mae: 0.0668 - val_loss: 8.6362e-04 - val_mae: 0.0243\n",
      "Epoch 19/300\n",
      "135/135 - 0s - loss: 0.0094 - mae: 0.0754 - val_loss: 8.4902e-04 - val_mae: 0.0241\n",
      "Epoch 20/300\n",
      "135/135 - 0s - loss: 0.0073 - mae: 0.0694 - val_loss: 0.0011 - val_mae: 0.0267\n",
      "Epoch 21/300\n",
      "135/135 - 0s - loss: 0.0078 - mae: 0.0729 - val_loss: 9.8824e-04 - val_mae: 0.0258\n",
      "Epoch 22/300\n",
      "135/135 - 0s - loss: 0.0110 - mae: 0.0807 - val_loss: 0.0011 - val_mae: 0.0273\n",
      "Epoch 23/300\n",
      "135/135 - 0s - loss: 0.0072 - mae: 0.0707 - val_loss: 7.9669e-04 - val_mae: 0.0235\n",
      "Epoch 24/300\n",
      "135/135 - 0s - loss: 0.0063 - mae: 0.0580 - val_loss: 0.0022 - val_mae: 0.0383\n",
      "Epoch 25/300\n",
      "135/135 - 0s - loss: 0.0059 - mae: 0.0610 - val_loss: 0.0012 - val_mae: 0.0294\n",
      "Epoch 26/300\n",
      "135/135 - 0s - loss: 0.0062 - mae: 0.0629 - val_loss: 0.0017 - val_mae: 0.0326\n",
      "Epoch 27/300\n",
      "135/135 - 0s - loss: 0.0061 - mae: 0.0630 - val_loss: 7.6278e-04 - val_mae: 0.0227\n",
      "Epoch 28/300\n",
      "135/135 - 0s - loss: 0.0044 - mae: 0.0532 - val_loss: 0.0010 - val_mae: 0.0244\n",
      "Epoch 29/300\n",
      "135/135 - 0s - loss: 0.0053 - mae: 0.0551 - val_loss: 7.4342e-04 - val_mae: 0.0228\n",
      "Epoch 30/300\n",
      "135/135 - 0s - loss: 0.0045 - mae: 0.0524 - val_loss: 0.0011 - val_mae: 0.0284\n",
      "Epoch 31/300\n",
      "135/135 - 0s - loss: 0.0056 - mae: 0.0605 - val_loss: 9.9408e-04 - val_mae: 0.0235\n",
      "Epoch 32/300\n",
      "135/135 - 0s - loss: 0.0043 - mae: 0.0493 - val_loss: 7.1514e-04 - val_mae: 0.0222\n",
      "Epoch 33/300\n",
      "135/135 - 0s - loss: 0.0039 - mae: 0.0491 - val_loss: 8.2129e-04 - val_mae: 0.0222\n",
      "Epoch 34/300\n",
      "135/135 - 0s - loss: 0.0055 - mae: 0.0544 - val_loss: 0.0011 - val_mae: 0.0276\n",
      "Epoch 35/300\n",
      "135/135 - 0s - loss: 0.0044 - mae: 0.0535 - val_loss: 0.0014 - val_mae: 0.0297\n",
      "Epoch 36/300\n",
      "135/135 - 0s - loss: 0.0034 - mae: 0.0442 - val_loss: 6.6062e-04 - val_mae: 0.0220\n",
      "Epoch 37/300\n",
      "135/135 - 0s - loss: 0.0039 - mae: 0.0506 - val_loss: 6.3584e-04 - val_mae: 0.0211\n",
      "Epoch 38/300\n",
      "135/135 - 0s - loss: 0.0035 - mae: 0.0435 - val_loss: 7.3359e-04 - val_mae: 0.0216\n",
      "Epoch 39/300\n",
      "135/135 - 0s - loss: 0.0041 - mae: 0.0507 - val_loss: 5.4983e-04 - val_mae: 0.0195\n",
      "Epoch 40/300\n",
      "135/135 - 0s - loss: 0.0028 - mae: 0.0409 - val_loss: 7.2057e-04 - val_mae: 0.0210\n",
      "Epoch 41/300\n",
      "135/135 - 0s - loss: 0.0026 - mae: 0.0406 - val_loss: 5.7177e-04 - val_mae: 0.0201\n",
      "Epoch 42/300\n",
      "135/135 - 0s - loss: 0.0030 - mae: 0.0438 - val_loss: 9.2813e-04 - val_mae: 0.0234\n",
      "Epoch 43/300\n",
      "135/135 - 0s - loss: 0.0033 - mae: 0.0454 - val_loss: 5.6855e-04 - val_mae: 0.0198\n",
      "Epoch 44/300\n",
      "135/135 - 0s - loss: 0.0035 - mae: 0.0477 - val_loss: 5.9801e-04 - val_mae: 0.0191\n",
      "Epoch 45/300\n",
      "135/135 - 0s - loss: 0.0029 - mae: 0.0431 - val_loss: 5.9168e-04 - val_mae: 0.0207\n",
      "Epoch 46/300\n",
      "135/135 - 0s - loss: 0.0030 - mae: 0.0435 - val_loss: 6.0366e-04 - val_mae: 0.0194\n",
      "Epoch 47/300\n",
      "135/135 - 0s - loss: 0.0033 - mae: 0.0446 - val_loss: 5.5098e-04 - val_mae: 0.0189\n",
      "Epoch 48/300\n",
      "135/135 - 0s - loss: 0.0026 - mae: 0.0412 - val_loss: 5.5910e-04 - val_mae: 0.0191\n",
      "Epoch 49/300\n",
      "135/135 - 0s - loss: 0.0027 - mae: 0.0408 - val_loss: 9.0136e-04 - val_mae: 0.0232\n",
      "Epoch 50/300\n",
      "135/135 - 0s - loss: 0.0031 - mae: 0.0419 - val_loss: 6.1536e-04 - val_mae: 0.0207\n",
      "Epoch 51/300\n",
      "135/135 - 0s - loss: 0.0029 - mae: 0.0428 - val_loss: 0.0010 - val_mae: 0.0248\n",
      "Epoch 52/300\n",
      "135/135 - 0s - loss: 0.0026 - mae: 0.0404 - val_loss: 5.7864e-04 - val_mae: 0.0197\n",
      "Epoch 53/300\n",
      "135/135 - 0s - loss: 0.0030 - mae: 0.0445 - val_loss: 5.8164e-04 - val_mae: 0.0197\n",
      "Epoch 54/300\n",
      "135/135 - 0s - loss: 0.0022 - mae: 0.0384 - val_loss: 6.0483e-04 - val_mae: 0.0205\n",
      "Epoch 55/300\n",
      "135/135 - 0s - loss: 0.0033 - mae: 0.0456 - val_loss: 6.5107e-04 - val_mae: 0.0204\n",
      "Epoch 56/300\n",
      "135/135 - 0s - loss: 0.0021 - mae: 0.0348 - val_loss: 6.0198e-04 - val_mae: 0.0195\n",
      "Epoch 57/300\n",
      "135/135 - 0s - loss: 0.0019 - mae: 0.0336 - val_loss: 5.6202e-04 - val_mae: 0.0190\n",
      "Epoch 58/300\n",
      "135/135 - 0s - loss: 0.0023 - mae: 0.0374 - val_loss: 6.9334e-04 - val_mae: 0.0202\n",
      "Epoch 59/300\n",
      "135/135 - 0s - loss: 0.0018 - mae: 0.0326 - val_loss: 5.9825e-04 - val_mae: 0.0198\n",
      "Epoch 60/300\n",
      "135/135 - 0s - loss: 0.0018 - mae: 0.0337 - val_loss: 7.3644e-04 - val_mae: 0.0208\n",
      "Epoch 61/300\n",
      "135/135 - 0s - loss: 0.0016 - mae: 0.0308 - val_loss: 5.2794e-04 - val_mae: 0.0193\n",
      "Epoch 62/300\n",
      "135/135 - 0s - loss: 0.0023 - mae: 0.0383 - val_loss: 7.6670e-04 - val_mae: 0.0222\n",
      "Epoch 63/300\n",
      "135/135 - 0s - loss: 0.0025 - mae: 0.0380 - val_loss: 5.5396e-04 - val_mae: 0.0199\n",
      "Epoch 64/300\n",
      "135/135 - 0s - loss: 0.0026 - mae: 0.0397 - val_loss: 0.0013 - val_mae: 0.0296\n",
      "Epoch 65/300\n",
      "135/135 - 0s - loss: 0.0025 - mae: 0.0404 - val_loss: 7.2329e-04 - val_mae: 0.0227\n",
      "Epoch 66/300\n",
      "135/135 - 0s - loss: 0.0029 - mae: 0.0442 - val_loss: 0.0013 - val_mae: 0.0289\n",
      "Epoch 67/300\n",
      "135/135 - 0s - loss: 0.0027 - mae: 0.0420 - val_loss: 9.7182e-04 - val_mae: 0.0264\n",
      "Epoch 68/300\n",
      "135/135 - 0s - loss: 0.0026 - mae: 0.0413 - val_loss: 0.0017 - val_mae: 0.0339\n",
      "Epoch 69/300\n",
      "135/135 - 0s - loss: 0.0028 - mae: 0.0408 - val_loss: 8.1727e-04 - val_mae: 0.0241\n",
      "Epoch 70/300\n",
      "135/135 - 0s - loss: 0.0027 - mae: 0.0414 - val_loss: 0.0013 - val_mae: 0.0283\n",
      "Epoch 71/300\n",
      "135/135 - 0s - loss: 0.0020 - mae: 0.0345 - val_loss: 7.5637e-04 - val_mae: 0.0234\n",
      "Epoch 72/300\n",
      "135/135 - 0s - loss: 0.0019 - mae: 0.0350 - val_loss: 0.0015 - val_mae: 0.0308\n",
      "Epoch 73/300\n",
      "135/135 - 0s - loss: 0.0020 - mae: 0.0347 - val_loss: 6.5290e-04 - val_mae: 0.0215\n",
      "Epoch 74/300\n",
      "135/135 - 0s - loss: 0.0021 - mae: 0.0363 - val_loss: 7.9276e-04 - val_mae: 0.0210\n",
      "Epoch 75/300\n",
      "135/135 - 0s - loss: 0.0020 - mae: 0.0337 - val_loss: 6.3665e-04 - val_mae: 0.0201\n",
      "Epoch 76/300\n",
      "135/135 - 0s - loss: 0.0018 - mae: 0.0340 - val_loss: 6.6238e-04 - val_mae: 0.0203\n",
      "Epoch 77/300\n",
      "135/135 - 0s - loss: 0.0019 - mae: 0.0351 - val_loss: 5.6441e-04 - val_mae: 0.0198\n",
      "Epoch 78/300\n",
      "135/135 - 0s - loss: 0.0019 - mae: 0.0328 - val_loss: 8.5463e-04 - val_mae: 0.0226\n",
      "Epoch 79/300\n",
      "135/135 - 0s - loss: 0.0014 - mae: 0.0294 - val_loss: 5.5183e-04 - val_mae: 0.0188\n",
      "Epoch 80/300\n",
      "135/135 - 0s - loss: 0.0014 - mae: 0.0283 - val_loss: 6.4363e-04 - val_mae: 0.0197\n",
      "Epoch 81/300\n",
      "135/135 - 0s - loss: 0.0014 - mae: 0.0281 - val_loss: 5.5903e-04 - val_mae: 0.0187\n",
      "Epoch 82/300\n",
      "135/135 - 0s - loss: 0.0019 - mae: 0.0353 - val_loss: 5.2956e-04 - val_mae: 0.0182\n",
      "Epoch 83/300\n",
      "135/135 - 0s - loss: 0.0014 - mae: 0.0296 - val_loss: 5.2053e-04 - val_mae: 0.0182\n",
      "Epoch 84/300\n",
      "135/135 - 0s - loss: 0.0016 - mae: 0.0325 - val_loss: 4.5333e-04 - val_mae: 0.0177\n",
      "Epoch 85/300\n",
      "135/135 - 0s - loss: 0.0016 - mae: 0.0306 - val_loss: 5.2260e-04 - val_mae: 0.0181\n",
      "Epoch 86/300\n",
      "135/135 - 0s - loss: 0.0018 - mae: 0.0349 - val_loss: 5.4628e-04 - val_mae: 0.0188\n",
      "Epoch 87/300\n",
      "135/135 - 0s - loss: 0.0016 - mae: 0.0310 - val_loss: 5.0102e-04 - val_mae: 0.0183\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 88/300\n",
      "135/135 - 0s - loss: 0.0016 - mae: 0.0320 - val_loss: 6.2127e-04 - val_mae: 0.0199\n",
      "Epoch 89/300\n",
      "135/135 - 0s - loss: 0.0015 - mae: 0.0291 - val_loss: 5.1344e-04 - val_mae: 0.0186\n",
      "Epoch 90/300\n",
      "135/135 - 0s - loss: 0.0020 - mae: 0.0367 - val_loss: 0.0014 - val_mae: 0.0316\n",
      "Epoch 91/300\n",
      "135/135 - 0s - loss: 0.0023 - mae: 0.0369 - val_loss: 5.0838e-04 - val_mae: 0.0190\n",
      "Epoch 92/300\n",
      "135/135 - 0s - loss: 0.0016 - mae: 0.0318 - val_loss: 6.6320e-04 - val_mae: 0.0203\n",
      "Epoch 93/300\n",
      "135/135 - 0s - loss: 0.0018 - mae: 0.0347 - val_loss: 5.1505e-04 - val_mae: 0.0190\n",
      "Epoch 94/300\n",
      "135/135 - 0s - loss: 0.0016 - mae: 0.0334 - val_loss: 5.5718e-04 - val_mae: 0.0184\n",
      "Epoch 95/300\n",
      "135/135 - 0s - loss: 0.0017 - mae: 0.0331 - val_loss: 9.4680e-04 - val_mae: 0.0241\n",
      "Epoch 96/300\n",
      "135/135 - 0s - loss: 0.0017 - mae: 0.0326 - val_loss: 5.6067e-04 - val_mae: 0.0198\n",
      "Epoch 97/300\n",
      "135/135 - 0s - loss: 0.0016 - mae: 0.0305 - val_loss: 5.3436e-04 - val_mae: 0.0182\n",
      "Epoch 98/300\n",
      "135/135 - 0s - loss: 0.0016 - mae: 0.0317 - val_loss: 7.0373e-04 - val_mae: 0.0209\n",
      "Epoch 99/300\n",
      "135/135 - 0s - loss: 0.0013 - mae: 0.0293 - val_loss: 4.9889e-04 - val_mae: 0.0181\n",
      "Epoch 100/300\n",
      "135/135 - 0s - loss: 0.0015 - mae: 0.0299 - val_loss: 5.4305e-04 - val_mae: 0.0184\n",
      "Epoch 101/300\n",
      "135/135 - 0s - loss: 0.0018 - mae: 0.0347 - val_loss: 6.7163e-04 - val_mae: 0.0196\n",
      "Epoch 102/300\n",
      "135/135 - 0s - loss: 0.0015 - mae: 0.0302 - val_loss: 5.7727e-04 - val_mae: 0.0203\n",
      "Epoch 103/300\n",
      "135/135 - 0s - loss: 0.0022 - mae: 0.0376 - val_loss: 6.5403e-04 - val_mae: 0.0196\n",
      "Epoch 104/300\n",
      "135/135 - 0s - loss: 0.0013 - mae: 0.0284 - val_loss: 5.9720e-04 - val_mae: 0.0202\n",
      "Epoch 105/300\n",
      "135/135 - 0s - loss: 0.0016 - mae: 0.0299 - val_loss: 6.1421e-04 - val_mae: 0.0194\n",
      "Epoch 106/300\n",
      "135/135 - 0s - loss: 0.0014 - mae: 0.0304 - val_loss: 5.9453e-04 - val_mae: 0.0190\n",
      "Epoch 107/300\n",
      "135/135 - 0s - loss: 0.0016 - mae: 0.0302 - val_loss: 5.8588e-04 - val_mae: 0.0206\n",
      "Epoch 108/300\n",
      "135/135 - 0s - loss: 0.0013 - mae: 0.0289 - val_loss: 5.3708e-04 - val_mae: 0.0183\n",
      "Epoch 109/300\n",
      "135/135 - 0s - loss: 0.0014 - mae: 0.0291 - val_loss: 4.7772e-04 - val_mae: 0.0182\n",
      "Epoch 110/300\n",
      "135/135 - 0s - loss: 0.0020 - mae: 0.0357 - val_loss: 9.3848e-04 - val_mae: 0.0245\n",
      "Epoch 111/300\n",
      "135/135 - 0s - loss: 0.0012 - mae: 0.0266 - val_loss: 7.0986e-04 - val_mae: 0.0222\n",
      "Epoch 112/300\n",
      "135/135 - 0s - loss: 0.0018 - mae: 0.0340 - val_loss: 0.0011 - val_mae: 0.0269\n",
      "Epoch 113/300\n",
      "135/135 - 0s - loss: 0.0014 - mae: 0.0287 - val_loss: 7.8596e-04 - val_mae: 0.0238\n",
      "Epoch 114/300\n",
      "135/135 - 0s - loss: 0.0017 - mae: 0.0334 - val_loss: 8.3126e-04 - val_mae: 0.0225\n",
      "Epoch 115/300\n",
      "135/135 - 0s - loss: 0.0015 - mae: 0.0300 - val_loss: 5.3966e-04 - val_mae: 0.0183\n",
      "Epoch 116/300\n",
      "135/135 - 0s - loss: 0.0014 - mae: 0.0291 - val_loss: 4.6957e-04 - val_mae: 0.0184\n",
      "Epoch 117/300\n",
      "135/135 - 0s - loss: 0.0016 - mae: 0.0310 - val_loss: 6.5416e-04 - val_mae: 0.0197\n",
      "Epoch 118/300\n",
      "135/135 - 0s - loss: 0.0012 - mae: 0.0255 - val_loss: 5.9308e-04 - val_mae: 0.0204\n",
      "Epoch 119/300\n",
      "135/135 - 0s - loss: 0.0011 - mae: 0.0262 - val_loss: 5.4807e-04 - val_mae: 0.0183\n",
      "Epoch 120/300\n",
      "135/135 - 0s - loss: 0.0012 - mae: 0.0282 - val_loss: 4.4760e-04 - val_mae: 0.0173\n",
      "Epoch 121/300\n",
      "135/135 - 0s - loss: 0.0016 - mae: 0.0303 - val_loss: 4.4569e-04 - val_mae: 0.0171\n",
      "Epoch 122/300\n",
      "135/135 - 0s - loss: 0.0014 - mae: 0.0290 - val_loss: 7.0395e-04 - val_mae: 0.0204\n",
      "Epoch 123/300\n",
      "135/135 - 0s - loss: 0.0011 - mae: 0.0257 - val_loss: 4.5502e-04 - val_mae: 0.0176\n",
      "Epoch 124/300\n",
      "135/135 - 0s - loss: 0.0012 - mae: 0.0265 - val_loss: 5.2463e-04 - val_mae: 0.0178\n",
      "Epoch 125/300\n",
      "135/135 - 0s - loss: 0.0013 - mae: 0.0283 - val_loss: 5.3575e-04 - val_mae: 0.0178\n",
      "Epoch 126/300\n",
      "135/135 - 0s - loss: 0.0013 - mae: 0.0267 - val_loss: 4.9494e-04 - val_mae: 0.0185\n",
      "Epoch 127/300\n",
      "135/135 - 0s - loss: 0.0019 - mae: 0.0343 - val_loss: 0.0012 - val_mae: 0.0272\n",
      "Epoch 128/300\n",
      "135/135 - 0s - loss: 8.9922e-04 - mae: 0.0229 - val_loss: 5.2436e-04 - val_mae: 0.0193\n",
      "Epoch 129/300\n",
      "135/135 - 0s - loss: 0.0014 - mae: 0.0303 - val_loss: 8.5146e-04 - val_mae: 0.0229\n",
      "Epoch 130/300\n",
      "135/135 - 0s - loss: 0.0014 - mae: 0.0291 - val_loss: 7.2090e-04 - val_mae: 0.0225\n",
      "Epoch 131/300\n",
      "135/135 - 0s - loss: 0.0015 - mae: 0.0302 - val_loss: 7.4875e-04 - val_mae: 0.0214\n",
      "Epoch 132/300\n",
      "135/135 - 0s - loss: 0.0012 - mae: 0.0258 - val_loss: 5.5566e-04 - val_mae: 0.0181\n",
      "Epoch 133/300\n",
      "135/135 - 0s - loss: 0.0012 - mae: 0.0269 - val_loss: 4.4762e-04 - val_mae: 0.0172\n",
      "Epoch 134/300\n",
      "135/135 - 0s - loss: 9.7720e-04 - mae: 0.0252 - val_loss: 4.6218e-04 - val_mae: 0.0179\n",
      "Epoch 135/300\n",
      "135/135 - 0s - loss: 0.0014 - mae: 0.0290 - val_loss: 9.7344e-04 - val_mae: 0.0249\n",
      "Epoch 136/300\n",
      "135/135 - 0s - loss: 9.4084e-04 - mae: 0.0233 - val_loss: 5.4729e-04 - val_mae: 0.0184\n",
      "Epoch 137/300\n",
      "135/135 - 0s - loss: 0.0012 - mae: 0.0261 - val_loss: 6.4578e-04 - val_mae: 0.0216\n",
      "Epoch 138/300\n",
      "135/135 - 0s - loss: 0.0016 - mae: 0.0306 - val_loss: 9.4727e-04 - val_mae: 0.0239\n",
      "Epoch 139/300\n",
      "135/135 - 0s - loss: 0.0014 - mae: 0.0291 - val_loss: 5.1307e-04 - val_mae: 0.0188\n",
      "Epoch 140/300\n",
      "135/135 - 0s - loss: 9.7642e-04 - mae: 0.0243 - val_loss: 5.3568e-04 - val_mae: 0.0181\n",
      "Epoch 141/300\n",
      "135/135 - 0s - loss: 0.0012 - mae: 0.0275 - val_loss: 7.4494e-04 - val_mae: 0.0208\n",
      "Epoch 142/300\n",
      "135/135 - 0s - loss: 7.7664e-04 - mae: 0.0225 - val_loss: 4.8521e-04 - val_mae: 0.0181\n",
      "Epoch 143/300\n",
      "135/135 - 0s - loss: 9.1167e-04 - mae: 0.0229 - val_loss: 4.7989e-04 - val_mae: 0.0176\n",
      "Epoch 144/300\n",
      "135/135 - 0s - loss: 0.0011 - mae: 0.0264 - val_loss: 4.5989e-04 - val_mae: 0.0174\n",
      "Epoch 145/300\n",
      "135/135 - 0s - loss: 0.0011 - mae: 0.0258 - val_loss: 6.0815e-04 - val_mae: 0.0188\n",
      "Epoch 146/300\n",
      "135/135 - 0s - loss: 7.3321e-04 - mae: 0.0213 - val_loss: 4.7286e-04 - val_mae: 0.0178\n",
      "Epoch 147/300\n",
      "135/135 - 0s - loss: 0.0012 - mae: 0.0257 - val_loss: 6.3959e-04 - val_mae: 0.0192\n",
      "Epoch 148/300\n",
      "135/135 - 0s - loss: 8.5627e-04 - mae: 0.0237 - val_loss: 4.5769e-04 - val_mae: 0.0175\n",
      "Epoch 149/300\n",
      "135/135 - 0s - loss: 0.0011 - mae: 0.0251 - val_loss: 4.4459e-04 - val_mae: 0.0171\n",
      "Epoch 150/300\n",
      "135/135 - 0s - loss: 0.0014 - mae: 0.0273 - val_loss: 4.8220e-04 - val_mae: 0.0177\n",
      "Epoch 151/300\n",
      "135/135 - 0s - loss: 8.5471e-04 - mae: 0.0219 - val_loss: 5.2401e-04 - val_mae: 0.0183\n",
      "Epoch 152/300\n",
      "135/135 - 0s - loss: 0.0012 - mae: 0.0256 - val_loss: 5.0642e-04 - val_mae: 0.0180\n",
      "Epoch 153/300\n",
      "135/135 - 0s - loss: 0.0011 - mae: 0.0253 - val_loss: 6.0113e-04 - val_mae: 0.0192\n",
      "Epoch 154/300\n",
      "135/135 - 0s - loss: 7.8427e-04 - mae: 0.0212 - val_loss: 5.3492e-04 - val_mae: 0.0188\n",
      "Epoch 155/300\n",
      "135/135 - 0s - loss: 0.0013 - mae: 0.0269 - val_loss: 5.3280e-04 - val_mae: 0.0190\n",
      "Epoch 156/300\n",
      "135/135 - 0s - loss: 9.6568e-04 - mae: 0.0249 - val_loss: 5.4858e-04 - val_mae: 0.0185\n",
      "Epoch 157/300\n",
      "135/135 - 0s - loss: 0.0013 - mae: 0.0284 - val_loss: 7.7102e-04 - val_mae: 0.0217\n",
      "Epoch 158/300\n",
      "135/135 - 0s - loss: 0.0010 - mae: 0.0250 - val_loss: 5.2016e-04 - val_mae: 0.0182\n",
      "Epoch 159/300\n",
      "135/135 - 0s - loss: 0.0010 - mae: 0.0244 - val_loss: 5.0263e-04 - val_mae: 0.0189\n",
      "Epoch 160/300\n",
      "135/135 - 0s - loss: 0.0012 - mae: 0.0257 - val_loss: 6.1586e-04 - val_mae: 0.0186\n",
      "Epoch 161/300\n",
      "135/135 - 0s - loss: 0.0015 - mae: 0.0300 - val_loss: 5.8334e-04 - val_mae: 0.0186\n",
      "Epoch 162/300\n",
      "135/135 - 0s - loss: 0.0010 - mae: 0.0234 - val_loss: 5.7278e-04 - val_mae: 0.0199\n",
      "Epoch 163/300\n",
      "135/135 - 0s - loss: 9.6414e-04 - mae: 0.0251 - val_loss: 6.7291e-04 - val_mae: 0.0201\n",
      "Epoch 164/300\n",
      "135/135 - 0s - loss: 7.7293e-04 - mae: 0.0222 - val_loss: 6.5103e-04 - val_mae: 0.0198\n",
      "Epoch 165/300\n",
      "135/135 - 0s - loss: 0.0011 - mae: 0.0256 - val_loss: 4.6593e-04 - val_mae: 0.0180\n",
      "Epoch 166/300\n",
      "135/135 - 0s - loss: 0.0013 - mae: 0.0283 - val_loss: 9.6575e-04 - val_mae: 0.0244\n",
      "Epoch 167/300\n",
      "135/135 - 0s - loss: 0.0012 - mae: 0.0255 - val_loss: 5.1320e-04 - val_mae: 0.0184\n",
      "Epoch 168/300\n",
      "135/135 - 0s - loss: 9.0805e-04 - mae: 0.0239 - val_loss: 4.6413e-04 - val_mae: 0.0176\n",
      "Epoch 169/300\n",
      "135/135 - 0s - loss: 0.0013 - mae: 0.0278 - val_loss: 0.0013 - val_mae: 0.0295\n",
      "Epoch 170/300\n",
      "135/135 - 0s - loss: 0.0011 - mae: 0.0248 - val_loss: 4.6091e-04 - val_mae: 0.0179\n",
      "Epoch 171/300\n",
      "135/135 - 0s - loss: 0.0011 - mae: 0.0265 - val_loss: 6.6793e-04 - val_mae: 0.0220\n",
      "Epoch 172/300\n",
      "135/135 - 0s - loss: 0.0015 - mae: 0.0299 - val_loss: 0.0014 - val_mae: 0.0309\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 173/300\n",
      "135/135 - 0s - loss: 0.0011 - mae: 0.0261 - val_loss: 4.6649e-04 - val_mae: 0.0177\n",
      "Epoch 174/300\n",
      "135/135 - 0s - loss: 7.5867e-04 - mae: 0.0213 - val_loss: 5.0185e-04 - val_mae: 0.0182\n",
      "Epoch 175/300\n",
      "135/135 - 0s - loss: 0.0010 - mae: 0.0242 - val_loss: 0.0012 - val_mae: 0.0277\n",
      "Epoch 176/300\n",
      "135/135 - 0s - loss: 0.0012 - mae: 0.0252 - val_loss: 4.6358e-04 - val_mae: 0.0177\n",
      "Epoch 177/300\n",
      "135/135 - 0s - loss: 8.9040e-04 - mae: 0.0226 - val_loss: 4.7618e-04 - val_mae: 0.0176\n",
      "Epoch 178/300\n",
      "135/135 - 0s - loss: 0.0011 - mae: 0.0262 - val_loss: 7.6446e-04 - val_mae: 0.0210\n",
      "Epoch 179/300\n",
      "135/135 - 0s - loss: 0.0013 - mae: 0.0242 - val_loss: 4.9778e-04 - val_mae: 0.0181\n",
      "Epoch 180/300\n",
      "135/135 - 0s - loss: 8.5625e-04 - mae: 0.0235 - val_loss: 5.7171e-04 - val_mae: 0.0187\n",
      "Epoch 181/300\n",
      "135/135 - 0s - loss: 0.0011 - mae: 0.0245 - val_loss: 7.4666e-04 - val_mae: 0.0213\n",
      "Epoch 182/300\n",
      "135/135 - 0s - loss: 0.0011 - mae: 0.0257 - val_loss: 9.2072e-04 - val_mae: 0.0261\n",
      "Epoch 183/300\n",
      "135/135 - 0s - loss: 0.0016 - mae: 0.0320 - val_loss: 5.0618e-04 - val_mae: 0.0184\n",
      "Epoch 184/300\n",
      "135/135 - 0s - loss: 0.0014 - mae: 0.0305 - val_loss: 7.7861e-04 - val_mae: 0.0216\n",
      "Epoch 185/300\n",
      "135/135 - 0s - loss: 9.8582e-04 - mae: 0.0251 - val_loss: 5.3882e-04 - val_mae: 0.0183\n",
      "Epoch 186/300\n",
      "135/135 - 0s - loss: 8.5733e-04 - mae: 0.0224 - val_loss: 5.3771e-04 - val_mae: 0.0183\n",
      "Epoch 187/300\n",
      "135/135 - 0s - loss: 7.6414e-04 - mae: 0.0214 - val_loss: 5.1530e-04 - val_mae: 0.0186\n",
      "Epoch 188/300\n",
      "135/135 - 0s - loss: 0.0015 - mae: 0.0277 - val_loss: 5.3544e-04 - val_mae: 0.0183\n",
      "Epoch 189/300\n",
      "135/135 - 0s - loss: 9.0601e-04 - mae: 0.0236 - val_loss: 8.8590e-04 - val_mae: 0.0242\n",
      "Epoch 190/300\n",
      "135/135 - 0s - loss: 8.3304e-04 - mae: 0.0232 - val_loss: 5.1555e-04 - val_mae: 0.0190\n",
      "Epoch 191/300\n",
      "135/135 - 0s - loss: 0.0011 - mae: 0.0265 - val_loss: 0.0012 - val_mae: 0.0292\n",
      "Epoch 192/300\n",
      "135/135 - 0s - loss: 8.7844e-04 - mae: 0.0234 - val_loss: 4.2993e-04 - val_mae: 0.0172\n",
      "Epoch 193/300\n",
      "135/135 - 0s - loss: 0.0012 - mae: 0.0250 - val_loss: 0.0011 - val_mae: 0.0281\n",
      "Epoch 194/300\n",
      "135/135 - 0s - loss: 0.0016 - mae: 0.0319 - val_loss: 6.7430e-04 - val_mae: 0.0200\n",
      "Epoch 195/300\n",
      "135/135 - 0s - loss: 0.0012 - mae: 0.0263 - val_loss: 6.2617e-04 - val_mae: 0.0196\n",
      "Epoch 196/300\n",
      "135/135 - 0s - loss: 9.5877e-04 - mae: 0.0242 - val_loss: 4.7338e-04 - val_mae: 0.0176\n",
      "Epoch 197/300\n",
      "135/135 - 0s - loss: 7.1959e-04 - mae: 0.0211 - val_loss: 5.4494e-04 - val_mae: 0.0183\n",
      "Epoch 198/300\n",
      "135/135 - 0s - loss: 8.5066e-04 - mae: 0.0226 - val_loss: 5.0195e-04 - val_mae: 0.0174\n",
      "Epoch 199/300\n",
      "135/135 - 0s - loss: 7.2526e-04 - mae: 0.0224 - val_loss: 4.8221e-04 - val_mae: 0.0171\n",
      "Epoch 200/300\n",
      "135/135 - 0s - loss: 8.5749e-04 - mae: 0.0222 - val_loss: 4.7523e-04 - val_mae: 0.0171\n",
      "Epoch 201/300\n",
      "135/135 - 0s - loss: 9.7069e-04 - mae: 0.0236 - val_loss: 4.5506e-04 - val_mae: 0.0171\n",
      "Epoch 202/300\n",
      "135/135 - 0s - loss: 9.3114e-04 - mae: 0.0235 - val_loss: 5.2145e-04 - val_mae: 0.0180\n",
      "Epoch 203/300\n",
      "135/135 - 0s - loss: 8.6766e-04 - mae: 0.0234 - val_loss: 5.3223e-04 - val_mae: 0.0178\n",
      "Epoch 204/300\n",
      "135/135 - 0s - loss: 8.3021e-04 - mae: 0.0211 - val_loss: 4.2021e-04 - val_mae: 0.0161\n",
      "Epoch 205/300\n",
      "135/135 - 0s - loss: 8.0091e-04 - mae: 0.0216 - val_loss: 4.2810e-04 - val_mae: 0.0166\n",
      "Epoch 206/300\n",
      "135/135 - 0s - loss: 7.6131e-04 - mae: 0.0226 - val_loss: 5.1383e-04 - val_mae: 0.0188\n",
      "Epoch 207/300\n",
      "135/135 - 0s - loss: 9.3608e-04 - mae: 0.0244 - val_loss: 5.1363e-04 - val_mae: 0.0175\n",
      "Epoch 208/300\n",
      "135/135 - 0s - loss: 9.1125e-04 - mae: 0.0234 - val_loss: 5.9933e-04 - val_mae: 0.0188\n",
      "Epoch 209/300\n",
      "135/135 - 0s - loss: 9.2133e-04 - mae: 0.0233 - val_loss: 5.0619e-04 - val_mae: 0.0186\n",
      "Epoch 210/300\n",
      "135/135 - 0s - loss: 9.1246e-04 - mae: 0.0239 - val_loss: 4.2738e-04 - val_mae: 0.0173\n",
      "Epoch 211/300\n",
      "135/135 - 0s - loss: 0.0017 - mae: 0.0334 - val_loss: 0.0016 - val_mae: 0.0353\n",
      "Epoch 212/300\n",
      "135/135 - 0s - loss: 0.0013 - mae: 0.0272 - val_loss: 6.1364e-04 - val_mae: 0.0194\n",
      "Epoch 213/300\n",
      "135/135 - 0s - loss: 0.0011 - mae: 0.0263 - val_loss: 0.0010 - val_mae: 0.0274\n",
      "Epoch 214/300\n",
      "135/135 - 0s - loss: 0.0015 - mae: 0.0311 - val_loss: 6.4655e-04 - val_mae: 0.0197\n",
      "Epoch 215/300\n",
      "135/135 - 0s - loss: 0.0011 - mae: 0.0262 - val_loss: 6.3060e-04 - val_mae: 0.0198\n",
      "Epoch 216/300\n",
      "135/135 - 0s - loss: 7.3240e-04 - mae: 0.0224 - val_loss: 5.8622e-04 - val_mae: 0.0190\n",
      "Epoch 217/300\n",
      "135/135 - 0s - loss: 9.9740e-04 - mae: 0.0244 - val_loss: 6.8973e-04 - val_mae: 0.0228\n",
      "Epoch 218/300\n",
      "135/135 - 0s - loss: 0.0014 - mae: 0.0290 - val_loss: 5.5962e-04 - val_mae: 0.0205\n",
      "Epoch 219/300\n",
      "135/135 - 0s - loss: 0.0015 - mae: 0.0311 - val_loss: 0.0016 - val_mae: 0.0339\n",
      "Epoch 220/300\n",
      "135/135 - 0s - loss: 0.0015 - mae: 0.0303 - val_loss: 0.0019 - val_mae: 0.0369\n",
      "Epoch 221/300\n",
      "135/135 - 0s - loss: 0.0013 - mae: 0.0308 - val_loss: 0.0014 - val_mae: 0.0331\n",
      "Epoch 222/300\n",
      "135/135 - 0s - loss: 0.0016 - mae: 0.0329 - val_loss: 5.1650e-04 - val_mae: 0.0194\n",
      "Epoch 223/300\n",
      "135/135 - 0s - loss: 0.0015 - mae: 0.0305 - val_loss: 8.0858e-04 - val_mae: 0.0225\n",
      "Epoch 224/300\n",
      "135/135 - 0s - loss: 0.0011 - mae: 0.0255 - val_loss: 7.3926e-04 - val_mae: 0.0216\n",
      "Epoch 225/300\n",
      "135/135 - 0s - loss: 9.3782e-04 - mae: 0.0233 - val_loss: 4.4854e-04 - val_mae: 0.0168\n",
      "Epoch 226/300\n",
      "135/135 - 0s - loss: 8.6327e-04 - mae: 0.0229 - val_loss: 7.0228e-04 - val_mae: 0.0228\n",
      "Epoch 227/300\n",
      "135/135 - 0s - loss: 0.0010 - mae: 0.0256 - val_loss: 5.0342e-04 - val_mae: 0.0178\n",
      "Epoch 228/300\n",
      "135/135 - 0s - loss: 0.0011 - mae: 0.0265 - val_loss: 9.4810e-04 - val_mae: 0.0248\n",
      "Epoch 229/300\n",
      "135/135 - 0s - loss: 8.2357e-04 - mae: 0.0230 - val_loss: 4.0483e-04 - val_mae: 0.0162\n",
      "Epoch 230/300\n",
      "135/135 - 0s - loss: 7.6660e-04 - mae: 0.0216 - val_loss: 4.4030e-04 - val_mae: 0.0167\n",
      "Epoch 231/300\n",
      "135/135 - 0s - loss: 0.0011 - mae: 0.0248 - val_loss: 0.0013 - val_mae: 0.0299\n",
      "Epoch 232/300\n",
      "135/135 - 0s - loss: 9.8083e-04 - mae: 0.0249 - val_loss: 4.8146e-04 - val_mae: 0.0173\n",
      "Epoch 233/300\n",
      "135/135 - 0s - loss: 9.9124e-04 - mae: 0.0239 - val_loss: 0.0010 - val_mae: 0.0278\n",
      "Epoch 234/300\n",
      "135/135 - 0s - loss: 0.0016 - mae: 0.0327 - val_loss: 5.4067e-04 - val_mae: 0.0182\n",
      "Epoch 235/300\n",
      "135/135 - 0s - loss: 0.0014 - mae: 0.0295 - val_loss: 0.0019 - val_mae: 0.0381\n",
      "Epoch 236/300\n",
      "135/135 - 0s - loss: 0.0010 - mae: 0.0253 - val_loss: 4.2050e-04 - val_mae: 0.0165\n",
      "Epoch 237/300\n",
      "135/135 - 0s - loss: 8.4349e-04 - mae: 0.0219 - val_loss: 4.4972e-04 - val_mae: 0.0179\n",
      "Epoch 238/300\n",
      "135/135 - 0s - loss: 9.5706e-04 - mae: 0.0243 - val_loss: 8.7141e-04 - val_mae: 0.0247\n",
      "Epoch 239/300\n",
      "135/135 - 0s - loss: 0.0013 - mae: 0.0280 - val_loss: 6.8992e-04 - val_mae: 0.0205\n",
      "Epoch 240/300\n",
      "135/135 - 0s - loss: 9.8380e-04 - mae: 0.0248 - val_loss: 9.1764e-04 - val_mae: 0.0242\n",
      "Epoch 241/300\n",
      "135/135 - 0s - loss: 0.0015 - mae: 0.0272 - val_loss: 5.4975e-04 - val_mae: 0.0193\n",
      "Epoch 242/300\n",
      "135/135 - 0s - loss: 9.5901e-04 - mae: 0.0225 - val_loss: 5.0715e-04 - val_mae: 0.0188\n",
      "Epoch 243/300\n",
      "135/135 - 0s - loss: 0.0012 - mae: 0.0263 - val_loss: 4.7053e-04 - val_mae: 0.0179\n",
      "Epoch 244/300\n",
      "135/135 - 0s - loss: 0.0014 - mae: 0.0291 - val_loss: 6.5355e-04 - val_mae: 0.0202\n",
      "Epoch 245/300\n",
      "135/135 - 0s - loss: 0.0011 - mae: 0.0251 - val_loss: 6.2392e-04 - val_mae: 0.0196\n",
      "Epoch 246/300\n",
      "135/135 - 0s - loss: 8.3266e-04 - mae: 0.0214 - val_loss: 3.8252e-04 - val_mae: 0.0158\n",
      "Epoch 247/300\n",
      "135/135 - 0s - loss: 7.4253e-04 - mae: 0.0201 - val_loss: 4.5505e-04 - val_mae: 0.0174\n",
      "Epoch 248/300\n",
      "135/135 - 0s - loss: 0.0010 - mae: 0.0235 - val_loss: 4.5050e-04 - val_mae: 0.0168\n",
      "Epoch 249/300\n",
      "135/135 - 0s - loss: 0.0011 - mae: 0.0245 - val_loss: 6.1303e-04 - val_mae: 0.0194\n",
      "Epoch 250/300\n",
      "135/135 - 0s - loss: 8.6902e-04 - mae: 0.0234 - val_loss: 8.8145e-04 - val_mae: 0.0238\n",
      "Epoch 251/300\n",
      "135/135 - 0s - loss: 6.8080e-04 - mae: 0.0203 - val_loss: 5.2861e-04 - val_mae: 0.0176\n",
      "Epoch 252/300\n",
      "135/135 - 0s - loss: 7.8778e-04 - mae: 0.0204 - val_loss: 4.6826e-04 - val_mae: 0.0182\n",
      "Epoch 253/300\n",
      "135/135 - 0s - loss: 9.6811e-04 - mae: 0.0248 - val_loss: 8.0089e-04 - val_mae: 0.0239\n",
      "Epoch 254/300\n",
      "135/135 - 0s - loss: 0.0017 - mae: 0.0336 - val_loss: 5.6220e-04 - val_mae: 0.0183\n",
      "Epoch 255/300\n",
      "135/135 - 0s - loss: 0.0014 - mae: 0.0300 - val_loss: 9.9186e-04 - val_mae: 0.0257\n",
      "Epoch 256/300\n",
      "135/135 - 0s - loss: 9.4673e-04 - mae: 0.0243 - val_loss: 5.3238e-04 - val_mae: 0.0179\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 257/300\n",
      "135/135 - 0s - loss: 7.9331e-04 - mae: 0.0222 - val_loss: 3.9430e-04 - val_mae: 0.0161\n",
      "Epoch 258/300\n",
      "135/135 - 0s - loss: 8.9712e-04 - mae: 0.0233 - val_loss: 5.5445e-04 - val_mae: 0.0199\n",
      "Epoch 259/300\n",
      "135/135 - 0s - loss: 0.0011 - mae: 0.0262 - val_loss: 4.9739e-04 - val_mae: 0.0175\n",
      "Epoch 260/300\n",
      "135/135 - 0s - loss: 0.0010 - mae: 0.0248 - val_loss: 4.7318e-04 - val_mae: 0.0168\n",
      "Epoch 261/300\n",
      "135/135 - 0s - loss: 7.5543e-04 - mae: 0.0207 - val_loss: 7.6692e-04 - val_mae: 0.0218\n",
      "Epoch 262/300\n",
      "135/135 - 0s - loss: 0.0010 - mae: 0.0240 - val_loss: 4.3788e-04 - val_mae: 0.0164\n",
      "Epoch 263/300\n",
      "135/135 - 0s - loss: 7.2998e-04 - mae: 0.0212 - val_loss: 3.8588e-04 - val_mae: 0.0161\n",
      "Epoch 264/300\n",
      "135/135 - 0s - loss: 5.8578e-04 - mae: 0.0202 - val_loss: 4.8235e-04 - val_mae: 0.0181\n",
      "Epoch 265/300\n",
      "135/135 - 0s - loss: 9.5724e-04 - mae: 0.0229 - val_loss: 4.9028e-04 - val_mae: 0.0179\n",
      "Epoch 266/300\n",
      "135/135 - 0s - loss: 8.7913e-04 - mae: 0.0237 - val_loss: 4.4616e-04 - val_mae: 0.0175\n",
      "Epoch 267/300\n",
      "135/135 - 0s - loss: 0.0014 - mae: 0.0286 - val_loss: 7.3441e-04 - val_mae: 0.0217\n",
      "Epoch 268/300\n",
      "135/135 - 0s - loss: 8.2698e-04 - mae: 0.0227 - val_loss: 6.1274e-04 - val_mae: 0.0196\n",
      "Epoch 269/300\n",
      "135/135 - 0s - loss: 8.9374e-04 - mae: 0.0230 - val_loss: 3.8329e-04 - val_mae: 0.0157\n",
      "Epoch 270/300\n",
      "135/135 - 0s - loss: 0.0011 - mae: 0.0230 - val_loss: 6.2220e-04 - val_mae: 0.0208\n",
      "Epoch 271/300\n",
      "135/135 - 0s - loss: 0.0011 - mae: 0.0235 - val_loss: 4.8650e-04 - val_mae: 0.0175\n",
      "Epoch 272/300\n",
      "135/135 - 0s - loss: 6.9088e-04 - mae: 0.0216 - val_loss: 5.1556e-04 - val_mae: 0.0188\n",
      "Epoch 273/300\n",
      "135/135 - 0s - loss: 8.9352e-04 - mae: 0.0227 - val_loss: 8.8408e-04 - val_mae: 0.0259\n",
      "Epoch 274/300\n",
      "135/135 - 0s - loss: 0.0015 - mae: 0.0301 - val_loss: 4.5029e-04 - val_mae: 0.0171\n",
      "Epoch 275/300\n",
      "135/135 - 0s - loss: 0.0014 - mae: 0.0287 - val_loss: 9.4009e-04 - val_mae: 0.0248\n",
      "Epoch 276/300\n",
      "135/135 - 0s - loss: 0.0014 - mae: 0.0290 - val_loss: 0.0016 - val_mae: 0.0349\n",
      "Epoch 277/300\n",
      "135/135 - 0s - loss: 8.6396e-04 - mae: 0.0235 - val_loss: 6.6030e-04 - val_mae: 0.0224\n",
      "Epoch 278/300\n",
      "135/135 - 0s - loss: 0.0013 - mae: 0.0292 - val_loss: 8.9461e-04 - val_mae: 0.0257\n",
      "Epoch 279/300\n",
      "135/135 - 0s - loss: 0.0015 - mae: 0.0319 - val_loss: 5.7691e-04 - val_mae: 0.0191\n",
      "Epoch 280/300\n",
      "135/135 - 0s - loss: 0.0018 - mae: 0.0329 - val_loss: 6.5735e-04 - val_mae: 0.0196\n",
      "Epoch 281/300\n",
      "135/135 - 0s - loss: 0.0010 - mae: 0.0258 - val_loss: 7.7538e-04 - val_mae: 0.0220\n",
      "Epoch 282/300\n",
      "135/135 - 0s - loss: 7.0102e-04 - mae: 0.0200 - val_loss: 4.8051e-04 - val_mae: 0.0170\n",
      "Epoch 283/300\n",
      "135/135 - 0s - loss: 6.6671e-04 - mae: 0.0205 - val_loss: 4.7001e-04 - val_mae: 0.0170\n",
      "Epoch 284/300\n",
      "135/135 - 0s - loss: 7.9606e-04 - mae: 0.0216 - val_loss: 3.9918e-04 - val_mae: 0.0165\n",
      "Epoch 285/300\n",
      "135/135 - 0s - loss: 7.2170e-04 - mae: 0.0213 - val_loss: 3.9260e-04 - val_mae: 0.0165\n",
      "Epoch 286/300\n",
      "135/135 - 0s - loss: 0.0010 - mae: 0.0253 - val_loss: 5.9124e-04 - val_mae: 0.0187\n",
      "Epoch 287/300\n",
      "135/135 - 0s - loss: 6.3875e-04 - mae: 0.0205 - val_loss: 4.8276e-04 - val_mae: 0.0167\n",
      "Epoch 288/300\n",
      "135/135 - 0s - loss: 7.2854e-04 - mae: 0.0217 - val_loss: 5.4189e-04 - val_mae: 0.0188\n",
      "Epoch 289/300\n",
      "135/135 - 0s - loss: 8.1253e-04 - mae: 0.0215 - val_loss: 4.8915e-04 - val_mae: 0.0184\n",
      "Epoch 290/300\n",
      "135/135 - 0s - loss: 7.9679e-04 - mae: 0.0225 - val_loss: 4.4577e-04 - val_mae: 0.0171\n",
      "Epoch 291/300\n",
      "135/135 - 0s - loss: 7.4563e-04 - mae: 0.0216 - val_loss: 4.1396e-04 - val_mae: 0.0161\n",
      "Epoch 292/300\n",
      "135/135 - 0s - loss: 9.0528e-04 - mae: 0.0223 - val_loss: 4.0072e-04 - val_mae: 0.0158\n",
      "Epoch 293/300\n",
      "135/135 - 0s - loss: 8.6551e-04 - mae: 0.0225 - val_loss: 4.2021e-04 - val_mae: 0.0160\n",
      "Epoch 294/300\n",
      "135/135 - 0s - loss: 7.4310e-04 - mae: 0.0211 - val_loss: 4.2632e-04 - val_mae: 0.0160\n",
      "Epoch 295/300\n",
      "135/135 - 0s - loss: 7.7390e-04 - mae: 0.0209 - val_loss: 5.6929e-04 - val_mae: 0.0194\n",
      "Epoch 296/300\n",
      "135/135 - 0s - loss: 6.3809e-04 - mae: 0.0198 - val_loss: 4.3060e-04 - val_mae: 0.0168\n",
      "Epoch 297/300\n",
      "135/135 - 0s - loss: 8.2602e-04 - mae: 0.0213 - val_loss: 4.5386e-04 - val_mae: 0.0177\n",
      "Epoch 298/300\n",
      "135/135 - 0s - loss: 0.0012 - mae: 0.0281 - val_loss: 4.3318e-04 - val_mae: 0.0170\n",
      "Epoch 299/300\n",
      "135/135 - 0s - loss: 0.0011 - mae: 0.0267 - val_loss: 4.5545e-04 - val_mae: 0.0167\n",
      "Epoch 300/300\n",
      "135/135 - 0s - loss: 9.6613e-04 - mae: 0.0232 - val_loss: 6.6167e-04 - val_mae: 0.0203\n"
     ]
    }
   ],
   "source": [
    "model = tf.keras.models.Sequential([\n",
    "  tf.keras.layers.Conv1D(filters=32, \n",
    "                         kernel_size=5,\n",
    "                         strides=1,\n",
    "                         padding=\"causal\",\n",
    "                         activation='relu',\n",
    "                         input_shape=(X_train.shape[1], X_train.shape[2])),\n",
    "  tf.keras.layers.LSTM(50,return_sequences=True),\n",
    "  tf.keras.layers.LSTM(50),\n",
    "  tf.keras.layers.Dense(128,activation='relu'),\n",
    "  tf.keras.layers.Dropout(0.1),\n",
    "  tf.keras.layers.Dense(1),\n",
    "  tf.keras.layers.Lambda(lambda x: x * 400)\n",
    "])\n",
    "\n",
    "#model.summary()\n",
    "\n",
    "model.compile(loss='mse',\n",
    "                  optimizer=Adam(lr = 1e-3),\n",
    "                  metrics=[\"mae\"]\n",
    "                  )\n",
    "\n",
    "\n",
    "history = model.fit(X_train, Y_train, epochs=300, batch_size=30, validation_data=(X_test, Y_test), \n",
    "                     verbose=2, shuffle=False)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "# make predictions\n",
    "train_predict = model.predict(X_train)\n",
    "test_predict = model.predict(X_test)\n",
    "# invert predictions\n",
    "train_predict = scaler.inverse_transform(train_predict)\n",
    "Y_train_Sca = scaler.inverse_transform([Y_train])\n",
    "test_predict = scaler.inverse_transform(test_predict)\n",
    "Y_test_Sca = scaler.inverse_transform([Y_test])\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Mean Absolute Error: 3.133602991419179\n",
      "Train Root Mean Squared Error: 3.8327178070210586\n",
      "Test Mean Absolute Error: 3.805154383088861\n",
      "Test Root Mean Squared Error: 4.830273340845651\n"
     ]
    }
   ],
   "source": [
    "print('Train Mean Absolute Error:', mean_absolute_error(Y_train_Sca[0], train_predict[:,0]))\n",
    "print('Train Root Mean Squared Error:',np.sqrt(mean_squared_error(Y_train_Sca[0], train_predict[:,0])))\n",
    "print('Test Mean Absolute Error:', mean_absolute_error(Y_test_Sca[0], test_predict[:,0]))\n",
    "print('Test Root Mean Squared Error:',np.sqrt(mean_squared_error(Y_test_Sca[0], test_predict[:,0])))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x191918c4c08>]"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(Y_train_Sca[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x1919a6b6348>]"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(train_predict[:,0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 93.75329 ,  94.683586,  94.69918 ,  94.89928 ,  95.118385,\n",
       "        98.35572 , 102.4501  , 109.24607 , 113.40583 , 118.87469 ,\n",
       "       127.69629 , 130.40967 , 134.40865 , 140.09383 , 145.93364 ,\n",
       "       153.88995 , 164.75761 , 170.74039 , 175.33017 , 179.20808 ,\n",
       "       182.50626 , 181.09012 , 183.71758 , 183.77104 , 181.11047 ,\n",
       "       177.67332 , 177.41835 , 177.3623  , 171.86888 , 167.812   ,\n",
       "       158.21857 , 153.17033 , 150.12234 , 147.37704 , 139.77525 ,\n",
       "       133.10187 , 125.03928 , 119.34273 , 114.31351 , 111.726685,\n",
       "       107.71659 , 105.22999 , 102.43676 , 100.50353 , 101.066765,\n",
       "       100.68381 , 101.77319 ,  99.83956 , 101.077774, 102.27349 ,\n",
       "       104.67652 , 105.746574, 107.62142 , 112.612724, 119.98809 ,\n",
       "       126.263985, 130.82869 , 137.2197  , 138.77022 , 144.0147  ,\n",
       "       150.8653  , 158.19229 , 165.67294 , 167.95082 , 169.47685 ,\n",
       "       172.78555 , 176.3503  , 174.562   , 170.32104 , 169.49216 ,\n",
       "       168.53835 , 167.47574 , 165.51013 , 160.27599 , 153.87698 ,\n",
       "       147.84239 , 140.616   , 136.43192 , 133.89375 , 130.0356  ,\n",
       "       121.788475, 116.69035 , 113.837585, 111.77507 , 109.17239 ,\n",
       "       108.425285, 106.60323 , 105.5631  , 106.75955 , 108.003456,\n",
       "       107.92511 , 107.63977 , 107.68064 , 108.59221 , 108.86833 ,\n",
       "       109.39736 , 109.600784, 109.678696, 111.80219 , 113.40142 ,\n",
       "       114.958435, 116.37737 , 118.74909 , 120.35659 , 120.88439 ,\n",
       "       123.99058 , 130.55786 , 135.43712 , 142.0805  , 148.5856  ,\n",
       "       152.95189 , 156.05742 , 160.60603 , 165.63199 , 169.94304 ,\n",
       "       172.9063  , 173.62035 , 175.2296  , 178.75215 , 175.53813 ,\n",
       "       172.53198 , 173.2193  , 168.65167 , 161.48201 , 159.4705  ,\n",
       "       159.95734 , 157.39249 , 151.26259 , 145.93604 , 139.62009 ,\n",
       "       136.75288 , 132.27992 , 129.06421 , 125.71607 , 121.433365],\n",
       "      dtype=float32)"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_predict[:,0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x1919a701c88>]"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAD4CAYAAAAXUaZHAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3dd3SU15nH8e+jjioCNZBQAYTopohejCu4xI5jJ4YkjuNG7Ni7zp7EiR1vkt1NyG6SXafsmmTxusWxcWzjQhw3jGk2YBBdFBVAHdS7UJ27f2iEBQgkpBm9U57POTpo7rwz+h2QHq7uve+9YoxBKaWUZ/GxOoBSSinH0+KulFIeSIu7Ukp5IC3uSinlgbS4K6WUB/KzOgBAVFSUSU5OtjqGUkq5lT179lQYY6J7es4lintycjIZGRlWx1BKKbciIvkXe06HZZRSygNpcVdKKQ+kxV0ppTyQFnellPJAWtyVUsoDaXFXSikPpMVdKaU8kEusc1dKeYYNR0qpbGhh4shwxsWGEeTva3Ukr6XFXSnlEOv2FPH91w+cfezrI4yJDuHrsxP59oIUC5N5Jy3uSqkB21tQzRNvHmL+mOGsum0Kx07VceRUHVuzy/m3d48wd8xwxseFWx3Tq+iYu1JqQE7XNvOdl/YQFxHE01+fQUpUCDdMGcH3r0/jxXtnExbkz6q/H0VPfRtcWtyVUv3W3NbBd17KoKmlnWe+lU5kSMA5zw8NDuDRa1LZllPB5qxyi1J6Jy3uSql+Mcbw+LqDHCiq5bd3TiMtLqzH6745N4mUqBB+8fcjtHXYBjml99LirpTql8ziOt7eX8I/XpPK9ZPiLnpdgJ8PT9wwnuPljby6q2AQE3o3Le5KqX7ZlVcFwIrZo3q99rqJscwbPZynNmRTe6bN2dEUWtyVUv20J7+K+KFDGBExpNdrRYQnb5pAzZk2nt6UOwjpVK/FXUSeE5EyEcns1vZXEdlv/8gTkf3dnntCRHJFJEtEljoruFLKOsYYMvKqSU+O7PNrJsdHcPuMBF74LI+6Zu29O1tfeu4vAMu6Nxhj7jTGTDPGTAPWAW8CiMhEYDkwyf6a1SKit6gp5WGKqs9QVt9CelLfizvAV2cm0NphY3tuhZOSqS69FndjzFagqqfnRESArwFr7U23Aq8aY1qMMSeBXGC2g7IqpVzEbvt4e3rysMt63YykSMIC/XRZ5CAY6Jj7IqDUGJNjfxwPFHZ7vsjedgERWSkiGSKSUV6u/9BKuZOM/GrCAv0YF9vz8seL8ff1YWFqFJuzyvWmJicbaHFfwRe9dgDp4Zoe/wWNMWuMMenGmPTo6B4P71ZKuag9edVMT4rE16enH/lLW5IWzem6ZrJK652QTHXpd3EXET/gK8BfuzUXAd3XRSUAJf39Gkop11Pb1EZWaf1lj7d3uXJcDIAOzTjZQHru1wLHjDFF3drWA8tFJFBEUoBUYNdAAiqlXMvegmqAy1op011cRBDj48LYnFXmyFjqPH1ZCrkW2AGkiUiRiNxnf2o55w7JYIw5DLwGHAE+AB42xnQ4NrJSykoZ+VX4+gjTRg3t93ssSYshI6+ael0S6TR9WS2zwhgzwhjjb4xJMMY8a2//tjHmTz1cv8oYM8YYk2aMed8ZoZVS1snIq2bSyHCCA/q/Y/iStGjabYbPcisdmEx1p3eoKqX6rLXdxv7CGmb2c7y9y8ykSEID/diSrUMzzqLFXSnVZ4dLamlptzHrMte3n8/f14cFY4ezRZdEOo0Wd6VUn+3Jt0+mDrDnDp3j7iW1zeSUNQz4vdSFtLgrpfpsd14Vo4YNISY8aMDvtSSt8/4WXTXjHFrclVJ9YoxhT3416UkDG5LpMiJiCGmxYbre3Um0uCul+iS/somKhtZ+r2/vyZK0aHbnVVHd2Nrj86drm/nle0cvuQ98W4eNvIpGh2XyFP1fy6SU8ipdh3M4qucOsGxyHGu2nWDRrzdx+4x47pqXzNiYUMrqmlm9+Tiv7CrAZjO02wwvbs/ju0vGcPf8ZIL8fSmta+aVzwtYu6uAsvoWXr5/DgvGRjksm7sTV5ipTk9PNxkZGVbHUEpdwqOv7uOz3Ap2/fhafPqxp8zFHCis4cUdebx74BStHTZmJA7lcEkd7TbDHTMSeOTqsdQ1t/GbD7PYnFXOiIggpiZEsPFoGe02w5K0aLJP1xMRHMC7/7CwX/vduCsR2WOMSe/pOe25K6V6ZbMZtuVUcOW4aIcWdoArRg3lqVHT+PGNE/jr7kLe2V/MTVNH8Og1qSQNDzl73Qv3zGbniUp+/cExdp2s4tvzk/nm3CSSo0J4Z38xj766n3V7i/haeu/H/nkD7bkrpXqVWVzLzf/9KU997Qq+MiPB6jgXMMZw2+rtlNScYfNjSwZ096w7uVTPXSdUlVK92prTuaJloYuOaYsIP7l5AmX1LazZesLqOC5Bi7tSqlfbsisYHxfmkPXtzjIzaRg3TRnB/245QWlds9VxLKfFXSl1SU2t7WTkV3HlONc/VOdHy8bTYTP854dZVkexnBZ3pdQl7TxRSVuHYVGq6xf3xOHBfHtBMm/sLeLzE96946QWd6XUJW3NriDQz8ehNy8508NXjSVxWDDfem4X7x06ZXUcy2hxV0pd0raccuaMHk6Qv6/VUfokYog/bz40n8nxEXz35b3875bjXrnzpBZ3pdRFFdec4Xh5I4tTXXOVzMUMDw3k5fvncPPUEfz7+8f48VuHaG23WR1rUHnHYlClVL98al8CudgNJlPPF+Tvyx+WTydpeDBPbzrOq7sLiRjiz7DgACJDArhhchz3LxptdUyn0eKulLqordkVxIYHkhoTanWUfvHxER5bOp70pGHsL6yhuqmVqsZWcssaWPXeURalRpMWF2Z1TKfQ4q6U6lGHzfBpbgXXTYxFxL33a7lqfAxXjY85+7imqZWFv9rEHzbm8PQ3ZliYzHl6HXMXkedEpExEMs9r/wcRyRKRwyLy627tT4hIrv25pc4IrZRyvkPFtdSeaWORm42398XQ4ADuWZDM3w+d4tjpOqvjOEVfJlRfAJZ1bxCRq4BbganGmEnAf9rbJwLLgUn216wWEfeYYldKnWNrdjkiuMX69v64b2EKYYF+/P7jHKujOEWvxd0YsxWoOq/5IeA/jDEt9mu6zsm6FXjVGNNijDkJ5AKzHZhXKTUIbDbDW/uKmZkYybCQAKvjOEVX7/39zNMcKTm3926M4eipOrdeQtnfpZDjgEUi8rmIbBGRWfb2eKCw23VF9rYLiMhKEckQkYzycs87Zstmc99vCqW25VZwsqKRu+YlWR3Fqe5bOJqwID/+sPGL3ntpXTN3P7+bG36/jbf3F1uYbmD6W9z9gEhgLvAY8Jp0zrj0NOvSY5UzxqwxxqQbY9Kjoz3r1778ykbSV33M6s25VkdRql9e2pFHVGgAyybHWR3FqSKC/bl3QQofHD7N4ZJaPsg8xdLfbWXXyUqiQgN5YXu+1RH7rb/FvQh403TaBdiAKHt7953yE4CSgUV0P2/sKaKqsZVff5DlseN5ynMVVjWx8VgZK2YnEujn+VNm9y5MISzIj28/v5sH/7KXUZHB/P0fF/EPV4/lQGENB4tqrI7YL/0t7m8DVwOIyDggAKgA1gPLRSRQRFKAVGCXI4K6C5vN8ObeYhaMHc7tMxL47cfZPPVRlluP3Snv8ped+fiI8PU5iVZHGRQRQ/x58MoxVDa08MhVY3nzu/MZEx3KbTPiCQ7w5c873LP33us6dxFZCywBokSkCPgZ8BzwnH15ZCtwt+msXodF5DXgCNAOPGyM6XBWeFe0K6+K4pozPLY0jVuuGImfj/CHT3JptxkeW5rm9uuFlWdrbuvgrxmFXD8xlhERQ6yOM2i+u2QMy2eNYnho4Nm28CB/bpsezxt7injyxglEutnEcq/F3Riz4iJPffMi168CVg0klDt7c28RIQG+XD8pFh8f4d+/MgUfH2H15uNEBgfwwGLPvd1Zub/1B0qoaWrjW/OSrY4yqETknMLe5a55Sbz8eQGv7ylk5eIxA/46p2ubiQoNwM/X+dt66cZhDnSmtYP3Dp3mhikjzp7h6OMjrPryZK4ZH8PvN+ZQ2dBicUqlemaM4c878hgXG8rc0cOsjuMSxseFMzt5GH/ZWTDgFXCfn6hkwa8+Ye3uwt4vdgAt7g604WgpDS3tfGXGuas/fXyEJ24cT1NrO09vOm5ROqUubV9hDZnFdXxrXrIOH3Zz17wkCqqa2JJz6SXbh0tqeXVXQY/za9WNrTz66n46bIbtuRXOinoOLe4O9ObeIkZGBDE3ZfgFz42NCeOOmQn8ZWc+RdVNFqRT6tJe+CyPsEA/bpve460pXmvppDiiQgN56RITq+v2FHHb6u08/uYhnnw785xevjGGx944SGVjC1ckRLA7r3pQFlhocXeQsvpmtmaX8+Xp8fj49Nzr+d6140DgqQ3ZDvmaW7PLOVDonsu0lGvZfryC9QdK+PrcREICdT/B7gL8fPj67FFsyiqjoPLcjll7h41/+9sRvv/6AWYkDuW+hSm88nkBP1x3kA57gX9xex4fHy3l8Rsm8LVZo6hoaCG/0vkdPP1XdJD1+0uwGS4Ykulu5NAh3DM/mTXbTrBy8WjGx4X362sZY1i9+Ti/+TCL4SEBfPKDJUQM8e9vdOXl6pvbeOz1g6REhfC9a8ZZHcclrZiTyOrNx7nut1uYNmooc1KGMT0pkv/bdoLPciv59vxknrxpAn4+QniQP7/9OJvWdhv3L0rhl+8d4+rxMdy7IJmcsgYAdudVkRwV4tTM2nN3kHV7i7kiIYKxMZfeG/qhJWMIC/Tj1x/073T29g4bP34rk998mMWStGiqm1r53ceO+U1AeadfvHuUU7Vn+M+vXsGQAM+/aak/RkQM4a/fmcddc5M409bB05uPc8/zu9mdV81v7pjKv9wyCX9fH0SER69N5UfLxrP+QAm3/3E7kSH+/OaOqYgIY6NDiRjiT0ZetdMza8/dAY6equPoqTr+9ZZJvV47NDiAB5eM4dcfZLHrZBWzU/q+KqGptZ1HXtnHJ8fK+O6SMTy2NI2fvJPJn3fks3xWosceOqCc55Njpfw1o5CHloxhZpJ7HIBtlZlJkWf/jhpa2tlXUE380CGMjr7wIJOHlowhyN+H327I5nd3Tj+7zNLHR0hPimR3/vl7MTqeFncHePbTkwT6+fClK0b26fp75qfw4vY87lyzg9AAP0KD/AgL8mPk0CHcNj2epZPizjmMuLmtgw8yT/OnLcfJLq3nF1+ezDfndm7o9P3r0nj34Cl+tj6TtQ/M1VUOqs+qG1v50bpDjI8L43vXplodx62EBvr1uhXyPQtSuHte8gVzcOnJw9h4rIzKhpYe19Y7ihb3AcqvbOStfcXcPS+5z1ujDgnw5cV7Z/PeodPUN7dR39xOfXMbR07V8eir+xka7M9XpidwzYQYPj5aylv7iqlpaiNxWDDPfCudaybEnn2vyJAAfnB9Gv/8diZ/P3SKm6f27T8YpX62/jDVja28cM8sr9hDxgo9La6YldzZ+9+TX831k5y3MZsW9wF6elMufj7Cg1de3p2n4+PCL5hQtdkM249XsnZXAS/tzOO5z07i7yssnRTHitmJzBs9vMdvlhWzE1m7q4BVfz/K1eNjzt5ApdTFHCisYf2BEr53bSqTRkZYHcerTI6PIMDXhwwt7q6roLKJN/cW8825ScSEBw34/Xx8hIWpUSxMjaKioYXd9jH53n518/UR/vWWSdzxpx2s3nScHyxNG3AW5dn+79OThAX6cf8i3Q5jsAX5+zI1IYLdec4dd9fVMgPw9KZcfHyEh5YMfM+J80WFBnLDlBF9HpNLTx7GV6bHs2brCfIrGx2eR3mO4pozvHfoFCvmJBKqa9otkZ48jMziWs60Om9fRS3u/VRY1cS6vUWsmDWKWAf02h3h8RvG4+8r/PzdI1ZHUS7shc9OAvDt+cnWBvFis5IjaeswHHDiXvFa3Ptp9eZcfER40Am99v6KCQ/iH69J5eOjZWzKKuv9Bcrr1De38equQm6aMoKRQ71nS19X07Wkck++89a7a3Hvh6LqJl7PKOLOWaNcbs/rexakMDoqhJ//7Qit7Tar4ygX89fdhdS3tHP/ohSro3i1ocEBjIsNdeq4uxb3fnj+szxEcMpY+0AF+Pnwky9N5ERFI8/bf/1WCjrvbn7+szxmpwxjasJQq+N4vZlJw9iTX312DxpH0+LeD5/lVjB39HCX/bX2qrQYrp0Qwx825lBa12x1HOUiPjh8muKaM9y/UHvtrmBWciT1ze1kl9Y75f21uF+m2jNtZJXWk57k2ocZ/PNNE2nrMPzq/WNWR1EuwBjDM9tOkjw8+Jyb4JR1ZiV31pAMJw3NaHG/TPsKqjEG0pNdex+O5KgQ7l+Uwpv7isksrrU6jrLYwaJaDhTWcO/CFHwvsiW1GlwJkUOIDQ9kt5M2EdPifpn25Ffj6yNMG+X6Y5YP2negXL051+ooymL7CjoLyFIn3hGpLo+I8LMvTeJuJy1J1eJ+mTLyqpkwIswtDjQID/LnW/OTeD/zNLn2faSVd8opayA8yI+YMOdtVKUu341TRjhtN85ei7uIPCciZSKS2a3tX0SkWET22z9u7PbcEyKSKyJZIrLUKakt0tZhY39hjcuPt3d3z4IUAnx9+N8tenarN8spbWBcbJjuGupF+tJzfwFY1kP7b40x0+wf7wGIyERgOTDJ/prVIuIx280dPVXHmbYOt9r3Oio0kBWzE3lrXzHFNWesjqMsYIwhu6ye1NgL9x1XnqvX4m6M2Qr0dTr3VuBVY0yLMeYkkAvMHkA+l9J1eoqrT6ae74HFnZtDPbP1xDntpXXN3PXs5zz3qa6H92QVDa3UNLWR2sspYcqzDGTM/REROWgftumqdvFAYbdriuxtFxCRlSKSISIZ5eXlA4gxePbkd5684mp3pfYm3n4IyNpdBVQ0tABw7HQdtz39GdtyKvjthmzqmtssTqmcJce+jnpcrBZ3b9Lf4v5HYAwwDTgF/Je9vacBvR5vvzLGrDHGpBtj0qOjL32iiSswxpCRX+V2vfYuDy4ZQ2uHjec/O8nW7HLu+OMOOozh17dPpb6lnZd3FlgdUTlJ16HMOizjXfpV3I0xpcaYDmOMDXiGL4ZeioBR3S5NAEoGFtE1FFWfobSuhXQ3Gm/vbkx0KDdMjuPZT09yzwu7SYgcwtsPL+Brs0axKDWK5z47SXOb87YfVdbJLq3XlTJeqF/FXURGdHt4G9C1kmY9sFxEAkUkBUgFdg0somvIsB9oO9ONVsqc77tLxtLWYVg4NorXH5x3dnjpwSvHUF7fwpt7iy1OqJwhp6yBVF0p43V6XawtImuBJUCUiBQBPwOWiMg0Oodc8oDvABhjDovIa8ARoB142BjjEd3BjLxqwgL9SItz33HLyfERbPvhVcSGB51zl+L8McOZEh/Bmq3HuXPWKL2D0YMYY8gprWfZZL15ydv0WtyNMSt6aH72EtevAlYNJJQr2pNfzbTEoW5f+Hra7Eyk8zSp7768lw8Pn+bGKSN6eKVyR5WNrVQ3tTFWV8p4Hb1DtQ/cZbOwgVg6KY6UqBD+uPk4xjhnC1I1+LLPrpTRyVRvo8W9D9xls7CB8PURVi4ezaHiWrYfr7Q6jnKQnNLOlTK6DNL7aHHvA3faLGwgbpseT3RYIL/4+1GOl+teNJ4gp6yeMF0p45W0uPeBO20WNhBB/r78/NbJFFU1sex3W/mP94/R2NJudSw1ANm6p4zX0uLei3Y33CxsIJZNjuOTHyzhlivi+dOW41z71BbeP3TK6liqn3LLGkiN0fF2b6TFvRfHTtdzpq2DGW5681J/RIcF8l9fu4I3HpxHZHAAD728l5++k0lbhx647U4qGlqoamwlVcfbvZIW917stR9yMCPRs8fbe5KePIz1jyzggUUp/HlHPt/4v8/P7k2jXF/XZKr23L2TFvde7MmvJjY8kHgXPQzb2fx8fXjypon87s5pHCis4Zb//pRDRXpsnzvIKdMNw7yZFvde7C2oZmZSpNdPSH15ejzrHpqPiHDHn7af3WlQua6c0gbCgvyIDdeVMt5Ii/sllNU3U1h1hhmJ3jPefimT4yN4/cF5tLTb2HiszOo4qhfZpfWkxoR6fcfEW2lxv4S9+TUAXjWZ2puRQ4eQEhXC3nznnNiuHCenrEGHZLyYFvdL2FtQTYCvD5NGhlsdxaXMSIxkb0G1blPgwirtK2XG6mSq19Lifgl786uZkhBBoJ/HHAPrEDOTIqloaKWgqsnqKOoisnXbAa+nxf0iWtttHCyu9colkL3pOiB8jw7NuKxc+0oZPX3Je2lxv4jDJbW0ttvOFjL1hdSYUMIC/bS4u7B9BTUMCwkgLjzI6ijKIlrcL6KrcOlKmQv5+AjTkyK1uLsom82wJbucRalRulLGi2lxv4i9BdUkRA4hRns+PZqZGElWaT31zW1WR1HnySyppbKxlSVprn/wvHIeLe49MMawJ79ah2QuYWZSJMbA/sIaq6Oo82zOKkcEFqdqcfdmWtx7UFLbTGldiw7JXMIVoyLwEZ1UdUWbs8qYGh/B8FC9M9WbaXHvQdcNOtpzv7iwIH/S4sK1uLuYmqZW9hfWcGVajNVRlMW8vrg3tLTz3xtz7Efpdd6Usye/miH+voyP0zXClzIjcSj7C2rosOnNTK5ia04FNoOOt6vei7uIPCciZSKS2cNzPxARIyJR3dqeEJFcEckSkaWODuxo6/YU8V8bsrlt9XaW/W4bz316kp0nKrliVAR+vl7/f98lzUyKpL6l/ezug8p6m7PKiAz254oEvT/D2/Wler0ALDu/UURGAdcBBd3aJgLLgUn216wWEZe+vXNTVhlJw4P5969MIcjfh3979wjHTtfreHsf6M1MrsVmM2zNLmdRajS+ProE0tv1WtyNMVuBqh6e+i3wQ6D77+S3Aq8aY1qMMSeBXGC2I4I6Q3NbBzuOV3JVWgwrZifyziMLee8fF/FP147jm3OTrI7n8hKHBRMVGqDF3UUcLqmjokGXQKpO/TrxWURuAYqNMQfOu0kiHtjZ7XGRva2n91gJrARITEzsT4wB23mikpZ2G1eN/2LyaeLIcCbqRmF9IiKdm4hpcXcJm7M6t2FePE6Lu+rHhKqIBANPAj/t6eke2nqcbTPGrDHGpBtj0qOjrflm3JxVTpC/D3NSvOPwa2eYmRRJXmUTlXr8nuU2Z5czNSGCKF0CqejfapkxQApwQETygARgr4jE0dlTH9Xt2gSgZKAhnWVzVhnzRg8nyN+lpwVcWte4+94CvZnJSjVNrewrqGaJ9tqV3WUXd2PMIWNMjDEm2RiTTGdBn2GMOQ2sB5aLSKCIpACpwC6HJnaQkxWN5FU2nTMkoy7f5PgIAv18+PhIqdVRvFrXEkhd36669GUp5FpgB5AmIkUict/FrjXGHAZeA44AHwAPG2M6HBXWkbrGJ5eM0x+GgQjy9+XOWaNYt7eIomrd390qm46VMTTYn2mjdAmk6tSX1TIrjDEjjDH+xpgEY8yz5z2fbIyp6PZ4lTFmjDEmzRjzvjNCO8LmrHJGR4eQODzY6ihu76ElY/AR4elNx62O4pVOlDfwtwMl3DRlhC6BVGd55V06Z1o72HGiUnvtDjIiYgjLZ4/i9YxCCvV0pkH3y/eOEuTvy/euHWd1FOVCvLK47zxRSWu7TdcDO9B3l4zFR4TVm3OtjuJVtuWU8/HRMh65eizRYbpKRn3BK4v75qwyhvj7MluXQDpMXEQQK2aP4vWMIu29D5L2Dhs/f/cIicOCuWdBstVxlIvxuuJujGFTVjnzx+gSSEd7SHvvg+qVXQVklzbw4xsn6CHu6gJeV9xPVjRSUNXEEl0C6XDaex88tU1tPLUhm3mjh7N0UqzVcZQL8rrivjmrHEBv9nCSrt777zfmWB3Fo/1uYzZ1Z9r46Zcm6jmpqkdeV9y3ZHcugRw1TJdAOkNcRBD3LEzmjT1FbMkutzqOR8o6Xc9LO/K5c1YiE0boPkiqZ15V3JvbOvj8ZKWeLelk/3TtOFJjQvnhGweobdIDtB3JGMNP3s4kNMiPx5amWR1HuTCvKu4ZedU0t9m4UodknCrI35envjaNyoZW/uVvh62O41He3FvMrrwqHl82nmEhAVbHUS7Mq4r71pxyAnx9mDNal0A625SECB65eixv7Svmg8xTVsfxCLVNbfzyvaNMTxzK19JH9f4C5dW8q7hnl5OeHElwQL+2sVeX6eGrxjIlPoIfv5VJeb1uCTxQv/noGNVNrfziy5Px0W0GVC+8priX1jVz7HS9HmQwiPx9fXjqa1fQ0NLO4+sO0t5hszqS2zpQWMPLnxdw9/xkJo2MsDqOcgNeU9y35XTubbYoNaqXK5UjpcaG8eSNE9h4rIyVL+2hqbXd6khup8Nm+Oe3M4kKDeSfrtP9Y1TfeE1x35pdTlRoIBPidOnYYLt7fjKrbpvM5qwyVqzZSYWe2nRZtuWUc6i4liduGE94kL/VcZSb8IribrMZPs2tYHFqlI5VWuQbc5L437vSySqt5/Y/bievotHqSG7j46OlDPH35cYpI6yOotyIVxT3zJJaqhpbWTROh2SsdN3EWF55YC51Z9q4/Y/bdZK1D4wxfHK0jEWpUboXkrosXlHcvxhv18lUq81IjOSl++ZQ2djK+7pEsldHTtVRUtvMtRN0/xh1ebyiuG/JLmfSyHA9Fd5FTI6PYHR0CB8d1nNXe/PxkTJE0LN+1WXz+OJe39zG3vxq7bW7mOsnxrHzRCW1Z3R7gkvZeKyUaaOG6kEc6rJ5fHHfeaKKdpthsY63u5TrJsbSbjNnDypXFyqta+ZgUa0Oyah+6bW4i8hzIlImIpnd2n4uIgdFZL+IfCQiI7s994SI5IpIlogsdVbwvtqaXU5wgC8zkyKtjqK6mT5qKFGhgXx0RIdmLmbj0c7/+LS4q/7oS8/9BWDZeW2/McZMNcZMA94FfgogIhOB5cAk+2tWi4hlU/w2m2HDkVLmj4nSk2pcjI+PcN3EGLZkldPS3mF1HJe08Wgpo4YNYVxsqNVRlBvqtbgbY81QpA0AABGZSURBVLYCVee11XV7GAIY++e3Aq8aY1qMMSeBXGC2g7Jetl15VZyua+aWaSN7v1gNuusmxtLQ0s6O45VWR3E5Z1o7+DS3gmvGx+phHKpf+j3mLiKrRKQQ+Ab2njsQDxR2u6zI3tbT61eKSIaIZJSXO+dQh3f2lzDE35drJ+hKA1c0f0wUwQG+bNChmQt8mltBS7uN6ybqkIzqn34Xd2PMk8aYUcDLwCP25p66GKaHNowxa4wx6caY9Ohox69kaW238d6hU1w/KVZ3gXRRQf6+XDkumo+PlmKz9fht4rU+PlJKWKAfs5J1e2rVP45YLfMKcLv98yKg+0bTCUCJA77GZduWU07tmTZu1SEZl3bdxFhK61o4WFxrdRSXYbMZNh4r48q0aAL8PH5Bm3KSfn3niEhqt4e3AMfsn68HlotIoIikAKnAroFF7J939pcQGeyv69td3NXjY/D1ETYcOW11FJdxoKiGioYWXSWjBqQvSyHXAjuANBEpEpH7gP8QkUwROQhcDzwKYIw5DLwGHAE+AB42xgz6UojGlnY2HCnlxikj8PfVno8rGxocwOzkYTru3s27B0/h7yssSdOOieq/XgejjTEremh+9hLXrwJWDSTUQH18tJQzbR3cOq3HuVzlYq6fFMu//u0IeRWNJEeFWB3HUs1tHazbW8T1k+IYGqxnpKr+88hu7Tv7SxgZEUS63rjkFrpWhPx5R77FSaz3QeZpapra+PrsRKujKDfnccW9urGVrdnlfGnaSN273U0kRAbzjTmJPL/9JHvyq3p/gQd7ZVcBycODmTd6uNVRlJvzuOL+XuYp2m2GW6/QIRl38sSNExgZMYTH3jhIc5t33rGaW1bPrpNVLJ+dqB0TNWAeV9zf2V9CakwoE0aEWR1FXYbQQD9+dftUTpQ38tuPs62OY4m1uwrx9xXumJlgdRTlATyquFc2tLA7r4qbp47UW7bd0MLUKFbMHsUzW0+wr6Da6jiD6uxE6sQ4PXdAOYRHFfdtORUYA1eN1yVk7urHN04gLjyIH3rZ8MzZidQ5OpGqHMOjivuW7HKGhQQweWSE1VFUP4UF+fPvt08lp6yBP24+bnWcQfPKrgKSdCJVOZDHFHebzbAtp5zFqVE6GeXmrhwXzbUTYnn58wLaO2xWx3G6sxOps3QiVTmOxxT3I6fqqGho5Uq9q88jfDU9gYqGlrOHm3uyV3cV4uejE6nKsTymuG/J7tw2WPeS8QxXpcUQGezPG3uLrI7idBuOlrIoNUrPSVUO5TnFPaucyfHhutLAQwT4+XDrtHg2HCmltslzD9E+WdFIfmUTV43XMweUY3lEca9rbmNPQTVXjtNeuye5fUYCre023j1kya7Rg6LrgPAl47S4K8fyiOK+PbeCDpvhSv0B8SiT48MZFxvKuj2eOzSzOauc0VEhJA4PtjqK8jAeUdy3ZJcTFujH9MShVkdRDiQi3D4jgb0FNZwob7A6jsM1t3Ww80SlLgJQTuH2xd0Yw5aschaMjdK92z3QbdPj8RF4c2+x1VEcbseJSlrabSxJ0984leO5fTXMLWugpLZZez8eKiY8iEWp0by1r9jjzlndklVOkL8Pc1L0nFTleG5f3LuWQC7WyVSPdfvMBIprzrDzRKXVURxqc1YZ80YPJ8jf1+ooygN5RHEfGxNK/NAhVkdRTnL9xFjCgvw8as17XkUjeZVNOiSjnMati/uZ1g4+P1mlSyA9XJC/L7dNj+ed/SV8lusZd6yeXQKpw4nKSdy6uB8oqqG13abF3Qv8cNl4xkSH8PAreymobLI6zoBtzi4nJSqEpOHefWasch63Lu5zRw9n5xPXMGe0Tkh5utBAP9bclY7NZlj5UgaNLe1WR+q35rYOdhyv1F67cqpei7uIPCciZSKS2a3tNyJyTEQOishbIjK023NPiEiuiGSJyFJnBe8SFxFEoJ9OSHmD5KgQ/ufrM8gureexNw5gjHuuntmpSyDVIOhLz/0FYNl5bRuAycaYqUA28ASAiEwElgOT7K9ZLSJaeZXDLB4XzeM3jOe9Q6d5elOu1XH6ZbMugVSDoNfibozZClSd1/aRMabr9+KdQNdepbcCrxpjWowxJ4FcYLYD8yrFA4tGc+u0kfznR9m8trvQ6jiXbWtOuS6BVE7niDH3e4H37Z/HA91/2orsbRcQkZUikiEiGeXl5Q6IobyFiPCr26eyeFw0P3rzoFsV+LrmNk6UN5KerL125VwDKu4i8iTQDrzc1dTDZT0OjBpj1hhj0o0x6dHROrGkLk+Qvy9r7prJolT3KvBHSuoAmDQy3OIkytP1u7iLyN3AzcA3zBczW0XAqG6XJQCeu1+rslRPBb65rYOy+maOlzdwoLCGM62udcj24bPFXc/5Vc7l158Xicgy4EfAlcaY7ouO1wOviMhTwEggFdg14JRKXURXgV/50h5+uO4gP1x38Jznvz4nkV/eNsWidBc6XFxLbHignrqknK7X4i4ia4ElQJSIFAE/o3N1TCCwQUQAdhpjHjTGHBaR14AjdA7XPGyMca2uk/I4XQX+pR35tHbYCA/yI3yIP+v3l7B+fwk/uWkiQwJcY/LycEmd9trVoOi1uBtjVvTQ/Owlrl8FrBpIKKUuV5C/Lw8sHn1OW0xYEBuPlfHh4dN8eXqP8/qDqrmtg9zyBpZOirU6ivICbn2HqlKXMidlGAmRQ1jnIhuOHTtdT4fNMFF77moQaHFXHsvHR/jKjAQ+za2gpOaM1XHILK4FOo8PVMrZtLgrj3b7jHiMgbf2WX+S0+GSWiKG+Ov21GpQaHFXHi1peAizk4exbk+R5XvRHC6pY3J8OPZFCEo5lRZ35fHumJnAiYpG9hXWWJahrcPGsVP1ulJGDRot7srj3TAljiB/H97YY93Eam5ZA60dNr0zVQ0aLe7K44UF+bNsUhx/O1BCc5s1t110TaZqz10NFi3uyivcMXMU9c3tbDhSasnXP1xSR3CALylRevKSGhxa3JVXmDdmOCMigngtw5oNxg6X1DJhRDi+PjqZqgaHFnflFXx9hG/OTWJbTgWHimoH9WvbbIYjJXVM1vF2NYi0uCuv8a15SYQH+fHfn+QM6tfNq2yksbVDx9vVoNLirrxGWJA/9y5M4aMjpWf3VR8MmV3b/OqdqWoQaXFXXuWe+SmEBvrxP5sGr/d+uKQWf18hNSZs0L6mUlrclVeJCPbn2/OTeT/zNNml9YPyNQ8X15EWF0aAn/64qcGj323K69y7MIUh/r78zye5Tv9axhgOl9QyaYSOt6vBpcVdeZ1hIQHcNS+Jdw+WcLy8walfK7u0geqmNiYnaHFXg0uLu/JKDywaTYCfD09vcm7v/ZltJwjy9+GmKSOc+nWUOp8Wd+WVokID+cacJN7ZX0JBZVPvL+iHkpozvL2vmOWzEhkWEuCUr6HUxWhxV17rgUWj8RVhzbbjTnn/Zz89iQHuX5TilPdX6lK0uCuvFRcRxO0zE3gto4iyumaHvndNUytrdxVwyxUjSYgMduh7K9UXvRZ3EXlORMpEJLNb21dF5LCI2EQk/bzrnxCRXBHJEpGlzgitlKM8eOVo2jtsPPvpSYe+75935NPU2sF3rhzd+8VKOUFfeu4vAMvOa8sEvgJs7d4oIhOB5cAk+2tWi4jvwGMq5RxJw0O4eepI/rIzn9qmNoe855nWDl7YnsfV42MYH6d3pSpr9FrcjTFbgarz2o4aY7J6uPxW4FVjTIsx5iSQC8x2SFKlnOShJWNobO3gxR15fbreGENpXTPbcsp57tOTPLP1xDmTsq9lFFLV2MqDV45xTmCl+sDPwe8XD+zs9rjI3qaUy5owIpxrxsfw/GcnuX9RCsEBF/+x+M2Hx3hpRz51ze3ntK967yhTEyK4eeoI/rwjn5lJkcxKjnR2dKUuytHFvafNqns8lVhEVgIrARITEx0cQ6nL892rxnL7H7ezdlch9y3seXVLblk9T286zqLUKK6dEEtqbCjjYsNobuvgvUOnePfgKX753jEAfvalSXoQtrKUo4t7ETCq2+MEoKSnC40xa4A1AOnp6dYeS6+83sykSOakDOOZrSe4a25Sj/vA/GlL5w1Jv7tzGsNDA895buXiMaxcPIb8ykZyShu4ZkLMYEVXqkeOXgq5HlguIoEikgKkArsc/DWUcopHrh7L6bpmntl24oLnut+QdH5h7y5peAjXTozVXruyXF+WQq4FdgBpIlIkIveJyG0iUgTMA/4uIh8CGGMOA68BR4APgIeNMdacSKzUZVqUGs2NU+L4/cacC/ac6Sr4ekOSche9DssYY1Zc5Km3LnL9KmDVQEIpZZV/uWUSn+VW8vi6g/x15Tx8fISqxlZe3VXILdP0hiTlPvQOVaW6iQkL4smbJrA7r5qXdxUA8OL2PM60dfCQLm1UbkSLu1Ln+erMBBaOjeJX7x8jt6yBF3fkcd3EWFJj9SQl5T60uCt1HhHhl7dNocNm+OqftlPT1MZDS7TXrtyLFnelepA4PJjvXz+O6qY25o4exoxEvSFJuRdHr3NXymPcsyCFhpZ2PWhDuSUt7kpdhK+P8L1rx1kdQ6l+0WEZpZTyQFrclVLKA2lxV0opD6TFXSmlPJAWd6WU8kBa3JVSygNpcVdKKQ+kxV0ppTyQGGP9IUgiUg7kD+AtooAKB8VxJs3pWO6SE9wnq+Z0LGfnTDLGRPf0hEsU94ESkQxjTLrVOXqjOR3LXXKC+2TVnI5lZU4dllFKKQ+kxV0ppTyQpxT3NVYH6CPN6VjukhPcJ6vmdCzLcnrEmLtSSqlzeUrPXSmlVDda3JVSygO5dXEXkWUikiUiuSLyuNV5uhOR50SkTEQyu7UNE5ENIpJj/9PSs9tEZJSIbBKRoyJyWEQedcWc9kxBIrJLRA7Ys/6rq2YFEBFfEdknIu/aH7tcThHJE5FDIrJfRDJcNSeAiAwVkTdE5Jj9+3Weq2UVkTT732XXR52IfM+qnG5b3EXEF3gauAGYCKwQkYnWpjrHC8Cy89oeBzYaY1KBjfbHVmoHvm+MmQDMBR62/x26Wk6AFuBqY8wVwDRgmYjMxTWzAjwKHO322FVzXmWMmdZtLbar5vw98IExZjxwBZ1/ty6V1RiTZf+7nAbMBJqAt7AqpzHGLT+AecCH3R4/ATxhda7zMiYDmd0eZwEj7J+PALKsznhe3neA69wgZzCwF5jjilmBBDp/iK8G3nXVf3sgD4g6r80Vc4YDJ7EvAHHlrN2yXQ98ZmVOt+25A/FAYbfHRfY2VxZrjDkFYP8zxuI8Z4lIMjAd+BwXzWkf6tgPlAEbjDGumvV3wA8BW7c2V8xpgI9EZI+IrLS3uWLO0UA58Lx9qOv/RCQE18zaZTmw1v65JTndubhLD226rrMfRCQUWAd8zxhTZ3WeizHGdJjOX3kTgNkiMtnqTOcTkZuBMmPMHquz9MECY8wMOoc2HxaRxVYHugg/YAbwR2PMdKAR1xkuuoCIBAC3AK9bmcOdi3sRMKrb4wSgxKIsfVUqIiMA7H+WWZwHEfGns7C/bIx5097scjm7M8bUAJvpnNNwtawLgFtEJA94FbhaRP6C6+XEGFNi/7OMzrHh2bhgTjp/1ovsv6kBvEFnsXfFrND5n+VeY0yp/bElOd25uO8GUkUkxf4/5XJgvcWZerMeuNv++d10jnFbRkQEeBY4aox5qttTLpUTQESiRWSo/fMhwLXAMVwsqzHmCWNMgjEmmc7vyU+MMd/ExXKKSIiIhHV9TucYcSYulhPAGHMaKBSRNHvTNcARXDCr3Qq+GJIBq3JaPfEwwEmLG4Fs4DjwpNV5zsu2FjgFtNHZ87gPGE7nRFuO/c9hFmdcSOdQ1kFgv/3jRlfLac86Fdhnz5oJ/NTe7nJZu2VewhcTqi6Vk85x7AP2j8NdPz+ulrNb3mlAhv3f/20g0hWz0jnZXwlEdGuzJKduP6CUUh7InYdllFJKXYQWd6WU8kBa3JVSygNpcVdKKQ+kxV0ppTyQFnellPJAWtyVUsoD/T9G/XdG8NtwJwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(Y_test_Sca[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x1919f2893c8>]"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(test_predict[:,0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,4))\n",
    "plt.plot(history.history['loss'], label='Train Loss')\n",
    "plt.plot(history.history['val_loss'], label='Test Loss')\n",
    "plt.title('model loss')\n",
    "plt.ylabel('loss')\n",
    "plt.xlabel('epochs')\n",
    "plt.legend(loc='upper right')\n",
    "plt.show();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "aa=[x for x in range(73)]\n",
    "plt.figure(figsize=(8,4))\n",
    "plt.plot(aa, Y_test_Sca[0][:73], marker='.', label=\"actual\")\n",
    "plt.plot(aa, test_predict[:,0][:73], 'r', label=\"prediction\")\n",
    "# plt.tick_params(left=False, labelleft=True) #remove ticks\n",
    "plt.tight_layout()\n",
    "sns.despine(top=True)\n",
    "plt.subplots_adjust(left=0.07)\n",
    "plt.ylabel('Global Sale', size=15)\n",
    "plt.xlabel('Time step', size=15)\n",
    "plt.legend(fontsize=15)\n",
    "plt.show();"
   ]
  },
  {
   "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": "3.7.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}