LSTM autoencoder-future-value.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 284,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd \n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from collections import Counter\n",
    "import os, pickle\n",
    "import seaborn as sns\n",
    "from imblearn.over_sampling import SMOTE\n",
    "from sklearn.ensemble import RandomForestClassifier, GradientBoostingClassifier, ExtraTreesClassifier\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.externals import joblib\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras import Model ,models, layers, optimizers\n",
    "from tensorflow.keras import callbacks\n",
    "from sklearn import metrics"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Import & Preprocess"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "oz_data = open('./data/onehr.data','r')\n",
    "oz_data = oz_data.readlines()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# to dataframe\n",
    "cols = [i for i in range(0, len(oz_data[0].split(',')))]\n",
    "oz_df = pd.DataFrame(columns=cols)\n",
    "for cnt, rows in enumerate(oz_data):\n",
    "    row = rows.split(',')\n",
    "    oz_df.loc[cnt] = row\n",
    "    \n",
    "oz_df = oz_df.rename(columns={0:'date', 73:'target'})\n",
    "oz_df['target'] = oz_df['target'].apply(lambda x: x.split('.')[0])"
   ]
  },
  {
   "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>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>10</th>\n",
       "      <th>...</th>\n",
       "      <th>64</th>\n",
       "      <th>65</th>\n",
       "      <th>66</th>\n",
       "      <th>67</th>\n",
       "      <th>68</th>\n",
       "      <th>69</th>\n",
       "      <th>70</th>\n",
       "      <th>71</th>\n",
       "      <th>72</th>\n",
       "      <th>target</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1/1/1998</th>\n",
       "      <td>0.8</td>\n",
       "      <td>1.8</td>\n",
       "      <td>2.4</td>\n",
       "      <td>2.1</td>\n",
       "      <td>2</td>\n",
       "      <td>2.1</td>\n",
       "      <td>1.5</td>\n",
       "      <td>1.7</td>\n",
       "      <td>1.9</td>\n",
       "      <td>2.3</td>\n",
       "      <td>...</td>\n",
       "      <td>0.15</td>\n",
       "      <td>10.67</td>\n",
       "      <td>-1.56</td>\n",
       "      <td>5795</td>\n",
       "      <td>-12.1</td>\n",
       "      <td>17.9</td>\n",
       "      <td>10330</td>\n",
       "      <td>-55</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1/2/1998</th>\n",
       "      <td>2.8</td>\n",
       "      <td>3.2</td>\n",
       "      <td>3.3</td>\n",
       "      <td>2.7</td>\n",
       "      <td>3.3</td>\n",
       "      <td>3.2</td>\n",
       "      <td>2.9</td>\n",
       "      <td>2.8</td>\n",
       "      <td>3.1</td>\n",
       "      <td>3.4</td>\n",
       "      <td>...</td>\n",
       "      <td>0.48</td>\n",
       "      <td>8.39</td>\n",
       "      <td>3.84</td>\n",
       "      <td>5805</td>\n",
       "      <td>14.05</td>\n",
       "      <td>29</td>\n",
       "      <td>10275</td>\n",
       "      <td>-55</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1/3/1998</th>\n",
       "      <td>2.9</td>\n",
       "      <td>2.8</td>\n",
       "      <td>2.6</td>\n",
       "      <td>2.1</td>\n",
       "      <td>2.2</td>\n",
       "      <td>2.5</td>\n",
       "      <td>2.5</td>\n",
       "      <td>2.7</td>\n",
       "      <td>2.2</td>\n",
       "      <td>2.5</td>\n",
       "      <td>...</td>\n",
       "      <td>0.6</td>\n",
       "      <td>6.94</td>\n",
       "      <td>9.8</td>\n",
       "      <td>5790</td>\n",
       "      <td>17.9</td>\n",
       "      <td>41.3</td>\n",
       "      <td>10235</td>\n",
       "      <td>-40</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1/4/1998</th>\n",
       "      <td>4.7</td>\n",
       "      <td>3.8</td>\n",
       "      <td>3.7</td>\n",
       "      <td>3.8</td>\n",
       "      <td>2.9</td>\n",
       "      <td>3.1</td>\n",
       "      <td>2.8</td>\n",
       "      <td>2.5</td>\n",
       "      <td>2.4</td>\n",
       "      <td>3.1</td>\n",
       "      <td>...</td>\n",
       "      <td>0.49</td>\n",
       "      <td>8.73</td>\n",
       "      <td>10.54</td>\n",
       "      <td>5775</td>\n",
       "      <td>31.15</td>\n",
       "      <td>51.7</td>\n",
       "      <td>10195</td>\n",
       "      <td>-40</td>\n",
       "      <td>2.08</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1/5/1998</th>\n",
       "      <td>2.6</td>\n",
       "      <td>2.1</td>\n",
       "      <td>1.6</td>\n",
       "      <td>1.4</td>\n",
       "      <td>0.9</td>\n",
       "      <td>1.5</td>\n",
       "      <td>1.2</td>\n",
       "      <td>1.4</td>\n",
       "      <td>1.3</td>\n",
       "      <td>1.4</td>\n",
       "      <td>...</td>\n",
       "      <td>?</td>\n",
       "      <td>?</td>\n",
       "      <td>?</td>\n",
       "      <td>?</td>\n",
       "      <td>?</td>\n",
       "      <td>?</td>\n",
       "      <td>?</td>\n",
       "      <td>?</td>\n",
       "      <td>0.58</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12/27/2004</th>\n",
       "      <td>0.3</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0.3</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.4</td>\n",
       "      <td>1.3</td>\n",
       "      <td>2.2</td>\n",
       "      <td>...</td>\n",
       "      <td>0.07</td>\n",
       "      <td>7.93</td>\n",
       "      <td>-4.41</td>\n",
       "      <td>5800</td>\n",
       "      <td>-25.6</td>\n",
       "      <td>21.8</td>\n",
       "      <td>10295</td>\n",
       "      <td>65</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12/28/2004</th>\n",
       "      <td>1</td>\n",
       "      <td>1.4</td>\n",
       "      <td>1.1</td>\n",
       "      <td>1.7</td>\n",
       "      <td>1.5</td>\n",
       "      <td>1.7</td>\n",
       "      <td>1.8</td>\n",
       "      <td>1.5</td>\n",
       "      <td>2.1</td>\n",
       "      <td>2.4</td>\n",
       "      <td>...</td>\n",
       "      <td>0.04</td>\n",
       "      <td>5.95</td>\n",
       "      <td>-1.14</td>\n",
       "      <td>5845</td>\n",
       "      <td>-19.4</td>\n",
       "      <td>19.1</td>\n",
       "      <td>10310</td>\n",
       "      <td>15</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12/29/2004</th>\n",
       "      <td>0.8</td>\n",
       "      <td>0.8</td>\n",
       "      <td>1.2</td>\n",
       "      <td>0.9</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.6</td>\n",
       "      <td>0.8</td>\n",
       "      <td>1.1</td>\n",
       "      <td>1.5</td>\n",
       "      <td>1.5</td>\n",
       "      <td>...</td>\n",
       "      <td>0.06</td>\n",
       "      <td>7.8</td>\n",
       "      <td>-0.64</td>\n",
       "      <td>5845</td>\n",
       "      <td>-9.6</td>\n",
       "      <td>35.2</td>\n",
       "      <td>10275</td>\n",
       "      <td>-35</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12/30/2004</th>\n",
       "      <td>1.3</td>\n",
       "      <td>0.9</td>\n",
       "      <td>1.5</td>\n",
       "      <td>1.2</td>\n",
       "      <td>1.6</td>\n",
       "      <td>1.8</td>\n",
       "      <td>1.1</td>\n",
       "      <td>1</td>\n",
       "      <td>1.9</td>\n",
       "      <td>2</td>\n",
       "      <td>...</td>\n",
       "      <td>0.25</td>\n",
       "      <td>7.72</td>\n",
       "      <td>-0.89</td>\n",
       "      <td>5845</td>\n",
       "      <td>-19.6</td>\n",
       "      <td>34.2</td>\n",
       "      <td>10245</td>\n",
       "      <td>-30</td>\n",
       "      <td>0.05</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12/31/2004</th>\n",
       "      <td>1.5</td>\n",
       "      <td>1.3</td>\n",
       "      <td>1.8</td>\n",
       "      <td>1.4</td>\n",
       "      <td>1.2</td>\n",
       "      <td>1.7</td>\n",
       "      <td>1.6</td>\n",
       "      <td>1.4</td>\n",
       "      <td>1.6</td>\n",
       "      <td>3</td>\n",
       "      <td>...</td>\n",
       "      <td>0.54</td>\n",
       "      <td>13.07</td>\n",
       "      <td>9.15</td>\n",
       "      <td>5820</td>\n",
       "      <td>1.95</td>\n",
       "      <td>39.35</td>\n",
       "      <td>10220</td>\n",
       "      <td>-25</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2536 rows × 73 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "              1    2    3    4    5    6    7    8    9   10  ...    64  \\\n",
       "date                                                          ...         \n",
       "1/1/1998    0.8  1.8  2.4  2.1    2  2.1  1.5  1.7  1.9  2.3  ...  0.15   \n",
       "1/2/1998    2.8  3.2  3.3  2.7  3.3  3.2  2.9  2.8  3.1  3.4  ...  0.48   \n",
       "1/3/1998    2.9  2.8  2.6  2.1  2.2  2.5  2.5  2.7  2.2  2.5  ...   0.6   \n",
       "1/4/1998    4.7  3.8  3.7  3.8  2.9  3.1  2.8  2.5  2.4  3.1  ...  0.49   \n",
       "1/5/1998    2.6  2.1  1.6  1.4  0.9  1.5  1.2  1.4  1.3  1.4  ...     ?   \n",
       "...         ...  ...  ...  ...  ...  ...  ...  ...  ...  ...  ...   ...   \n",
       "12/27/2004  0.3  0.4  0.5  0.5  0.2  0.3  0.4  0.4  1.3  2.2  ...  0.07   \n",
       "12/28/2004    1  1.4  1.1  1.7  1.5  1.7  1.8  1.5  2.1  2.4  ...  0.04   \n",
       "12/29/2004  0.8  0.8  1.2  0.9  0.4  0.6  0.8  1.1  1.5  1.5  ...  0.06   \n",
       "12/30/2004  1.3  0.9  1.5  1.2  1.6  1.8  1.1    1  1.9    2  ...  0.25   \n",
       "12/31/2004  1.5  1.3  1.8  1.4  1.2  1.7  1.6  1.4  1.6    3  ...  0.54   \n",
       "\n",
       "               65     66    67     68     69     70   71    72 target  \n",
       "date                                                                   \n",
       "1/1/1998    10.67  -1.56  5795  -12.1   17.9  10330  -55     0      0  \n",
       "1/2/1998     8.39   3.84  5805  14.05     29  10275  -55     0      0  \n",
       "1/3/1998     6.94    9.8  5790   17.9   41.3  10235  -40     0      0  \n",
       "1/4/1998     8.73  10.54  5775  31.15   51.7  10195  -40  2.08      0  \n",
       "1/5/1998        ?      ?     ?      ?      ?      ?    ?  0.58      0  \n",
       "...           ...    ...   ...    ...    ...    ...  ...   ...    ...  \n",
       "12/27/2004   7.93  -4.41  5800  -25.6   21.8  10295   65     0      0  \n",
       "12/28/2004   5.95  -1.14  5845  -19.4   19.1  10310   15     0      0  \n",
       "12/29/2004    7.8  -0.64  5845   -9.6   35.2  10275  -35     0      0  \n",
       "12/30/2004   7.72  -0.89  5845  -19.6   34.2  10245  -30  0.05      0  \n",
       "12/31/2004  13.07   9.15  5820   1.95  39.35  10220  -25     0      0  \n",
       "\n",
       "[2536 rows x 73 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "oz_df = oz_df.set_index('date')\n",
    "oz_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1         0.117902\n",
       "2         0.115142\n",
       "3         0.115931\n",
       "4         0.115142\n",
       "5         0.115536\n",
       "            ...   \n",
       "69        0.049290\n",
       "70        0.037461\n",
       "71        0.062697\n",
       "72        0.000789\n",
       "target    0.000000\n",
       "Length: 73, dtype: float64"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# ? >> np.nan\n",
    "for cols in oz_df.columns:\n",
    "    oz_df[cols] = oz_df[cols].replace({'?':np.nan})\n",
    "    \n",
    "# null check\n",
    "oz_df.isnull().sum() / len(oz_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>10</th>\n",
       "      <th>...</th>\n",
       "      <th>64</th>\n",
       "      <th>65</th>\n",
       "      <th>66</th>\n",
       "      <th>67</th>\n",
       "      <th>68</th>\n",
       "      <th>69</th>\n",
       "      <th>70</th>\n",
       "      <th>71</th>\n",
       "      <th>72</th>\n",
       "      <th>target</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1/1/1998</th>\n",
       "      <td>0.8</td>\n",
       "      <td>1.8</td>\n",
       "      <td>2.4</td>\n",
       "      <td>2.1</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.1</td>\n",
       "      <td>1.5</td>\n",
       "      <td>1.7</td>\n",
       "      <td>1.9</td>\n",
       "      <td>2.3</td>\n",
       "      <td>...</td>\n",
       "      <td>0.15</td>\n",
       "      <td>10.67</td>\n",
       "      <td>-1.56</td>\n",
       "      <td>5795.0</td>\n",
       "      <td>-12.10</td>\n",
       "      <td>17.90</td>\n",
       "      <td>10330.0</td>\n",
       "      <td>-55.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1/2/1998</th>\n",
       "      <td>2.8</td>\n",
       "      <td>3.2</td>\n",
       "      <td>3.3</td>\n",
       "      <td>2.7</td>\n",
       "      <td>3.3</td>\n",
       "      <td>3.2</td>\n",
       "      <td>2.9</td>\n",
       "      <td>2.8</td>\n",
       "      <td>3.1</td>\n",
       "      <td>3.4</td>\n",
       "      <td>...</td>\n",
       "      <td>0.48</td>\n",
       "      <td>8.39</td>\n",
       "      <td>3.84</td>\n",
       "      <td>5805.0</td>\n",
       "      <td>14.05</td>\n",
       "      <td>29.00</td>\n",
       "      <td>10275.0</td>\n",
       "      <td>-55.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1/3/1998</th>\n",
       "      <td>2.9</td>\n",
       "      <td>2.8</td>\n",
       "      <td>2.6</td>\n",
       "      <td>2.1</td>\n",
       "      <td>2.2</td>\n",
       "      <td>2.5</td>\n",
       "      <td>2.5</td>\n",
       "      <td>2.7</td>\n",
       "      <td>2.2</td>\n",
       "      <td>2.5</td>\n",
       "      <td>...</td>\n",
       "      <td>0.60</td>\n",
       "      <td>6.94</td>\n",
       "      <td>9.80</td>\n",
       "      <td>5790.0</td>\n",
       "      <td>17.90</td>\n",
       "      <td>41.30</td>\n",
       "      <td>10235.0</td>\n",
       "      <td>-40.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1/4/1998</th>\n",
       "      <td>4.7</td>\n",
       "      <td>3.8</td>\n",
       "      <td>3.7</td>\n",
       "      <td>3.8</td>\n",
       "      <td>2.9</td>\n",
       "      <td>3.1</td>\n",
       "      <td>2.8</td>\n",
       "      <td>2.5</td>\n",
       "      <td>2.4</td>\n",
       "      <td>3.1</td>\n",
       "      <td>...</td>\n",
       "      <td>0.49</td>\n",
       "      <td>8.73</td>\n",
       "      <td>10.54</td>\n",
       "      <td>5775.0</td>\n",
       "      <td>31.15</td>\n",
       "      <td>51.70</td>\n",
       "      <td>10195.0</td>\n",
       "      <td>-40.0</td>\n",
       "      <td>2.08</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1/5/1998</th>\n",
       "      <td>2.6</td>\n",
       "      <td>2.1</td>\n",
       "      <td>1.6</td>\n",
       "      <td>1.4</td>\n",
       "      <td>0.9</td>\n",
       "      <td>1.5</td>\n",
       "      <td>1.2</td>\n",
       "      <td>1.4</td>\n",
       "      <td>1.3</td>\n",
       "      <td>1.4</td>\n",
       "      <td>...</td>\n",
       "      <td>0.49</td>\n",
       "      <td>8.73</td>\n",
       "      <td>10.54</td>\n",
       "      <td>5775.0</td>\n",
       "      <td>31.15</td>\n",
       "      <td>51.70</td>\n",
       "      <td>10195.0</td>\n",
       "      <td>-40.0</td>\n",
       "      <td>0.58</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12/27/2004</th>\n",
       "      <td>0.3</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0.3</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.4</td>\n",
       "      <td>1.3</td>\n",
       "      <td>2.2</td>\n",
       "      <td>...</td>\n",
       "      <td>0.07</td>\n",
       "      <td>7.93</td>\n",
       "      <td>-4.41</td>\n",
       "      <td>5800.0</td>\n",
       "      <td>-25.60</td>\n",
       "      <td>21.80</td>\n",
       "      <td>10295.0</td>\n",
       "      <td>65.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12/28/2004</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.4</td>\n",
       "      <td>1.1</td>\n",
       "      <td>1.7</td>\n",
       "      <td>1.5</td>\n",
       "      <td>1.7</td>\n",
       "      <td>1.8</td>\n",
       "      <td>1.5</td>\n",
       "      <td>2.1</td>\n",
       "      <td>2.4</td>\n",
       "      <td>...</td>\n",
       "      <td>0.04</td>\n",
       "      <td>5.95</td>\n",
       "      <td>-1.14</td>\n",
       "      <td>5845.0</td>\n",
       "      <td>-19.40</td>\n",
       "      <td>19.10</td>\n",
       "      <td>10310.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12/29/2004</th>\n",
       "      <td>0.8</td>\n",
       "      <td>0.8</td>\n",
       "      <td>1.2</td>\n",
       "      <td>0.9</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.6</td>\n",
       "      <td>0.8</td>\n",
       "      <td>1.1</td>\n",
       "      <td>1.5</td>\n",
       "      <td>1.5</td>\n",
       "      <td>...</td>\n",
       "      <td>0.06</td>\n",
       "      <td>7.80</td>\n",
       "      <td>-0.64</td>\n",
       "      <td>5845.0</td>\n",
       "      <td>-9.60</td>\n",
       "      <td>35.20</td>\n",
       "      <td>10275.0</td>\n",
       "      <td>-35.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12/30/2004</th>\n",
       "      <td>1.3</td>\n",
       "      <td>0.9</td>\n",
       "      <td>1.5</td>\n",
       "      <td>1.2</td>\n",
       "      <td>1.6</td>\n",
       "      <td>1.8</td>\n",
       "      <td>1.1</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.9</td>\n",
       "      <td>2.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.25</td>\n",
       "      <td>7.72</td>\n",
       "      <td>-0.89</td>\n",
       "      <td>5845.0</td>\n",
       "      <td>-19.60</td>\n",
       "      <td>34.20</td>\n",
       "      <td>10245.0</td>\n",
       "      <td>-30.0</td>\n",
       "      <td>0.05</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12/31/2004</th>\n",
       "      <td>1.5</td>\n",
       "      <td>1.3</td>\n",
       "      <td>1.8</td>\n",
       "      <td>1.4</td>\n",
       "      <td>1.2</td>\n",
       "      <td>1.7</td>\n",
       "      <td>1.6</td>\n",
       "      <td>1.4</td>\n",
       "      <td>1.6</td>\n",
       "      <td>3.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.54</td>\n",
       "      <td>13.07</td>\n",
       "      <td>9.15</td>\n",
       "      <td>5820.0</td>\n",
       "      <td>1.95</td>\n",
       "      <td>39.35</td>\n",
       "      <td>10220.0</td>\n",
       "      <td>-25.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2536 rows × 73 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "              1    2    3    4    5    6    7    8    9   10  ...    64  \\\n",
       "date                                                          ...         \n",
       "1/1/1998    0.8  1.8  2.4  2.1  2.0  2.1  1.5  1.7  1.9  2.3  ...  0.15   \n",
       "1/2/1998    2.8  3.2  3.3  2.7  3.3  3.2  2.9  2.8  3.1  3.4  ...  0.48   \n",
       "1/3/1998    2.9  2.8  2.6  2.1  2.2  2.5  2.5  2.7  2.2  2.5  ...  0.60   \n",
       "1/4/1998    4.7  3.8  3.7  3.8  2.9  3.1  2.8  2.5  2.4  3.1  ...  0.49   \n",
       "1/5/1998    2.6  2.1  1.6  1.4  0.9  1.5  1.2  1.4  1.3  1.4  ...  0.49   \n",
       "...         ...  ...  ...  ...  ...  ...  ...  ...  ...  ...  ...   ...   \n",
       "12/27/2004  0.3  0.4  0.5  0.5  0.2  0.3  0.4  0.4  1.3  2.2  ...  0.07   \n",
       "12/28/2004  1.0  1.4  1.1  1.7  1.5  1.7  1.8  1.5  2.1  2.4  ...  0.04   \n",
       "12/29/2004  0.8  0.8  1.2  0.9  0.4  0.6  0.8  1.1  1.5  1.5  ...  0.06   \n",
       "12/30/2004  1.3  0.9  1.5  1.2  1.6  1.8  1.1  1.0  1.9  2.0  ...  0.25   \n",
       "12/31/2004  1.5  1.3  1.8  1.4  1.2  1.7  1.6  1.4  1.6  3.0  ...  0.54   \n",
       "\n",
       "               65     66      67     68     69       70    71    72  target  \n",
       "date                                                                         \n",
       "1/1/1998    10.67  -1.56  5795.0 -12.10  17.90  10330.0 -55.0  0.00     0.0  \n",
       "1/2/1998     8.39   3.84  5805.0  14.05  29.00  10275.0 -55.0  0.00     0.0  \n",
       "1/3/1998     6.94   9.80  5790.0  17.90  41.30  10235.0 -40.0  0.00     0.0  \n",
       "1/4/1998     8.73  10.54  5775.0  31.15  51.70  10195.0 -40.0  2.08     0.0  \n",
       "1/5/1998     8.73  10.54  5775.0  31.15  51.70  10195.0 -40.0  0.58     0.0  \n",
       "...           ...    ...     ...    ...    ...      ...   ...   ...     ...  \n",
       "12/27/2004   7.93  -4.41  5800.0 -25.60  21.80  10295.0  65.0  0.00     0.0  \n",
       "12/28/2004   5.95  -1.14  5845.0 -19.40  19.10  10310.0  15.0  0.00     0.0  \n",
       "12/29/2004   7.80  -0.64  5845.0  -9.60  35.20  10275.0 -35.0  0.00     0.0  \n",
       "12/30/2004   7.72  -0.89  5845.0 -19.60  34.20  10245.0 -30.0  0.05     0.0  \n",
       "12/31/2004  13.07   9.15  5820.0   1.95  39.35  10220.0 -25.0  0.00     0.0  \n",
       "\n",
       "[2536 rows x 73 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "for i in oz_df.columns:\n",
    "    oz_df[i] = oz_df[i].apply(lambda x: float(x))\n",
    "\n",
    "# fill NaN\n",
    "oz_df = oz_df.fillna(method='ffill')\n",
    "oz_df = oz_df.fillna(method='bfill')\n",
    "oz_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Counter({0.0: 2463, 1.0: 73})"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Counter(oz_df['target'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Train & Test split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 276,
   "metadata": {},
   "outputs": [],
   "source": [
    "# train & test split\n",
    "train = oz_df.loc['1/1/1998':'1/1/2002'].drop('target', axis=1)\n",
    "test = oz_df.loc['1/1/2002':].drop('target', axis=1)\n",
    "\n",
    "y_train = oz_df.loc['1/1/1998':'1/1/2002']['target']\n",
    "y_test = oz_df.loc['1/1/2002':]['target']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Scale(z-score)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 277,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Shape:  (1451, 72)\n",
      "Test Shape:  (1086, 72)\n"
     ]
    }
   ],
   "source": [
    "# scale\n",
    "scaled_train = (train - train.mean()) / train.std()\n",
    "scaled_test = (test - train.mean()) / train.std()\n",
    "\n",
    "print('Train Shape: ',scaled_train.shape)\n",
    "print('Test Shape: ',scaled_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. Classification model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1.1 Random Forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 278,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 240 candidates, totalling 720 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 12 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done  28 tasks      | elapsed:    0.6s\n",
      "[Parallel(n_jobs=-1)]: Done 328 tasks      | elapsed:    6.2s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RandomForestClassifier(bootstrap=True, class_weight='balanced',\n",
      "                       criterion='gini', max_depth=8, max_features='auto',\n",
      "                       max_leaf_nodes=2, min_impurity_decrease=0.0,\n",
      "                       min_impurity_split=None, min_samples_leaf=2,\n",
      "                       min_samples_split=2, min_weight_fraction_leaf=0.0,\n",
      "                       n_estimators=10, n_jobs=None, oob_score=False,\n",
      "                       random_state=None, verbose=0, warm_start=False)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done 720 out of 720 | elapsed:   13.9s finished\n"
     ]
    }
   ],
   "source": [
    "rf_params = {\n",
    "    'n_estimators': [10,50,100],\n",
    "    'max_depth': [i for i in range(4,13,2)],\n",
    "    'min_samples_leaf': [1,2,3,4],\n",
    "    'max_leaf_nodes': [i for i in range(2,6,1)]\n",
    "}\n",
    "rf_clf = RandomForestClassifier(class_weight='balanced')\n",
    "rf_grid = GridSearchCV(estimator=rf_clf, param_grid=rf_params, cv=3, \n",
    "                       scoring='recall', n_jobs=-1, verbose=1)\n",
    "\n",
    "rf_grid.fit(scaled_train, y_train)\n",
    "print(rf_grid.best_estimator_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 279,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Acc:  0.9016393442622951\n",
      "Train Recall:  0.11066398390342053\n",
      "Test Acc:  0.75\n",
      "Test Recall:  0.75\n"
     ]
    }
   ],
   "source": [
    "pred_train = rf_grid.predict(scaled_train)\n",
    "print('Train Acc: ', rf_grid.score(scaled_train, y_train))\n",
    "print('Train Recall: ', metrics.recall_score(y_train, pred_train))\n",
    "\n",
    "pred_test = rf_grid.predict(scaled_test)\n",
    "print('Test Acc: ', rf_grid.score(scaled_test, y_test))\n",
    "print('Test Recall: ',metrics.recall_score(y_test, pred_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 280,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       1.00      0.75      0.85      1074\n",
      "         1.0       0.03      0.75      0.06        12\n",
      "\n",
      "    accuracy                           0.75      1086\n",
      "   macro avg       0.51      0.75      0.46      1086\n",
      "weighted avg       0.99      0.75      0.85      1086\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 504x504 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "print(metrics.classification_report(y_test, pred_test))\n",
    "\n",
    "conf_matrix = metrics.confusion_matrix(y_test, pred_test)\n",
    "plt.figure(figsize=(7, 7))\n",
    "sns.heatmap(conf_matrix, xticklabels=[0, 1], yticklabels=[0, 1], annot=True, fmt='d')\n",
    "plt.title('Confusion Matrix')\n",
    "plt.xlabel('Predicted Class'); plt.ylabel('True Class')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2.2 Extra Trees"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 281,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 12 concurrent workers.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 240 candidates, totalling 720 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done  28 tasks      | elapsed:    0.3s\n",
      "[Parallel(n_jobs=-1)]: Done 328 tasks      | elapsed:    4.4s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ExtraTreesClassifier(bootstrap=False, class_weight='balanced', criterion='gini',\n",
      "                     max_depth=12, max_features='auto', max_leaf_nodes=2,\n",
      "                     min_impurity_decrease=0.0, min_impurity_split=None,\n",
      "                     min_samples_leaf=1, min_samples_split=2,\n",
      "                     min_weight_fraction_leaf=0.0, n_estimators=10, n_jobs=None,\n",
      "                     oob_score=False, random_state=None, verbose=0,\n",
      "                     warm_start=False)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done 720 out of 720 | elapsed:    9.7s finished\n"
     ]
    }
   ],
   "source": [
    "ext_params = {\n",
    "    'n_estimators': [10,50,100],\n",
    "    'max_depth': [i for i in range(4,13,2)],\n",
    "    'min_samples_leaf': [1,2,3,4],\n",
    "    'max_leaf_nodes': [i for i in range(2,6,1)]\n",
    "}\n",
    "ext_clf = ExtraTreesClassifier(class_weight='balanced')\n",
    "ext_grid = GridSearchCV(estimator=ext_clf, param_grid=ext_params, cv=3,\n",
    "                       scoring='recall', n_jobs=-1, verbose=1)\n",
    "\n",
    "ext_grid.fit(scaled_train, y_train)\n",
    "print(ext_grid.best_estimator_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 282,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Acc:  0.9344262295081968\n",
      "Train Recall:  0.0967741935483871\n",
      "Test Acc:  0.8333333333333334\n",
      "Test Recall:  0.8333333333333334\n"
     ]
    }
   ],
   "source": [
    "pred_train = ext_grid.predict(scaled_train)\n",
    "print('Train Acc: ', ext_grid.score(scaled_train, y_train))\n",
    "print('Train Recall: ', metrics.recall_score(y_train, pred_train))\n",
    "\n",
    "pred_test = ext_grid.predict(scaled_test)\n",
    "print('Test Acc: ', ext_grid.score(scaled_test, y_test))\n",
    "print('Test Recall: ',metrics.recall_score(y_test, pred_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 283,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       1.00      0.66      0.79      1074\n",
      "         1.0       0.03      0.83      0.05        12\n",
      "\n",
      "    accuracy                           0.66      1086\n",
      "   macro avg       0.51      0.75      0.42      1086\n",
      "weighted avg       0.99      0.66      0.78      1086\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 504x504 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "print(metrics.classification_report(y_test, pred_test))\n",
    "\n",
    "conf_matrix = metrics.confusion_matrix(y_test, pred_test)\n",
    "plt.figure(figsize=(7, 7))\n",
    "sns.heatmap(conf_matrix, xticklabels=[0, 1], yticklabels=[0, 1], annot=True, fmt='d')\n",
    "plt.title('Confusion Matrix')\n",
    "plt.xlabel('Predicted Class'); plt.ylabel('True Class')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 286,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['./model_t50/ExtraTreesGrid.pkl']"
      ]
     },
     "execution_count": 286,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# save model \n",
    "ext_grid_pickle = pickle.dumps(ext_grid)\n",
    "\n",
    "joblib.dump(ext_grid, './model_t50/ExtraTreesGrid.pkl')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. LSTM Autoencoder for create future value"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''\n",
    "reshape [samples, timesteps, features]\n",
    "[len(data.shape[0], 5, len(data.shape[1])]\n",
    "1-6    6-11    11-16\n",
    "2-7    7-12    12-17\n",
    "3-8    8-13    13-18\n",
    "4-9    9-14    14-19\n",
    "5-10  10-15    15-20\n",
    "'''\n",
    "def getSeriesData_ae(data, timesteps):\n",
    "    n_pairs = len(data) - timesteps\n",
    "    ec_seqences = [data[i : i+timesteps] for i in range(0, n_pairs, timesteps)]\n",
    "    dc_seqences = [data[i+timesteps : i+(timesteps*2)] for i in range(0, n_pairs-timesteps, timesteps)]\n",
    "    \n",
    "    return np.array(ec_seqences), np.array(dc_seqences)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "timesteps = 5\n",
    "input_dim = scaled_train.shape[1]\n",
    "\n",
    "train_ec, train_dc = getSeriesData_ae(scaled_train.values, timesteps)\n",
    "test_ec, test_dc = getSeriesData_ae(scaled_test.values, timesteps)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(433, 5, 72)\n",
      "(433, 5, 72)\n",
      "(72, 5, 72)\n",
      "(72, 5, 72)\n"
     ]
    }
   ],
   "source": [
    "train_ec = train_ec[:-1] # train_dc 길이를 맞춰주기 위해 마지막 시퀀스 제거 \n",
    "print(train_ec.shape)\n",
    "print(train_dc.shape)\n",
    "\n",
    "test_ec = test_ec[:-1] # test_dc 길이를 맞춰주기 위해 마지막 시퀀스 제거 \n",
    "print(test_ec.shape)\n",
    "print(test_dc.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_2 (InputLayer)         [(None, 5, 72)]           0         \n",
      "_________________________________________________________________\n",
      "lstm_2 (LSTM)                (None, 30)                12360     \n",
      "_________________________________________________________________\n",
      "repeat_vector (RepeatVector) (None, 5, 30)             0         \n",
      "_________________________________________________________________\n",
      "lstm_3 (LSTM)                (None, 5, 30)             7320      \n",
      "_________________________________________________________________\n",
      "time_distributed (TimeDistri (None, 5, 72)             2232      \n",
      "=================================================================\n",
      "Total params: 21,912\n",
      "Trainable params: 21,912\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "# define model \n",
    "input_layer = layers.Input(shape=(timesteps, input_dim))\n",
    "encoder = layers.LSTM(30, activation='relu')(input_layer)\n",
    "decoder = layers.RepeatVector(timesteps)(encoder)\n",
    "decoder = layers.LSTM(30, activation='relu', return_sequences=True)(decoder)\n",
    "decoder = layers.TimeDistributed(layers.Dense(input_dim))(decoder)\n",
    "\n",
    "model = Model(inputs=input_layer, outputs=decoder)\n",
    "\n",
    "# compile\n",
    "model.compile(optimizer='adam', loss='mse', metrics=['mse'])\n",
    "print(model.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 433 samples\n",
      "Epoch 1/500\n",
      "433/433 [==============================] - 2s 4ms/sample - loss: 0.9822 - mse: 0.9822\n",
      "Epoch 2/500\n",
      "433/433 [==============================] - 0s 364us/sample - loss: 0.9238 - mse: 0.9238\n",
      "Epoch 3/500\n",
      "433/433 [==============================] - 0s 362us/sample - loss: 0.8552 - mse: 0.8552\n",
      "Epoch 4/500\n",
      "433/433 [==============================] - 0s 364us/sample - loss: 0.7903 - mse: 0.7903\n",
      "Epoch 5/500\n",
      "433/433 [==============================] - 0s 369us/sample - loss: 0.7389 - mse: 0.7389\n",
      "Epoch 6/500\n",
      "433/433 [==============================] - 0s 359us/sample - loss: 0.7057 - mse: 0.7057\n",
      "Epoch 7/500\n",
      "433/433 [==============================] - 0s 369us/sample - loss: 0.6820 - mse: 0.6820\n",
      "Epoch 8/500\n",
      "433/433 [==============================] - 0s 362us/sample - loss: 0.6681 - mse: 0.6681\n",
      "Epoch 9/500\n",
      "433/433 [==============================] - 0s 368us/sample - loss: 0.6573 - mse: 0.6573\n",
      "Epoch 10/500\n",
      "433/433 [==============================] - 0s 364us/sample - loss: 0.6459 - mse: 0.6459\n",
      "Epoch 11/500\n",
      "433/433 [==============================] - 0s 366us/sample - loss: 0.6370 - mse: 0.6370\n",
      "Epoch 12/500\n",
      "433/433 [==============================] - 0s 362us/sample - loss: 0.6285 - mse: 0.6285\n",
      "Epoch 13/500\n",
      "433/433 [==============================] - 0s 371us/sample - loss: 0.6205 - mse: 0.6205\n",
      "Epoch 14/500\n",
      "433/433 [==============================] - 0s 362us/sample - loss: 0.6129 - mse: 0.6129\n",
      "Epoch 15/500\n",
      "433/433 [==============================] - 0s 371us/sample - loss: 0.6044 - mse: 0.6044\n",
      "Epoch 16/500\n",
      "433/433 [==============================] - 0s 378us/sample - loss: 0.5970 - mse: 0.5970\n",
      "Epoch 17/500\n",
      "433/433 [==============================] - 0s 366us/sample - loss: 0.5906 - mse: 0.5906\n",
      "Epoch 18/500\n",
      "433/433 [==============================] - 0s 371us/sample - loss: 0.5887 - mse: 0.5887\n",
      "Epoch 19/500\n",
      "433/433 [==============================] - 0s 366us/sample - loss: 0.5789 - mse: 0.5789\n",
      "Epoch 20/500\n",
      "433/433 [==============================] - 0s 362us/sample - loss: 0.5700 - mse: 0.5700\n",
      "Epoch 21/500\n",
      "433/433 [==============================] - 0s 364us/sample - loss: 0.5628 - mse: 0.5628\n",
      "Epoch 22/500\n",
      "433/433 [==============================] - 0s 371us/sample - loss: 0.5558 - mse: 0.5558\n",
      "Epoch 23/500\n",
      "433/433 [==============================] - 0s 359us/sample - loss: 0.5483 - mse: 0.5483\n",
      "Epoch 24/500\n",
      "433/433 [==============================] - 0s 362us/sample - loss: 0.5443 - mse: 0.5443\n",
      "Epoch 25/500\n",
      "433/433 [==============================] - 0s 366us/sample - loss: 0.5414 - mse: 0.5414\n",
      "Epoch 26/500\n",
      "433/433 [==============================] - 0s 364us/sample - loss: 0.5296 - mse: 0.5296\n",
      "Epoch 27/500\n",
      "433/433 [==============================] - 0s 366us/sample - loss: 0.5209 - mse: 0.5209\n",
      "Epoch 28/500\n",
      "433/433 [==============================] - 0s 366us/sample - loss: 0.5130 - mse: 0.5130\n",
      "Epoch 29/500\n",
      "433/433 [==============================] - 0s 362us/sample - loss: 0.5105 - mse: 0.5105\n",
      "Epoch 30/500\n",
      "433/433 [==============================] - 0s 364us/sample - loss: 0.5057 - mse: 0.5057\n",
      "Epoch 31/500\n",
      "433/433 [==============================] - 0s 366us/sample - loss: 0.4986 - mse: 0.4986\n",
      "Epoch 32/500\n",
      "433/433 [==============================] - 0s 364us/sample - loss: 0.4929 - mse: 0.4929\n",
      "Epoch 33/500\n",
      "433/433 [==============================] - 0s 364us/sample - loss: 0.4896 - mse: 0.4896\n",
      "Epoch 34/500\n",
      "433/433 [==============================] - 0s 359us/sample - loss: 0.4839 - mse: 0.4839\n",
      "Epoch 35/500\n",
      "433/433 [==============================] - 0s 364us/sample - loss: 0.4795 - mse: 0.4795\n",
      "Epoch 36/500\n",
      "433/433 [==============================] - 0s 366us/sample - loss: 0.4750 - mse: 0.4750\n",
      "Epoch 37/500\n",
      "433/433 [==============================] - 0s 366us/sample - loss: 0.4717 - mse: 0.4717\n",
      "Epoch 38/500\n",
      "433/433 [==============================] - 0s 364us/sample - loss: 0.4701 - mse: 0.4701\n",
      "Epoch 39/500\n",
      "433/433 [==============================] - 0s 364us/sample - loss: 0.4653 - mse: 0.4653\n",
      "Epoch 40/500\n",
      "433/433 [==============================] - 0s 364us/sample - loss: 0.4609 - mse: 0.4609\n",
      "Epoch 41/500\n",
      "433/433 [==============================] - 0s 366us/sample - loss: 0.4585 - mse: 0.4585\n",
      "Epoch 42/500\n",
      "433/433 [==============================] - 0s 366us/sample - loss: 0.4543 - mse: 0.4543\n",
      "Epoch 43/500\n",
      "433/433 [==============================] - 0s 369us/sample - loss: 0.4502 - mse: 0.4502\n",
      "Epoch 44/500\n",
      "433/433 [==============================] - 0s 375us/sample - loss: 0.4504 - mse: 0.4504\n",
      "Epoch 45/500\n",
      "433/433 [==============================] - 0s 369us/sample - loss: 0.4464 - mse: 0.4464\n",
      "Epoch 46/500\n",
      "433/433 [==============================] - 0s 359us/sample - loss: 0.4426 - mse: 0.4426\n",
      "Epoch 47/500\n",
      "433/433 [==============================] - 0s 366us/sample - loss: 0.4427 - mse: 0.4427\n",
      "Epoch 48/500\n",
      "433/433 [==============================] - 0s 371us/sample - loss: 0.4407 - mse: 0.4407\n",
      "Epoch 49/500\n",
      "433/433 [==============================] - 0s 369us/sample - loss: 0.4395 - mse: 0.4395\n",
      "Epoch 50/500\n",
      "433/433 [==============================] - 0s 373us/sample - loss: 0.4337 - mse: 0.4337\n",
      "Epoch 51/500\n",
      "433/433 [==============================] - 0s 373us/sample - loss: 0.4301 - mse: 0.4301\n",
      "Epoch 52/500\n",
      "433/433 [==============================] - 0s 364us/sample - loss: 0.4274 - mse: 0.4274\n",
      "Epoch 53/500\n",
      "433/433 [==============================] - 0s 368us/sample - loss: 0.4237 - mse: 0.4237\n",
      "Epoch 54/500\n",
      "433/433 [==============================] - 0s 369us/sample - loss: 0.4213 - mse: 0.4213\n",
      "Epoch 55/500\n",
      "433/433 [==============================] - 0s 359us/sample - loss: 0.4223 - mse: 0.4223\n",
      "Epoch 56/500\n",
      "433/433 [==============================] - 0s 364us/sample - loss: 0.4197 - mse: 0.4197\n",
      "Epoch 57/500\n",
      "433/433 [==============================] - 0s 366us/sample - loss: 0.4185 - mse: 0.4185\n",
      "Epoch 58/500\n",
      "433/433 [==============================] - 0s 359us/sample - loss: 0.4156 - mse: 0.4156\n",
      "Epoch 59/500\n",
      "433/433 [==============================] - 0s 366us/sample - loss: 0.4168 - mse: 0.4168\n",
      "Epoch 60/500\n",
      "433/433 [==============================] - 0s 369us/sample - loss: 0.4165 - mse: 0.4165\n",
      "Epoch 61/500\n",
      "433/433 [==============================] - 0s 408us/sample - loss: 0.4082 - mse: 0.4082\n",
      "Epoch 62/500\n",
      "433/433 [==============================] - 0s 415us/sample - loss: 0.4039 - mse: 0.4039\n",
      "Epoch 63/500\n",
      "433/433 [==============================] - 0s 366us/sample - loss: 0.4006 - mse: 0.4006\n",
      "Epoch 64/500\n",
      "433/433 [==============================] - 0s 359us/sample - loss: 0.3989 - mse: 0.3989\n",
      "Epoch 65/500\n",
      "433/433 [==============================] - 0s 362us/sample - loss: 0.3967 - mse: 0.3967\n",
      "Epoch 66/500\n",
      "433/433 [==============================] - 0s 364us/sample - loss: 0.3947 - mse: 0.3947\n",
      "Epoch 67/500\n",
      "433/433 [==============================] - 0s 364us/sample - loss: 0.3927 - mse: 0.3927\n",
      "Epoch 68/500\n",
      "433/433 [==============================] - 0s 362us/sample - loss: 0.3927 - mse: 0.3927\n",
      "Epoch 69/500\n",
      "433/433 [==============================] - 0s 373us/sample - loss: 0.3904 - mse: 0.3904\n",
      "Epoch 70/500\n",
      "433/433 [==============================] - 0s 359us/sample - loss: 0.3892 - mse: 0.3892\n",
      "Epoch 71/500\n",
      "433/433 [==============================] - 0s 371us/sample - loss: 0.3883 - mse: 0.3883\n",
      "Epoch 72/500\n",
      "433/433 [==============================] - 0s 364us/sample - loss: 0.3880 - mse: 0.3880\n",
      "Epoch 73/500\n",
      "433/433 [==============================] - 0s 369us/sample - loss: 0.3858 - mse: 0.3858\n",
      "Epoch 74/500\n",
      "433/433 [==============================] - 0s 366us/sample - loss: 0.3828 - mse: 0.3828\n",
      "Epoch 75/500\n",
      "433/433 [==============================] - 0s 371us/sample - loss: 0.3813 - mse: 0.3813\n",
      "Epoch 76/500\n",
      "433/433 [==============================] - 0s 364us/sample - loss: 0.3783 - mse: 0.3783\n",
      "Epoch 77/500\n",
      "433/433 [==============================] - 0s 362us/sample - loss: 0.3768 - mse: 0.3768\n",
      "Epoch 78/500\n",
      "433/433 [==============================] - 0s 362us/sample - loss: 0.3743 - mse: 0.3743\n",
      "Epoch 79/500\n",
      "433/433 [==============================] - 0s 364us/sample - loss: 0.3723 - mse: 0.3723\n",
      "Epoch 80/500\n",
      "433/433 [==============================] - 0s 366us/sample - loss: 0.3716 - mse: 0.3716\n",
      "Epoch 81/500\n",
      "433/433 [==============================] - 0s 364us/sample - loss: 0.3684 - mse: 0.3684\n",
      "Epoch 82/500\n",
      "433/433 [==============================] - 0s 366us/sample - loss: 0.3682 - mse: 0.3682\n",
      "Epoch 83/500\n",
      "433/433 [==============================] - 0s 373us/sample - loss: 0.3676 - mse: 0.3676\n",
      "Epoch 84/500\n",
      "433/433 [==============================] - 0s 366us/sample - loss: 0.3666 - mse: 0.3666\n",
      "Epoch 85/500\n",
      "433/433 [==============================] - 0s 364us/sample - loss: 0.3683 - mse: 0.3683\n",
      "Epoch 86/500\n",
      "433/433 [==============================] - 0s 369us/sample - loss: 0.3735 - mse: 0.3735\n",
      "Epoch 87/500\n",
      "433/433 [==============================] - 0s 368us/sample - loss: 0.3724 - mse: 0.3724\n",
      "Epoch 88/500\n",
      "433/433 [==============================] - 0s 382us/sample - loss: 0.3658 - mse: 0.3658\n",
      "Epoch 89/500\n",
      "433/433 [==============================] - 0s 373us/sample - loss: 0.3653 - mse: 0.3653\n",
      "Epoch 90/500\n",
      "433/433 [==============================] - 0s 371us/sample - loss: 0.3660 - mse: 0.3660\n",
      "Epoch 91/500\n",
      "433/433 [==============================] - 0s 373us/sample - loss: 0.3614 - mse: 0.3614\n",
      "Epoch 92/500\n",
      "433/433 [==============================] - 0s 373us/sample - loss: 0.3612 - mse: 0.3612\n",
      "Epoch 93/500\n",
      "433/433 [==============================] - 0s 368us/sample - loss: 0.3576 - mse: 0.3576\n",
      "Epoch 94/500\n",
      "433/433 [==============================] - 0s 373us/sample - loss: 0.3532 - mse: 0.3532\n",
      "Epoch 95/500\n",
      "433/433 [==============================] - 0s 366us/sample - loss: 0.3495 - mse: 0.3495\n",
      "Epoch 96/500\n",
      "433/433 [==============================] - 0s 371us/sample - loss: 0.3474 - mse: 0.3474\n",
      "Epoch 97/500\n",
      "433/433 [==============================] - 0s 369us/sample - loss: 0.3463 - mse: 0.3463\n",
      "Epoch 98/500\n",
      "433/433 [==============================] - 0s 366us/sample - loss: 0.3434 - mse: 0.3434\n",
      "Epoch 99/500\n",
      "433/433 [==============================] - 0s 366us/sample - loss: 0.3441 - mse: 0.3441\n",
      "Epoch 100/500\n",
      "433/433 [==============================] - 0s 369us/sample - loss: 0.3450 - mse: 0.3450\n",
      "Epoch 101/500\n",
      "433/433 [==============================] - 0s 364us/sample - loss: 0.3430 - mse: 0.3430\n",
      "Epoch 102/500\n",
      "433/433 [==============================] - 0s 368us/sample - loss: 0.3448 - mse: 0.3448\n",
      "Epoch 103/500\n",
      "433/433 [==============================] - 0s 373us/sample - loss: 0.3458 - mse: 0.3458\n",
      "Epoch 104/500\n",
      "433/433 [==============================] - 0s 369us/sample - loss: 0.3441 - mse: 0.3441\n",
      "Epoch 105/500\n",
      "433/433 [==============================] - 0s 378us/sample - loss: 0.3390 - mse: 0.3390\n",
      "Epoch 106/500\n",
      "433/433 [==============================] - 0s 373us/sample - loss: 0.3361 - mse: 0.3361\n",
      "Epoch 107/500\n",
      "433/433 [==============================] - 0s 378us/sample - loss: 0.3348 - mse: 0.3348\n",
      "Epoch 108/500\n",
      "433/433 [==============================] - 0s 373us/sample - loss: 0.3328 - mse: 0.3328\n",
      "Epoch 109/500\n",
      "433/433 [==============================] - 0s 368us/sample - loss: 0.3345 - mse: 0.3345\n",
      "Epoch 110/500\n",
      "433/433 [==============================] - 0s 373us/sample - loss: 0.3348 - mse: 0.3348\n",
      "Epoch 111/500\n",
      "433/433 [==============================] - 0s 412us/sample - loss: 0.3326 - mse: 0.3326\n",
      "Epoch 112/500\n",
      "433/433 [==============================] - 0s 375us/sample - loss: 0.3306 - mse: 0.3306\n",
      "Epoch 113/500\n",
      "433/433 [==============================] - 0s 394us/sample - loss: 0.3280 - mse: 0.3280\n",
      "Epoch 114/500\n",
      "433/433 [==============================] - 0s 385us/sample - loss: 0.3261 - mse: 0.3261\n",
      "Epoch 115/500\n",
      "433/433 [==============================] - 0s 372us/sample - loss: 0.3235 - mse: 0.3235\n",
      "Epoch 116/500\n",
      "433/433 [==============================] - 0s 381us/sample - loss: 0.3227 - mse: 0.3227\n",
      "Epoch 117/500\n",
      "433/433 [==============================] - 0s 368us/sample - loss: 0.3235 - mse: 0.3235\n",
      "Epoch 118/500\n",
      "433/433 [==============================] - 0s 369us/sample - loss: 0.3220 - mse: 0.3220\n",
      "Epoch 119/500\n",
      "433/433 [==============================] - 0s 371us/sample - loss: 0.3204 - mse: 0.3204\n",
      "Epoch 120/500\n",
      "433/433 [==============================] - 0s 387us/sample - loss: 0.3189 - mse: 0.3189\n",
      "Epoch 121/500\n",
      "433/433 [==============================] - 0s 366us/sample - loss: 0.3208 - mse: 0.3208\n",
      "Epoch 122/500\n",
      "433/433 [==============================] - 0s 375us/sample - loss: 0.3180 - mse: 0.3180\n",
      "Epoch 123/500\n",
      "433/433 [==============================] - 0s 387us/sample - loss: 0.3164 - mse: 0.3164\n",
      "Epoch 124/500\n",
      "433/433 [==============================] - 0s 382us/sample - loss: 0.3151 - mse: 0.3151\n",
      "Epoch 125/500\n",
      "433/433 [==============================] - 0s 372us/sample - loss: 0.3136 - mse: 0.3136\n",
      "Epoch 126/500\n",
      "433/433 [==============================] - 0s 387us/sample - loss: 0.3131 - mse: 0.3131\n",
      "Epoch 127/500\n",
      "433/433 [==============================] - 0s 380us/sample - loss: 0.3139 - mse: 0.3139\n",
      "Epoch 128/500\n",
      "433/433 [==============================] - 0s 382us/sample - loss: 0.3123 - mse: 0.3123\n",
      "Epoch 129/500\n",
      "433/433 [==============================] - 0s 389us/sample - loss: 0.3111 - mse: 0.3111\n",
      "Epoch 130/500\n",
      "433/433 [==============================] - 0s 380us/sample - loss: 0.3126 - mse: 0.3126\n",
      "Epoch 131/500\n",
      "433/433 [==============================] - 0s 387us/sample - loss: 0.3120 - mse: 0.3120\n",
      "Epoch 132/500\n",
      "433/433 [==============================] - 0s 371us/sample - loss: 0.3094 - mse: 0.3094\n",
      "Epoch 133/500\n",
      "433/433 [==============================] - 0s 383us/sample - loss: 0.3081 - mse: 0.3081\n",
      "Epoch 134/500\n",
      "433/433 [==============================] - 0s 373us/sample - loss: 0.3066 - mse: 0.3066\n",
      "Epoch 135/500\n",
      "433/433 [==============================] - 0s 373us/sample - loss: 0.3049 - mse: 0.3049\n",
      "Epoch 136/500\n",
      "433/433 [==============================] - 0s 373us/sample - loss: 0.3052 - mse: 0.3052\n",
      "Epoch 137/500\n",
      "433/433 [==============================] - 0s 378us/sample - loss: 0.3047 - mse: 0.3047\n",
      "Epoch 138/500\n",
      "433/433 [==============================] - 0s 373us/sample - loss: 0.3071 - mse: 0.3071\n",
      "Epoch 139/500\n",
      "433/433 [==============================] - 0s 378us/sample - loss: 0.3062 - mse: 0.3062\n",
      "Epoch 140/500\n",
      "433/433 [==============================] - 0s 375us/sample - loss: 0.3032 - mse: 0.3032\n",
      "Epoch 141/500\n",
      "433/433 [==============================] - 0s 382us/sample - loss: 0.3006 - mse: 0.3006\n",
      "Epoch 142/500\n",
      "433/433 [==============================] - 0s 375us/sample - loss: 0.3004 - mse: 0.3004\n",
      "Epoch 143/500\n",
      "433/433 [==============================] - 0s 371us/sample - loss: 0.3011 - mse: 0.3011\n",
      "Epoch 144/500\n",
      "433/433 [==============================] - 0s 385us/sample - loss: 0.3001 - mse: 0.3001\n",
      "Epoch 145/500\n",
      "433/433 [==============================] - 0s 375us/sample - loss: 0.2967 - mse: 0.2967\n",
      "Epoch 146/500\n",
      "433/433 [==============================] - 0s 375us/sample - loss: 0.2946 - mse: 0.2946\n",
      "Epoch 147/500\n",
      "433/433 [==============================] - 0s 378us/sample - loss: 0.2933 - mse: 0.2933\n",
      "Epoch 148/500\n",
      "433/433 [==============================] - 0s 373us/sample - loss: 0.2973 - mse: 0.2973\n",
      "Epoch 149/500\n",
      "433/433 [==============================] - 0s 373us/sample - loss: 0.2958 - mse: 0.2958\n",
      "Epoch 150/500\n",
      "433/433 [==============================] - 0s 373us/sample - loss: 0.2930 - mse: 0.2930\n",
      "Epoch 151/500\n",
      "433/433 [==============================] - 0s 366us/sample - loss: 0.2910 - mse: 0.2910\n",
      "Epoch 152/500\n",
      "433/433 [==============================] - 0s 387us/sample - loss: 0.2908 - mse: 0.2908\n",
      "Epoch 153/500\n",
      "433/433 [==============================] - 0s 387us/sample - loss: 0.2920 - mse: 0.2920\n",
      "Epoch 154/500\n",
      "433/433 [==============================] - 0s 378us/sample - loss: 0.2901 - mse: 0.2901\n",
      "Epoch 155/500\n",
      "433/433 [==============================] - 0s 378us/sample - loss: 0.2877 - mse: 0.2877\n",
      "Epoch 156/500\n",
      "433/433 [==============================] - 0s 378us/sample - loss: 0.2860 - mse: 0.2860\n",
      "Epoch 157/500\n",
      "433/433 [==============================] - 0s 380us/sample - loss: 0.2850 - mse: 0.2850\n",
      "Epoch 158/500\n",
      "433/433 [==============================] - 0s 380us/sample - loss: 0.2836 - mse: 0.2836\n",
      "Epoch 159/500\n",
      "433/433 [==============================] - 0s 380us/sample - loss: 0.2858 - mse: 0.2858\n",
      "Epoch 160/500\n",
      "433/433 [==============================] - 0s 369us/sample - loss: 0.2876 - mse: 0.2876\n",
      "Epoch 161/500\n",
      "433/433 [==============================] - 0s 378us/sample - loss: 0.2881 - mse: 0.2881\n",
      "Epoch 162/500\n",
      "433/433 [==============================] - 0s 375us/sample - loss: 0.2867 - mse: 0.2867\n",
      "Epoch 163/500\n",
      "433/433 [==============================] - 0s 375us/sample - loss: 0.2877 - mse: 0.2877\n",
      "Epoch 164/500\n",
      "433/433 [==============================] - 0s 373us/sample - loss: 0.2863 - mse: 0.2863\n",
      "Epoch 165/500\n",
      "433/433 [==============================] - 0s 389us/sample - loss: 0.2842 - mse: 0.2842\n",
      "Epoch 166/500\n",
      "433/433 [==============================] - 0s 378us/sample - loss: 0.2817 - mse: 0.2817\n",
      "Epoch 167/500\n",
      "433/433 [==============================] - 0s 369us/sample - loss: 0.2797 - mse: 0.2797\n",
      "Epoch 168/500\n",
      "433/433 [==============================] - 0s 387us/sample - loss: 0.2795 - mse: 0.2795\n",
      "Epoch 169/500\n",
      "433/433 [==============================] - 0s 382us/sample - loss: 0.2828 - mse: 0.2828\n",
      "Epoch 170/500\n",
      "433/433 [==============================] - 0s 378us/sample - loss: 0.2825 - mse: 0.2825\n",
      "Epoch 171/500\n",
      "433/433 [==============================] - 0s 380us/sample - loss: 0.2802 - mse: 0.2802\n",
      "Epoch 172/500\n",
      "433/433 [==============================] - 0s 375us/sample - loss: 0.2794 - mse: 0.2794\n",
      "Epoch 173/500\n",
      "433/433 [==============================] - 0s 373us/sample - loss: 0.2780 - mse: 0.2780\n",
      "Epoch 174/500\n",
      "433/433 [==============================] - 0s 378us/sample - loss: 0.2793 - mse: 0.2793\n",
      "Epoch 175/500\n",
      "433/433 [==============================] - 0s 380us/sample - loss: 0.2772 - mse: 0.2772\n",
      "Epoch 176/500\n",
      "433/433 [==============================] - 0s 389us/sample - loss: 0.2791 - mse: 0.2791\n",
      "Epoch 177/500\n",
      "433/433 [==============================] - 0s 378us/sample - loss: 0.2761 - mse: 0.2761\n",
      "Epoch 178/500\n",
      "433/433 [==============================] - 0s 410us/sample - loss: 0.2739 - mse: 0.2739\n",
      "Epoch 179/500\n",
      "433/433 [==============================] - 0s 382us/sample - loss: 0.2724 - mse: 0.2724\n",
      "Epoch 180/500\n",
      "433/433 [==============================] - 0s 382us/sample - loss: 0.2724 - mse: 0.2724\n",
      "Epoch 181/500\n",
      "433/433 [==============================] - 0s 389us/sample - loss: 0.2712 - mse: 0.2712\n",
      "Epoch 182/500\n",
      "433/433 [==============================] - 0s 378us/sample - loss: 0.2717 - mse: 0.2717\n",
      "Epoch 183/500\n",
      "433/433 [==============================] - 0s 382us/sample - loss: 0.2711 - mse: 0.2711\n",
      "Epoch 184/500\n",
      "433/433 [==============================] - 0s 378us/sample - loss: 0.2717 - mse: 0.2717\n",
      "Epoch 185/500\n",
      "433/433 [==============================] - 0s 378us/sample - loss: 0.2702 - mse: 0.2702\n",
      "Epoch 186/500\n",
      "433/433 [==============================] - 0s 382us/sample - loss: 0.2703 - mse: 0.2703\n",
      "Epoch 187/500\n",
      "433/433 [==============================] - 0s 378us/sample - loss: 0.2717 - mse: 0.2717\n",
      "Epoch 188/500\n",
      "433/433 [==============================] - 0s 373us/sample - loss: 0.2719 - mse: 0.2719\n",
      "Epoch 189/500\n",
      "433/433 [==============================] - 0s 382us/sample - loss: 0.2690 - mse: 0.2690\n",
      "Epoch 190/500\n",
      "433/433 [==============================] - 0s 392us/sample - loss: 0.2678 - mse: 0.2678\n",
      "Epoch 191/500\n",
      "433/433 [==============================] - 0s 422us/sample - loss: 0.2663 - mse: 0.2663\n",
      "Epoch 192/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2660 - mse: 0.2660\n",
      "Epoch 193/500\n",
      "433/433 [==============================] - 0s 405us/sample - loss: 0.2642 - mse: 0.2642\n",
      "Epoch 194/500\n",
      "433/433 [==============================] - 0s 380us/sample - loss: 0.2628 - mse: 0.2628\n",
      "Epoch 195/500\n",
      "433/433 [==============================] - 0s 378us/sample - loss: 0.2616 - mse: 0.2616\n",
      "Epoch 196/500\n",
      "433/433 [==============================] - 0s 389us/sample - loss: 0.2605 - mse: 0.2605\n",
      "Epoch 197/500\n",
      "433/433 [==============================] - 0s 405us/sample - loss: 0.2602 - mse: 0.2602\n",
      "Epoch 198/500\n",
      "433/433 [==============================] - 0s 392us/sample - loss: 0.2600 - mse: 0.2600\n",
      "Epoch 199/500\n",
      "433/433 [==============================] - 0s 392us/sample - loss: 0.2614 - mse: 0.2614\n",
      "Epoch 200/500\n",
      "433/433 [==============================] - 0s 385us/sample - loss: 0.2625 - mse: 0.2625\n",
      "Epoch 201/500\n",
      "433/433 [==============================] - 0s 380us/sample - loss: 0.2610 - mse: 0.2610\n",
      "Epoch 202/500\n",
      "433/433 [==============================] - 0s 380us/sample - loss: 0.2591 - mse: 0.2591\n",
      "Epoch 203/500\n",
      "433/433 [==============================] - 0s 389us/sample - loss: 0.2585 - mse: 0.2585\n",
      "Epoch 204/500\n",
      "433/433 [==============================] - 0s 394us/sample - loss: 0.2594 - mse: 0.2594\n",
      "Epoch 205/500\n",
      "433/433 [==============================] - 0s 387us/sample - loss: 0.2601 - mse: 0.2601\n",
      "Epoch 206/500\n",
      "433/433 [==============================] - 0s 392us/sample - loss: 0.2582 - mse: 0.2582\n",
      "Epoch 207/500\n",
      "433/433 [==============================] - 0s 385us/sample - loss: 0.2572 - mse: 0.2572\n",
      "Epoch 208/500\n",
      "433/433 [==============================] - 0s 465us/sample - loss: 0.2589 - mse: 0.2589\n",
      "Epoch 209/500\n",
      "433/433 [==============================] - 0s 408us/sample - loss: 0.2577 - mse: 0.2577\n",
      "Epoch 210/500\n",
      "433/433 [==============================] - 0s 419us/sample - loss: 0.2586 - mse: 0.2586\n",
      "Epoch 211/500\n",
      "433/433 [==============================] - 0s 392us/sample - loss: 0.2611 - mse: 0.2611\n",
      "Epoch 212/500\n",
      "433/433 [==============================] - 0s 378us/sample - loss: 0.2595 - mse: 0.2595\n",
      "Epoch 213/500\n",
      "433/433 [==============================] - 0s 398us/sample - loss: 0.2582 - mse: 0.2582\n",
      "Epoch 214/500\n",
      "433/433 [==============================] - 0s 391us/sample - loss: 0.2564 - mse: 0.2564\n",
      "Epoch 215/500\n",
      "433/433 [==============================] - 0s 385us/sample - loss: 0.2566 - mse: 0.2566\n",
      "Epoch 216/500\n",
      "433/433 [==============================] - 0s 408us/sample - loss: 0.2582 - mse: 0.2582\n",
      "Epoch 217/500\n",
      "433/433 [==============================] - 0s 397us/sample - loss: 0.2582 - mse: 0.2582\n",
      "Epoch 218/500\n",
      "433/433 [==============================] - 0s 397us/sample - loss: 0.2588 - mse: 0.2588\n",
      "Epoch 219/500\n",
      "433/433 [==============================] - 0s 390us/sample - loss: 0.2588 - mse: 0.2588\n",
      "Epoch 220/500\n",
      "433/433 [==============================] - 0s 380us/sample - loss: 0.2586 - mse: 0.2586\n",
      "Epoch 221/500\n",
      "433/433 [==============================] - 0s 373us/sample - loss: 0.2548 - mse: 0.2548\n",
      "Epoch 222/500\n",
      "433/433 [==============================] - 0s 387us/sample - loss: 0.2513 - mse: 0.2513\n",
      "Epoch 223/500\n",
      "433/433 [==============================] - 0s 419us/sample - loss: 0.2527 - mse: 0.2527\n",
      "Epoch 224/500\n",
      "433/433 [==============================] - 0s 440us/sample - loss: 0.2504 - mse: 0.2504\n",
      "Epoch 225/500\n",
      "433/433 [==============================] - 0s 435us/sample - loss: 0.2510 - mse: 0.2510\n",
      "Epoch 226/500\n",
      "433/433 [==============================] - 0s 443us/sample - loss: 0.2494 - mse: 0.2494\n",
      "Epoch 227/500\n",
      "433/433 [==============================] - 0s 438us/sample - loss: 0.2486 - mse: 0.2486\n",
      "Epoch 228/500\n",
      "433/433 [==============================] - 0s 422us/sample - loss: 0.2471 - mse: 0.2471\n",
      "Epoch 229/500\n",
      "433/433 [==============================] - 0s 394us/sample - loss: 0.2466 - mse: 0.2466\n",
      "Epoch 230/500\n",
      "433/433 [==============================] - 0s 377us/sample - loss: 0.2475 - mse: 0.2475\n",
      "Epoch 231/500\n",
      "433/433 [==============================] - 0s 394us/sample - loss: 0.2478 - mse: 0.2478\n",
      "Epoch 232/500\n",
      "433/433 [==============================] - 0s 384us/sample - loss: 0.2482 - mse: 0.2482\n",
      "Epoch 233/500\n",
      "433/433 [==============================] - 0s 388us/sample - loss: 0.2457 - mse: 0.2457\n",
      "Epoch 234/500\n",
      "433/433 [==============================] - 0s 431us/sample - loss: 0.2477 - mse: 0.2477\n",
      "Epoch 235/500\n",
      "433/433 [==============================] - 0s 390us/sample - loss: 0.2488 - mse: 0.2488\n",
      "Epoch 236/500\n",
      "433/433 [==============================] - 0s 401us/sample - loss: 0.2516 - mse: 0.2516\n",
      "Epoch 237/500\n",
      "433/433 [==============================] - 0s 398us/sample - loss: 0.2505 - mse: 0.2505\n",
      "Epoch 238/500\n",
      "433/433 [==============================] - 0s 382us/sample - loss: 0.2487 - mse: 0.2487\n",
      "Epoch 239/500\n",
      "433/433 [==============================] - 0s 391us/sample - loss: 0.2467 - mse: 0.2467\n",
      "Epoch 240/500\n",
      "433/433 [==============================] - 0s 386us/sample - loss: 0.2471 - mse: 0.2471\n",
      "Epoch 241/500\n",
      "433/433 [==============================] - 0s 388us/sample - loss: 0.2472 - mse: 0.2472\n",
      "Epoch 242/500\n",
      "433/433 [==============================] - 0s 394us/sample - loss: 0.2453 - mse: 0.2453\n",
      "Epoch 243/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2458 - mse: 0.2458\n",
      "Epoch 244/500\n",
      "433/433 [==============================] - 0s 391us/sample - loss: 0.2451 - mse: 0.2451\n",
      "Epoch 245/500\n",
      "433/433 [==============================] - 0s 391us/sample - loss: 0.2447 - mse: 0.2447\n",
      "Epoch 246/500\n",
      "433/433 [==============================] - 0s 387us/sample - loss: 0.2433 - mse: 0.2433\n",
      "Epoch 247/500\n",
      "433/433 [==============================] - 0s 392us/sample - loss: 0.2435 - mse: 0.2435\n",
      "Epoch 248/500\n",
      "433/433 [==============================] - 0s 384us/sample - loss: 0.2419 - mse: 0.2419\n",
      "Epoch 249/500\n",
      "433/433 [==============================] - 0s 389us/sample - loss: 0.2420 - mse: 0.2420\n",
      "Epoch 250/500\n",
      "433/433 [==============================] - 0s 377us/sample - loss: 0.2428 - mse: 0.2428\n",
      "Epoch 251/500\n",
      "433/433 [==============================] - 0s 397us/sample - loss: 0.2441 - mse: 0.2441\n",
      "Epoch 252/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2444 - mse: 0.2444\n",
      "Epoch 253/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2478 - mse: 0.2478\n",
      "Epoch 254/500\n",
      "433/433 [==============================] - 0s 413us/sample - loss: 0.2506 - mse: 0.2506\n",
      "Epoch 255/500\n",
      "433/433 [==============================] - 0s 388us/sample - loss: 0.2470 - mse: 0.2470\n",
      "Epoch 256/500\n",
      "433/433 [==============================] - 0s 385us/sample - loss: 0.2444 - mse: 0.2444\n",
      "Epoch 257/500\n",
      "433/433 [==============================] - 0s 382us/sample - loss: 0.2424 - mse: 0.2424\n",
      "Epoch 258/500\n",
      "433/433 [==============================] - 0s 380us/sample - loss: 0.2440 - mse: 0.2440\n",
      "Epoch 259/500\n",
      "433/433 [==============================] - 0s 387us/sample - loss: 0.2471 - mse: 0.2471\n",
      "Epoch 260/500\n",
      "433/433 [==============================] - 0s 375us/sample - loss: 0.2460 - mse: 0.2460\n",
      "Epoch 261/500\n",
      "433/433 [==============================] - 0s 402us/sample - loss: 0.2444 - mse: 0.2444\n",
      "Epoch 262/500\n",
      "433/433 [==============================] - 0s 389us/sample - loss: 0.2423 - mse: 0.2423\n",
      "Epoch 263/500\n",
      "433/433 [==============================] - 0s 380us/sample - loss: 0.2390 - mse: 0.2390\n",
      "Epoch 264/500\n",
      "433/433 [==============================] - 0s 393us/sample - loss: 0.2373 - mse: 0.2373\n",
      "Epoch 265/500\n",
      "433/433 [==============================] - 0s 388us/sample - loss: 0.2368 - mse: 0.2368\n",
      "Epoch 266/500\n",
      "433/433 [==============================] - 0s 393us/sample - loss: 0.2376 - mse: 0.2376\n",
      "Epoch 267/500\n",
      "433/433 [==============================] - 0s 382us/sample - loss: 0.2372 - mse: 0.2372\n",
      "Epoch 268/500\n",
      "433/433 [==============================] - 0s 379us/sample - loss: 0.2372 - mse: 0.2372\n",
      "Epoch 269/500\n",
      "433/433 [==============================] - 0s 393us/sample - loss: 0.2352 - mse: 0.2352\n",
      "Epoch 270/500\n",
      "433/433 [==============================] - 0s 397us/sample - loss: 0.2350 - mse: 0.2350\n",
      "Epoch 271/500\n",
      "433/433 [==============================] - 0s 386us/sample - loss: 0.2357 - mse: 0.2357\n",
      "Epoch 272/500\n",
      "433/433 [==============================] - 0s 393us/sample - loss: 0.2346 - mse: 0.2346\n",
      "Epoch 273/500\n",
      "433/433 [==============================] - 0s 386us/sample - loss: 0.2354 - mse: 0.2354\n",
      "Epoch 274/500\n",
      "433/433 [==============================] - 0s 383us/sample - loss: 0.2355 - mse: 0.2355\n",
      "Epoch 275/500\n",
      "433/433 [==============================] - 0s 394us/sample - loss: 0.2347 - mse: 0.2347\n",
      "Epoch 276/500\n",
      "433/433 [==============================] - 0s 391us/sample - loss: 0.2335 - mse: 0.2335\n",
      "Epoch 277/500\n",
      "433/433 [==============================] - 0s 391us/sample - loss: 0.2329 - mse: 0.2329\n",
      "Epoch 278/500\n",
      "433/433 [==============================] - 0s 395us/sample - loss: 0.2332 - mse: 0.2332\n",
      "Epoch 279/500\n",
      "433/433 [==============================] - 0s 398us/sample - loss: 0.2337 - mse: 0.2337\n",
      "Epoch 280/500\n",
      "433/433 [==============================] - 0s 392us/sample - loss: 0.2335 - mse: 0.2335\n",
      "Epoch 281/500\n",
      "433/433 [==============================] - 0s 379us/sample - loss: 0.2346 - mse: 0.2346\n",
      "Epoch 282/500\n",
      "433/433 [==============================] - 0s 403us/sample - loss: 0.2372 - mse: 0.2372\n",
      "Epoch 283/500\n",
      "433/433 [==============================] - 0s 385us/sample - loss: 0.2373 - mse: 0.2373\n",
      "Epoch 284/500\n",
      "433/433 [==============================] - 0s 378us/sample - loss: 0.2382 - mse: 0.2382\n",
      "Epoch 285/500\n",
      "433/433 [==============================] - 0s 387us/sample - loss: 0.2400 - mse: 0.2400\n",
      "Epoch 286/500\n",
      "433/433 [==============================] - 0s 378us/sample - loss: 0.2384 - mse: 0.2384\n",
      "Epoch 287/500\n",
      "433/433 [==============================] - 0s 393us/sample - loss: 0.2383 - mse: 0.2383\n",
      "Epoch 288/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2350 - mse: 0.2350\n",
      "Epoch 289/500\n",
      "433/433 [==============================] - 0s 392us/sample - loss: 0.2332 - mse: 0.2332\n",
      "Epoch 290/500\n",
      "433/433 [==============================] - 0s 380us/sample - loss: 0.2320 - mse: 0.2320\n",
      "Epoch 291/500\n",
      "433/433 [==============================] - 0s 398us/sample - loss: 0.2317 - mse: 0.2317\n",
      "Epoch 292/500\n",
      "433/433 [==============================] - 0s 384us/sample - loss: 0.2333 - mse: 0.2333\n",
      "Epoch 293/500\n",
      "433/433 [==============================] - 0s 401us/sample - loss: 0.2327 - mse: 0.2327\n",
      "Epoch 294/500\n",
      "433/433 [==============================] - 0s 405us/sample - loss: 0.2302 - mse: 0.2302\n",
      "Epoch 295/500\n",
      "433/433 [==============================] - 0s 403us/sample - loss: 0.2290 - mse: 0.2290\n",
      "Epoch 296/500\n",
      "433/433 [==============================] - 0s 389us/sample - loss: 0.2293 - mse: 0.2293\n",
      "Epoch 297/500\n",
      "433/433 [==============================] - 0s 391us/sample - loss: 0.2282 - mse: 0.2282\n",
      "Epoch 298/500\n",
      "433/433 [==============================] - 0s 387us/sample - loss: 0.2271 - mse: 0.2271\n",
      "Epoch 299/500\n",
      "433/433 [==============================] - 0s 404us/sample - loss: 0.2264 - mse: 0.2264\n",
      "Epoch 300/500\n",
      "433/433 [==============================] - 0s 393us/sample - loss: 0.2270 - mse: 0.2270\n",
      "Epoch 301/500\n",
      "433/433 [==============================] - 0s 399us/sample - loss: 0.2286 - mse: 0.2286\n",
      "Epoch 302/500\n",
      "433/433 [==============================] - 0s 400us/sample - loss: 0.2301 - mse: 0.2301\n",
      "Epoch 303/500\n",
      "433/433 [==============================] - 0s 385us/sample - loss: 0.2304 - mse: 0.2304\n",
      "Epoch 304/500\n",
      "433/433 [==============================] - 0s 378us/sample - loss: 0.2307 - mse: 0.2307\n",
      "Epoch 305/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2307 - mse: 0.2307\n",
      "Epoch 306/500\n",
      "433/433 [==============================] - 0s 394us/sample - loss: 0.2290 - mse: 0.2290\n",
      "Epoch 307/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2310 - mse: 0.2310\n",
      "Epoch 308/500\n",
      "433/433 [==============================] - 0s 382us/sample - loss: 0.2307 - mse: 0.2307\n",
      "Epoch 309/500\n",
      "433/433 [==============================] - 0s 369us/sample - loss: 0.2284 - mse: 0.2284\n",
      "Epoch 310/500\n",
      "433/433 [==============================] - 0s 417us/sample - loss: 0.2287 - mse: 0.2287\n",
      "Epoch 311/500\n",
      "433/433 [==============================] - 0s 408us/sample - loss: 0.2283 - mse: 0.2283\n",
      "Epoch 312/500\n",
      "433/433 [==============================] - 0s 398us/sample - loss: 0.2283 - mse: 0.2283\n",
      "Epoch 313/500\n",
      "433/433 [==============================] - 0s 389us/sample - loss: 0.2279 - mse: 0.2279\n",
      "Epoch 314/500\n",
      "433/433 [==============================] - ETA: 0s - loss: 0.2288 - mse: 0.228 - 0s 403us/sample - loss: 0.2272 - mse: 0.2272\n",
      "Epoch 315/500\n",
      "433/433 [==============================] - 0s 392us/sample - loss: 0.2285 - mse: 0.2285\n",
      "Epoch 316/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2281 - mse: 0.2281\n",
      "Epoch 317/500\n",
      "433/433 [==============================] - 0s 412us/sample - loss: 0.2282 - mse: 0.2282\n",
      "Epoch 318/500\n",
      "433/433 [==============================] - 0s 394us/sample - loss: 0.2276 - mse: 0.2276\n",
      "Epoch 319/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2269 - mse: 0.2269\n",
      "Epoch 320/500\n",
      "433/433 [==============================] - 0s 385us/sample - loss: 0.2254 - mse: 0.2254\n",
      "Epoch 321/500\n",
      "433/433 [==============================] - 0s 398us/sample - loss: 0.2226 - mse: 0.2226\n",
      "Epoch 322/500\n",
      "433/433 [==============================] - 0s 394us/sample - loss: 0.2217 - mse: 0.2217\n",
      "Epoch 323/500\n",
      "433/433 [==============================] - 0s 383us/sample - loss: 0.2220 - mse: 0.2220\n",
      "Epoch 324/500\n",
      "433/433 [==============================] - 0s 380us/sample - loss: 0.2219 - mse: 0.2219\n",
      "Epoch 325/500\n",
      "433/433 [==============================] - 0s 394us/sample - loss: 0.2214 - mse: 0.2214\n",
      "Epoch 326/500\n",
      "433/433 [==============================] - 0s 387us/sample - loss: 0.2212 - mse: 0.2212\n",
      "Epoch 327/500\n",
      "433/433 [==============================] - 0s 382us/sample - loss: 0.2231 - mse: 0.2231\n",
      "Epoch 328/500\n",
      "433/433 [==============================] - 0s 389us/sample - loss: 0.2256 - mse: 0.2256\n",
      "Epoch 329/500\n",
      "433/433 [==============================] - 0s 386us/sample - loss: 0.2267 - mse: 0.2267\n",
      "Epoch 330/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2258 - mse: 0.2258\n",
      "Epoch 331/500\n",
      "433/433 [==============================] - 0s 392us/sample - loss: 0.2248 - mse: 0.2248\n",
      "Epoch 332/500\n",
      "433/433 [==============================] - 0s 392us/sample - loss: 0.2266 - mse: 0.2266\n",
      "Epoch 333/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2256 - mse: 0.2256\n",
      "Epoch 334/500\n",
      "433/433 [==============================] - 0s 397us/sample - loss: 0.2242 - mse: 0.2242\n",
      "Epoch 335/500\n",
      "433/433 [==============================] - 0s 394us/sample - loss: 0.2248 - mse: 0.2248\n",
      "Epoch 336/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2269 - mse: 0.2269\n",
      "Epoch 337/500\n",
      "433/433 [==============================] - 0s 392us/sample - loss: 0.2269 - mse: 0.2269\n",
      "Epoch 338/500\n",
      "433/433 [==============================] - 0s 377us/sample - loss: 0.2267 - mse: 0.2267\n",
      "Epoch 339/500\n",
      "433/433 [==============================] - 0s 387us/sample - loss: 0.2269 - mse: 0.2269\n",
      "Epoch 340/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2241 - mse: 0.2241\n",
      "Epoch 341/500\n",
      "433/433 [==============================] - 0s 405us/sample - loss: 0.2214 - mse: 0.2214\n",
      "Epoch 342/500\n",
      "433/433 [==============================] - 0s 388us/sample - loss: 0.2211 - mse: 0.2211\n",
      "Epoch 343/500\n",
      "433/433 [==============================] - 0s 384us/sample - loss: 0.2197 - mse: 0.2197\n",
      "Epoch 344/500\n",
      "433/433 [==============================] - 0s 378us/sample - loss: 0.2200 - mse: 0.2200\n",
      "Epoch 345/500\n",
      "433/433 [==============================] - 0s 386us/sample - loss: 0.2206 - mse: 0.2206\n",
      "Epoch 346/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2207 - mse: 0.2207\n",
      "Epoch 347/500\n",
      "433/433 [==============================] - 0s 403us/sample - loss: 0.2189 - mse: 0.2189\n",
      "Epoch 348/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2181 - mse: 0.2181\n",
      "Epoch 349/500\n",
      "433/433 [==============================] - 0s 394us/sample - loss: 0.2183 - mse: 0.2183\n",
      "Epoch 350/500\n",
      "433/433 [==============================] - 0s 387us/sample - loss: 0.2173 - mse: 0.2173\n",
      "Epoch 351/500\n",
      "433/433 [==============================] - 0s 378us/sample - loss: 0.2174 - mse: 0.2174\n",
      "Epoch 352/500\n",
      "433/433 [==============================] - 0s 408us/sample - loss: 0.2176 - mse: 0.2176\n",
      "Epoch 353/500\n",
      "433/433 [==============================] - 0s 397us/sample - loss: 0.2180 - mse: 0.2180\n",
      "Epoch 354/500\n",
      "433/433 [==============================] - 0s 383us/sample - loss: 0.2180 - mse: 0.2180s - loss: 0.2227 - mse: 0.222\n",
      "Epoch 355/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2176 - mse: 0.2176\n",
      "Epoch 356/500\n",
      "433/433 [==============================] - 0s 387us/sample - loss: 0.2192 - mse: 0.2192\n",
      "Epoch 357/500\n",
      "433/433 [==============================] - 0s 382us/sample - loss: 0.2196 - mse: 0.2196\n",
      "Epoch 358/500\n",
      "433/433 [==============================] - 0s 438us/sample - loss: 0.2212 - mse: 0.2212\n",
      "Epoch 359/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2201 - mse: 0.2201\n",
      "Epoch 360/500\n",
      "433/433 [==============================] - 0s 382us/sample - loss: 0.2192 - mse: 0.2192\n",
      "Epoch 361/500\n",
      "433/433 [==============================] - 0s 394us/sample - loss: 0.2208 - mse: 0.2208\n",
      "Epoch 362/500\n",
      "433/433 [==============================] - 0s 387us/sample - loss: 0.2215 - mse: 0.2215\n",
      "Epoch 363/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2233 - mse: 0.2233\n",
      "Epoch 364/500\n",
      "433/433 [==============================] - 0s 419us/sample - loss: 0.2230 - mse: 0.2230\n",
      "Epoch 365/500\n",
      "433/433 [==============================] - 0s 378us/sample - loss: 0.2194 - mse: 0.2194\n",
      "Epoch 366/500\n",
      "433/433 [==============================] - 0s 391us/sample - loss: 0.2176 - mse: 0.2176\n",
      "Epoch 367/500\n",
      "433/433 [==============================] - 0s 417us/sample - loss: 0.2152 - mse: 0.2152\n",
      "Epoch 368/500\n",
      "433/433 [==============================] - 0s 382us/sample - loss: 0.2148 - mse: 0.2148\n",
      "Epoch 369/500\n",
      "433/433 [==============================] - 0s 410us/sample - loss: 0.2153 - mse: 0.2153\n",
      "Epoch 370/500\n",
      "433/433 [==============================] - 0s 410us/sample - loss: 0.2162 - mse: 0.2162\n",
      "Epoch 371/500\n",
      "433/433 [==============================] - 0s 390us/sample - loss: 0.2153 - mse: 0.2153\n",
      "Epoch 372/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2146 - mse: 0.2146\n",
      "Epoch 373/500\n",
      "433/433 [==============================] - 0s 384us/sample - loss: 0.2136 - mse: 0.2136\n",
      "Epoch 374/500\n",
      "433/433 [==============================] - 0s 389us/sample - loss: 0.2127 - mse: 0.2127\n",
      "Epoch 375/500\n",
      "433/433 [==============================] - 0s 387us/sample - loss: 0.2128 - mse: 0.2128\n",
      "Epoch 376/500\n",
      "433/433 [==============================] - 0s 394us/sample - loss: 0.2124 - mse: 0.2124\n",
      "Epoch 377/500\n",
      "433/433 [==============================] - 0s 386us/sample - loss: 0.2120 - mse: 0.2120\n",
      "Epoch 378/500\n",
      "433/433 [==============================] - 0s 415us/sample - loss: 0.2117 - mse: 0.2117\n",
      "Epoch 379/500\n",
      "433/433 [==============================] - 0s 405us/sample - loss: 0.2113 - mse: 0.2113\n",
      "Epoch 380/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2117 - mse: 0.2117\n",
      "Epoch 381/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2126 - mse: 0.2126\n",
      "Epoch 382/500\n",
      "433/433 [==============================] - 0s 405us/sample - loss: 0.2129 - mse: 0.2129\n",
      "Epoch 383/500\n",
      "433/433 [==============================] - 0s 398us/sample - loss: 0.2135 - mse: 0.2135\n",
      "Epoch 384/500\n",
      "433/433 [==============================] - 0s 412us/sample - loss: 0.2122 - mse: 0.2122\n",
      "Epoch 385/500\n",
      "433/433 [==============================] - 0s 400us/sample - loss: 0.2116 - mse: 0.2116\n",
      "Epoch 386/500\n",
      "433/433 [==============================] - 0s 405us/sample - loss: 0.2110 - mse: 0.2110\n",
      "Epoch 387/500\n",
      "433/433 [==============================] - 0s 398us/sample - loss: 0.2106 - mse: 0.2106\n",
      "Epoch 388/500\n",
      "433/433 [==============================] - 0s 380us/sample - loss: 0.2103 - mse: 0.2103\n",
      "Epoch 389/500\n",
      "433/433 [==============================] - 0s 412us/sample - loss: 0.2114 - mse: 0.2114\n",
      "Epoch 390/500\n",
      "433/433 [==============================] - 0s 398us/sample - loss: 0.2127 - mse: 0.2127\n",
      "Epoch 391/500\n",
      "433/433 [==============================] - 0s 385us/sample - loss: 0.2128 - mse: 0.2128\n",
      "Epoch 392/500\n",
      "433/433 [==============================] - 0s 405us/sample - loss: 0.2132 - mse: 0.2132\n",
      "Epoch 393/500\n",
      "433/433 [==============================] - 0s 462us/sample - loss: 0.2130 - mse: 0.2130\n",
      "Epoch 394/500\n",
      "433/433 [==============================] - 0s 440us/sample - loss: 0.2123 - mse: 0.2123\n",
      "Epoch 395/500\n",
      "433/433 [==============================] - 0s 410us/sample - loss: 0.2116 - mse: 0.2116\n",
      "Epoch 396/500\n",
      "433/433 [==============================] - 0s 405us/sample - loss: 0.2120 - mse: 0.2120\n",
      "Epoch 397/500\n",
      "433/433 [==============================] - 0s 415us/sample - loss: 0.2128 - mse: 0.2128\n",
      "Epoch 398/500\n",
      "433/433 [==============================] - 0s 435us/sample - loss: 0.2134 - mse: 0.2134\n",
      "Epoch 399/500\n",
      "433/433 [==============================] - 0s 437us/sample - loss: 0.2147 - mse: 0.2147\n",
      "Epoch 400/500\n",
      "433/433 [==============================] - 0s 438us/sample - loss: 0.2140 - mse: 0.2140\n",
      "Epoch 401/500\n",
      "433/433 [==============================] - 0s 422us/sample - loss: 0.2131 - mse: 0.2131\n",
      "Epoch 402/500\n",
      "433/433 [==============================] - 0s 382us/sample - loss: 0.2125 - mse: 0.2125\n",
      "Epoch 403/500\n",
      "433/433 [==============================] - 0s 385us/sample - loss: 0.2132 - mse: 0.2132\n",
      "Epoch 404/500\n",
      "433/433 [==============================] - 0s 415us/sample - loss: 0.2133 - mse: 0.2133\n",
      "Epoch 405/500\n",
      "433/433 [==============================] - 0s 398us/sample - loss: 0.2134 - mse: 0.2134\n",
      "Epoch 406/500\n",
      "433/433 [==============================] - 0s 408us/sample - loss: 0.2128 - mse: 0.2128\n",
      "Epoch 407/500\n",
      "433/433 [==============================] - 0s 401us/sample - loss: 0.2124 - mse: 0.2124\n",
      "Epoch 408/500\n",
      "433/433 [==============================] - 0s 376us/sample - loss: 0.2118 - mse: 0.2118\n",
      "Epoch 409/500\n",
      "433/433 [==============================] - 0s 424us/sample - loss: 0.2125 - mse: 0.2125\n",
      "Epoch 410/500\n",
      "433/433 [==============================] - 0s 468us/sample - loss: 0.2134 - mse: 0.2134\n",
      "Epoch 411/500\n",
      "433/433 [==============================] - 0s 412us/sample - loss: 0.2108 - mse: 0.2108\n",
      "Epoch 412/500\n",
      "433/433 [==============================] - 0s 394us/sample - loss: 0.2098 - mse: 0.2098\n",
      "Epoch 413/500\n",
      "433/433 [==============================] - 0s 392us/sample - loss: 0.2086 - mse: 0.2086\n",
      "Epoch 414/500\n",
      "433/433 [==============================] - 0s 412us/sample - loss: 0.2099 - mse: 0.2099\n",
      "Epoch 415/500\n",
      "433/433 [==============================] - 0s 401us/sample - loss: 0.2117 - mse: 0.2117\n",
      "Epoch 416/500\n",
      "433/433 [==============================] - 0s 419us/sample - loss: 0.2112 - mse: 0.2112\n",
      "Epoch 417/500\n",
      "433/433 [==============================] - 0s 428us/sample - loss: 0.2109 - mse: 0.2109\n",
      "Epoch 418/500\n",
      "433/433 [==============================] - 0s 412us/sample - loss: 0.2096 - mse: 0.2096\n",
      "Epoch 419/500\n",
      "433/433 [==============================] - 0s 408us/sample - loss: 0.2091 - mse: 0.2091\n",
      "Epoch 420/500\n",
      "433/433 [==============================] - 0s 401us/sample - loss: 0.2096 - mse: 0.2096\n",
      "Epoch 421/500\n",
      "433/433 [==============================] - 0s 403us/sample - loss: 0.2090 - mse: 0.2090\n",
      "Epoch 422/500\n",
      "433/433 [==============================] - 0s 385us/sample - loss: 0.2074 - mse: 0.2074\n",
      "Epoch 423/500\n",
      "433/433 [==============================] - 0s 392us/sample - loss: 0.2080 - mse: 0.2080\n",
      "Epoch 424/500\n",
      "433/433 [==============================] - 0s 408us/sample - loss: 0.2072 - mse: 0.2072\n",
      "Epoch 425/500\n",
      "433/433 [==============================] - 0s 392us/sample - loss: 0.2057 - mse: 0.2057\n",
      "Epoch 426/500\n",
      "433/433 [==============================] - 0s 394us/sample - loss: 0.2076 - mse: 0.2076\n",
      "Epoch 427/500\n",
      "433/433 [==============================] - 0s 408us/sample - loss: 0.2069 - mse: 0.2069\n",
      "Epoch 428/500\n",
      "433/433 [==============================] - 0s 410us/sample - loss: 0.2075 - mse: 0.2075\n",
      "Epoch 429/500\n",
      "433/433 [==============================] - 0s 412us/sample - loss: 0.2073 - mse: 0.2073\n",
      "Epoch 430/500\n",
      "433/433 [==============================] - 0s 435us/sample - loss: 0.2078 - mse: 0.2078\n",
      "Epoch 431/500\n",
      "433/433 [==============================] - 0s 417us/sample - loss: 0.2093 - mse: 0.2093\n",
      "Epoch 432/500\n",
      "433/433 [==============================] - 0s 405us/sample - loss: 0.2103 - mse: 0.2103\n",
      "Epoch 433/500\n",
      "433/433 [==============================] - 0s 398us/sample - loss: 0.2117 - mse: 0.2117\n",
      "Epoch 434/500\n",
      "433/433 [==============================] - 0s 389us/sample - loss: 0.2120 - mse: 0.2120\n",
      "Epoch 435/500\n",
      "433/433 [==============================] - 0s 398us/sample - loss: 0.2104 - mse: 0.2104\n",
      "Epoch 436/500\n",
      "433/433 [==============================] - 0s 403us/sample - loss: 0.2087 - mse: 0.2087\n",
      "Epoch 437/500\n",
      "433/433 [==============================] - 0s 394us/sample - loss: 0.2086 - mse: 0.2086\n",
      "Epoch 438/500\n",
      "433/433 [==============================] - 0s 394us/sample - loss: 0.2067 - mse: 0.2067\n",
      "Epoch 439/500\n",
      "433/433 [==============================] - 0s 417us/sample - loss: 0.2058 - mse: 0.2058\n",
      "Epoch 440/500\n",
      "433/433 [==============================] - 0s 405us/sample - loss: 0.2074 - mse: 0.2074\n",
      "Epoch 441/500\n",
      "433/433 [==============================] - 0s 392us/sample - loss: 0.2081 - mse: 0.2081\n",
      "Epoch 442/500\n",
      "433/433 [==============================] - 0s 389us/sample - loss: 0.2065 - mse: 0.2065\n",
      "Epoch 443/500\n",
      "433/433 [==============================] - 0s 392us/sample - loss: 0.2059 - mse: 0.2059\n",
      "Epoch 444/500\n",
      "433/433 [==============================] - 0s 389us/sample - loss: 0.2051 - mse: 0.2051\n",
      "Epoch 445/500\n",
      "433/433 [==============================] - 0s 392us/sample - loss: 0.2054 - mse: 0.2054\n",
      "Epoch 446/500\n",
      "433/433 [==============================] - 0s 412us/sample - loss: 0.2037 - mse: 0.2037\n",
      "Epoch 447/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2033 - mse: 0.2033\n",
      "Epoch 448/500\n",
      "433/433 [==============================] - 0s 389us/sample - loss: 0.2022 - mse: 0.2022\n",
      "Epoch 449/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2027 - mse: 0.2027\n",
      "Epoch 450/500\n",
      "433/433 [==============================] - 0s 389us/sample - loss: 0.2024 - mse: 0.2024\n",
      "Epoch 451/500\n",
      "433/433 [==============================] - 0s 389us/sample - loss: 0.2033 - mse: 0.2033\n",
      "Epoch 452/500\n",
      "433/433 [==============================] - 0s 398us/sample - loss: 0.2051 - mse: 0.2051\n",
      "Epoch 453/500\n",
      "433/433 [==============================] - 0s 392us/sample - loss: 0.2053 - mse: 0.2053\n",
      "Epoch 454/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2048 - mse: 0.2048\n",
      "Epoch 455/500\n",
      "433/433 [==============================] - 0s 385us/sample - loss: 0.2056 - mse: 0.2056\n",
      "Epoch 456/500\n",
      "433/433 [==============================] - 0s 398us/sample - loss: 0.2050 - mse: 0.2050\n",
      "Epoch 457/500\n",
      "433/433 [==============================] - 0s 403us/sample - loss: 0.2053 - mse: 0.2053\n",
      "Epoch 458/500\n",
      "433/433 [==============================] - 0s 385us/sample - loss: 0.2039 - mse: 0.2039\n",
      "Epoch 459/500\n",
      "433/433 [==============================] - 0s 401us/sample - loss: 0.2034 - mse: 0.2034\n",
      "Epoch 460/500\n",
      "433/433 [==============================] - 0s 419us/sample - loss: 0.2050 - mse: 0.2050\n",
      "Epoch 461/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2049 - mse: 0.2049\n",
      "Epoch 462/500\n",
      "433/433 [==============================] - 0s 398us/sample - loss: 0.2044 - mse: 0.2044\n",
      "Epoch 463/500\n",
      "433/433 [==============================] - 0s 408us/sample - loss: 0.2054 - mse: 0.2054\n",
      "Epoch 464/500\n",
      "433/433 [==============================] - 0s 387us/sample - loss: 0.2063 - mse: 0.2063\n",
      "Epoch 465/500\n",
      "433/433 [==============================] - 0s 389us/sample - loss: 0.2062 - mse: 0.2062\n",
      "Epoch 466/500\n",
      "433/433 [==============================] - 0s 408us/sample - loss: 0.2051 - mse: 0.2051\n",
      "Epoch 467/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2027 - mse: 0.2027\n",
      "Epoch 468/500\n",
      "433/433 [==============================] - 0s 403us/sample - loss: 0.2030 - mse: 0.2030\n",
      "Epoch 469/500\n",
      "433/433 [==============================] - 0s 410us/sample - loss: 0.2018 - mse: 0.2018\n",
      "Epoch 470/500\n",
      "433/433 [==============================] - 0s 419us/sample - loss: 0.2011 - mse: 0.2011\n",
      "Epoch 471/500\n",
      "433/433 [==============================] - 0s 438us/sample - loss: 0.2008 - mse: 0.2008\n",
      "Epoch 472/500\n",
      "433/433 [==============================] - 0s 415us/sample - loss: 0.2013 - mse: 0.2013\n",
      "Epoch 473/500\n",
      "433/433 [==============================] - 0s 410us/sample - loss: 0.2012 - mse: 0.2012\n",
      "Epoch 474/500\n",
      "433/433 [==============================] - 0s 418us/sample - loss: 0.2008 - mse: 0.2008\n",
      "Epoch 475/500\n",
      "433/433 [==============================] - 0s 409us/sample - loss: 0.2005 - mse: 0.2005\n",
      "Epoch 476/500\n",
      "433/433 [==============================] - 0s 430us/sample - loss: 0.2017 - mse: 0.2017\n",
      "Epoch 477/500\n",
      "433/433 [==============================] - 0s 408us/sample - loss: 0.2022 - mse: 0.2022\n",
      "Epoch 478/500\n",
      "433/433 [==============================] - 0s 405us/sample - loss: 0.2028 - mse: 0.2028\n",
      "Epoch 479/500\n",
      "433/433 [==============================] - 0s 401us/sample - loss: 0.2020 - mse: 0.2020\n",
      "Epoch 480/500\n",
      "433/433 [==============================] - 0s 424us/sample - loss: 0.2007 - mse: 0.2007\n",
      "Epoch 481/500\n",
      "433/433 [==============================] - 0s 431us/sample - loss: 0.2022 - mse: 0.2022\n",
      "Epoch 482/500\n",
      "433/433 [==============================] - 0s 417us/sample - loss: 0.2033 - mse: 0.2033\n",
      "Epoch 483/500\n",
      "433/433 [==============================] - 0s 412us/sample - loss: 0.2027 - mse: 0.2027\n",
      "Epoch 484/500\n",
      "433/433 [==============================] - 0s 395us/sample - loss: 0.2023 - mse: 0.2023\n",
      "Epoch 485/500\n",
      "433/433 [==============================] - 0s 430us/sample - loss: 0.2003 - mse: 0.2003\n",
      "Epoch 486/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2006 - mse: 0.2006\n",
      "Epoch 487/500\n",
      "433/433 [==============================] - 0s 403us/sample - loss: 0.2003 - mse: 0.2003\n",
      "Epoch 488/500\n",
      "433/433 [==============================] - 0s 401us/sample - loss: 0.2009 - mse: 0.2009\n",
      "Epoch 489/500\n",
      "433/433 [==============================] - 0s 403us/sample - loss: 0.2000 - mse: 0.2000\n",
      "Epoch 490/500\n",
      "433/433 [==============================] - 0s 401us/sample - loss: 0.1997 - mse: 0.1997\n",
      "Epoch 491/500\n",
      "433/433 [==============================] - 0s 398us/sample - loss: 0.1985 - mse: 0.1985\n",
      "Epoch 492/500\n",
      "433/433 [==============================] - 0s 408us/sample - loss: 0.1980 - mse: 0.1980\n",
      "Epoch 493/500\n",
      "433/433 [==============================] - 0s 410us/sample - loss: 0.1985 - mse: 0.1985\n",
      "Epoch 494/500\n",
      "433/433 [==============================] - 0s 419us/sample - loss: 0.1986 - mse: 0.1986\n",
      "Epoch 495/500\n",
      "433/433 [==============================] - 0s 398us/sample - loss: 0.1994 - mse: 0.1994\n",
      "Epoch 496/500\n",
      "433/433 [==============================] - 0s 428us/sample - loss: 0.1989 - mse: 0.1989\n",
      "Epoch 497/500\n",
      "433/433 [==============================] - 0s 445us/sample - loss: 0.1991 - mse: 0.1991\n",
      "Epoch 498/500\n",
      "433/433 [==============================] - 0s 387us/sample - loss: 0.2015 - mse: 0.2015\n",
      "Epoch 499/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2015 - mse: 0.2015\n",
      "Epoch 500/500\n",
      "433/433 [==============================] - 0s 396us/sample - loss: 0.2005 - mse: 0.2005\n"
     ]
    }
   ],
   "source": [
    "# fit model\n",
    "epochs = 500\n",
    "history = model.fit(train_ec, train_dc, epochs=epochs, verbose=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "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": [
    "loss = history.history['loss']\n",
    "plt.figure(figsize=(8,4))\n",
    "plt.plot(range(0, epochs), loss, label='Loss')\n",
    "plt.title(\"Loss between 't+1' and reconstruction of 't'\")\n",
    "plt.xlabel('epoch')\n",
    "plt.ylabel('loss')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "72/1 [================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================] - 0s 4ms/sample - loss: 1.6372 - mse: 1.1518\n",
      "mse of test:  1.1518134\n"
     ]
    }
   ],
   "source": [
    "# test score\n",
    "loss, mse = model.evaluate(test_ec, test_dc)\n",
    "print('mse of test: ',mse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "# reconstruction\n",
    "pred = model.predict(test_ec)\n",
    "\n",
    "# future value\n",
    "future_value = pred[-1:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 289,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[[ 5.06242633e-01,  7.04212308e-01,  6.05242312e-01,\n",
       "          7.09652722e-01,  8.11226547e-01,  8.13420713e-01,\n",
       "          7.24370658e-01,  7.34983146e-01,  7.15959728e-01,\n",
       "          5.89542985e-01,  2.01928034e-01, -6.98067248e-03,\n",
       "         -6.02119714e-02, -2.62317598e-01, -4.84356463e-01,\n",
       "         -7.66965985e-01, -8.77079189e-01, -9.23557997e-01,\n",
       "         -6.00969493e-01, -7.42234066e-02,  9.80935469e-02,\n",
       "          3.27369869e-01,  3.38637620e-01,  3.94332826e-01,\n",
       "         -7.49366522e-01,  7.99524039e-02, -2.73510265e+00,\n",
       "         -2.73084664e+00, -2.75348926e+00, -2.75083852e+00,\n",
       "         -2.74679112e+00, -2.74232721e+00, -2.72891450e+00,\n",
       "         -2.70686221e+00, -2.75673938e+00, -2.81607866e+00,\n",
       "         -2.84183264e+00, -2.85083723e+00, -2.83140516e+00,\n",
       "         -2.78511214e+00, -2.75357008e+00, -2.73249245e+00,\n",
       "         -2.69749403e+00, -2.63776779e+00, -2.55096817e+00,\n",
       "         -2.46145749e+00, -2.37710047e+00, -2.28212190e+00,\n",
       "         -2.20313430e+00, -2.10415721e+00, -2.83008909e+00,\n",
       "         -2.76654696e+00, -1.51746535e+00, -1.24576843e+00,\n",
       "          1.01283228e+00, -1.21266294e+00,  6.11202598e-01,\n",
       "         -1.40558636e+00,  6.92143142e-01,  1.73018789e+00,\n",
       "         -5.14596879e-01, -2.81615585e-01, -9.52388048e-01,\n",
       "          1.22745419e+00,  1.78146029e+00, -2.12555617e-01,\n",
       "         -7.08454072e-01, -7.17143893e-01, -1.71749187e+00,\n",
       "          2.38241434e+00,  1.81315875e+00,  1.57576233e-01],\n",
       "        [-1.57722905e-01,  2.62053758e-02, -5.15267104e-02,\n",
       "         -1.57566622e-01, -3.50320160e-01, -4.53987151e-01,\n",
       "         -5.01465380e-01, -3.87502611e-01, -3.29991639e-01,\n",
       "         -7.79037625e-02, -1.44481167e-01, -1.82892099e-01,\n",
       "          1.76722854e-02, -3.30878794e-03,  2.19589055e-01,\n",
       "          4.47903603e-01,  3.36191297e-01,  1.76385850e-01,\n",
       "         -1.88576445e-01,  3.07287499e-02,  1.13231994e-01,\n",
       "          4.18144464e-02, -1.46695763e-01, -2.65021145e-01,\n",
       "         -2.56521493e-01,  7.79921263e-02, -2.07486868e+00,\n",
       "         -2.08360648e+00, -2.09820127e+00, -2.10269141e+00,\n",
       "         -2.10292315e+00, -2.09428334e+00, -2.05032873e+00,\n",
       "         -1.93942142e+00, -1.83722377e+00, -1.68932831e+00,\n",
       "         -1.57580185e+00, -1.45265698e+00, -1.31789374e+00,\n",
       "         -1.24653101e+00, -1.17756677e+00, -1.10764849e+00,\n",
       "         -1.15093720e+00, -1.23013330e+00, -1.32701635e+00,\n",
       "         -1.38100731e+00, -1.42161548e+00, -1.45302963e+00,\n",
       "         -1.47676563e+00, -1.47279978e+00, -1.39823902e+00,\n",
       "         -1.68901193e+00, -1.24318290e+00, -1.58836997e+00,\n",
       "          1.57569730e+00, -7.11963296e-01,  1.08167922e+00,\n",
       "         -6.80594265e-01, -1.20910430e+00,  1.66539061e+00,\n",
       "         -8.45262706e-01,  2.39418536e-01, -6.13131106e-01,\n",
       "         -8.71904433e-01,  1.72828889e+00, -8.36989701e-01,\n",
       "         -1.42525256e-01, -1.77686179e+00, -1.99857163e+00,\n",
       "          2.24229407e+00, -8.27216133e-02, -8.68952215e-01],\n",
       "        [ 4.79412258e-01,  7.05452383e-01,  8.28844011e-01,\n",
       "          6.07791364e-01,  6.03362978e-01,  3.49181622e-01,\n",
       "         -6.01155311e-03, -1.61062300e-01, -3.11505377e-01,\n",
       "         -2.86948442e-01, -7.47698784e-01, -1.14468932e+00,\n",
       "         -1.53816271e+00, -1.98636031e+00, -2.19632030e+00,\n",
       "         -2.29871345e+00, -2.61461186e+00, -2.86118388e+00,\n",
       "         -3.01486015e+00, -2.54942632e+00, -2.05887938e+00,\n",
       "         -1.50955856e+00, -1.54613280e+00, -1.46889722e+00,\n",
       "         -1.48776507e+00, -1.12452793e+00, -2.21737075e+00,\n",
       "         -2.27877307e+00, -2.32634926e+00, -2.41197515e+00,\n",
       "         -2.46922922e+00, -2.51676607e+00, -2.57975674e+00,\n",
       "         -2.63122630e+00, -2.67383885e+00, -2.60485506e+00,\n",
       "         -2.52689886e+00, -2.38267183e+00, -2.24130011e+00,\n",
       "         -2.08151269e+00, -1.88441360e+00, -1.77060223e+00,\n",
       "         -1.74405682e+00, -1.82722676e+00, -2.00250125e+00,\n",
       "         -2.14602280e+00, -2.24747324e+00, -2.33068871e+00,\n",
       "         -2.39107156e+00, -2.41729140e+00, -2.12268400e+00,\n",
       "         -2.37234688e+00, -2.16598535e+00, -9.46183681e-01,\n",
       "          2.66326356e+00, -2.37706208e+00,  1.37761033e+00,\n",
       "         -1.89498246e+00, -6.44134521e-01,  2.64578795e+00,\n",
       "         -2.26898003e+00, -9.43138823e-02, -1.27526498e+00,\n",
       "         -3.16304237e-01,  2.64232874e+00, -1.99918675e+00,\n",
       "         -7.69069850e-01, -1.23724043e+00, -1.57835484e+00,\n",
       "          2.94316506e+00,  1.43625307e+00, -1.31327033e+00],\n",
       "        [ 1.49664283e+00,  1.87029850e+00,  2.35795832e+00,\n",
       "          2.60669136e+00,  2.97563076e+00,  3.25727177e+00,\n",
       "          3.36143637e+00,  3.69506049e+00,  4.47997427e+00,\n",
       "          4.61381197e+00,  4.29586124e+00,  3.79032445e+00,\n",
       "          3.54170775e+00,  3.12317204e+00,  2.94415331e+00,\n",
       "          2.92567229e+00,  2.40224409e+00,  1.88857341e+00,\n",
       "          1.11251438e+00,  1.03304088e+00,  1.07901800e+00,\n",
       "          1.57152987e+00,  1.46870351e+00,  1.20740581e+00,\n",
       "          2.88413525e+00,  3.77283049e+00, -2.46555424e+00,\n",
       "         -2.50013614e+00, -2.50571775e+00, -2.53680515e+00,\n",
       "         -2.54768133e+00, -2.56355095e+00, -2.57852793e+00,\n",
       "         -2.49791884e+00, -2.44259095e+00, -2.40246940e+00,\n",
       "         -2.40586662e+00, -2.31369424e+00, -2.22966027e+00,\n",
       "         -2.19797826e+00, -2.15071321e+00, -2.14116240e+00,\n",
       "         -2.15839100e+00, -2.15549469e+00, -2.13959861e+00,\n",
       "         -2.09947729e+00, -2.01096487e+00, -1.92918324e+00,\n",
       "         -1.86658454e+00, -1.79415965e+00, -2.43363118e+00,\n",
       "         -2.37448382e+00, -3.07130456e+00, -2.78687906e+00,\n",
       "         -6.43171370e-01, -3.49229503e+00,  3.21784282e+00,\n",
       "         -2.73219061e+00, -1.30721164e+00,  6.86235905e-01,\n",
       "         -4.27938366e+00,  7.84412205e-01, -2.12724638e+00,\n",
       "         -1.18201458e+00,  1.22602034e+00, -4.36428452e+00,\n",
       "         -6.68222785e-01, -2.67776608e+00, -3.32984614e+00,\n",
       "          5.32098770e+00,  2.57627225e+00, -7.36304402e-01],\n",
       "        [ 1.62290990e+00,  1.59775329e+00,  1.68978977e+00,\n",
       "          1.46878278e+00,  1.33854544e+00,  8.07692885e-01,\n",
       "          1.82931945e-02, -8.70940506e-01, -1.97986555e+00,\n",
       "         -2.03963995e+00, -2.56415796e+00, -3.05073547e+00,\n",
       "         -3.93391681e+00, -4.40248585e+00, -4.79465199e+00,\n",
       "         -5.15430021e+00, -5.41356897e+00, -5.46153879e+00,\n",
       "         -5.19540644e+00, -4.12307930e+00, -3.41674423e+00,\n",
       "         -2.42727995e+00, -1.99789405e+00, -1.59552598e+00,\n",
       "         -2.94467354e+00, -2.76182723e+00, -2.30197334e+00,\n",
       "         -2.29929042e+00, -2.30594254e+00, -2.36118865e+00,\n",
       "         -2.41879034e+00, -2.45624042e+00, -2.58722281e+00,\n",
       "         -2.85632896e+00, -3.09519935e+00, -3.08127022e+00,\n",
       "         -2.96267748e+00, -2.83261299e+00, -2.65410662e+00,\n",
       "         -2.42133212e+00, -2.11179876e+00, -1.99046123e+00,\n",
       "         -1.90615547e+00, -2.00612807e+00, -2.19852328e+00,\n",
       "         -2.34336901e+00, -2.43889666e+00, -2.54534149e+00,\n",
       "         -2.60624719e+00, -2.61565280e+00, -2.32947922e+00,\n",
       "         -2.57596636e+00, -2.28883696e+00,  1.68405902e-02,\n",
       "          5.11897850e+00,  2.78249979e-01, -2.54895902e+00,\n",
       "         -1.66206360e+00, -3.05979609e-01,  4.59495211e+00,\n",
       "          1.29929578e+00, -2.95477414e+00, -1.36989665e+00,\n",
       "          2.52499223e-01,  4.19453526e+00,  1.54018903e+00,\n",
       "         -2.44638062e+00, -9.86349583e-01, -1.08804154e+00,\n",
       "         -1.47532940e-01, -1.58767271e+00,  1.12755813e-01]]],\n",
       "      dtype=float32)"
      ]
     },
     "execution_count": 289,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "future_value"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[-0.1577229 ,  0.02620538, -0.05152671, -0.15756662, -0.35032016,\n",
       "        -0.45398715, -0.5014654 , -0.3875026 , -0.32999164, -0.07790376,\n",
       "        -0.14448117, -0.1828921 ,  0.01767229, -0.00330879,  0.21958905,\n",
       "         0.4479036 ,  0.3361913 ,  0.17638585, -0.18857644,  0.03072875,\n",
       "         0.11323199,  0.04181445, -0.14669576, -0.26502115, -0.2565215 ,\n",
       "         0.07799213, -2.0748687 , -2.0836065 , -2.0982013 , -2.1026914 ,\n",
       "        -2.1029232 , -2.0942833 , -2.0503287 , -1.9394214 , -1.8372238 ,\n",
       "        -1.6893283 , -1.5758018 , -1.452657  , -1.3178937 , -1.246531  ,\n",
       "        -1.1775668 , -1.1076485 , -1.1509372 , -1.2301333 , -1.3270164 ,\n",
       "        -1.3810073 , -1.4216155 , -1.4530296 , -1.4767656 , -1.4727998 ,\n",
       "        -1.398239  , -1.6890119 , -1.2431829 , -1.58837   ,  1.5756973 ,\n",
       "        -0.7119633 ,  1.0816792 , -0.68059427, -1.2091043 ,  1.6653906 ,\n",
       "        -0.8452627 ,  0.23941854, -0.6131311 , -0.87190443,  1.7282889 ,\n",
       "        -0.8369897 , -0.14252526, -1.7768618 , -1.9985716 ,  2.242294  ,\n",
       "        -0.08272161, -0.8689522 ]], dtype=float32)"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "future_value[:,1,:]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. Apply future value to clf model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 287,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load clf model\n",
    "ext_grid = joblib.load('./model_t50/ExtraTreesGrid.pkl')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 297,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.50624263,  0.7042123 ,  0.6052423 ,  0.7096527 ,  0.81122655,\n",
       "        0.8134207 ,  0.72437066,  0.73498315,  0.7159597 ,  0.589543  ,\n",
       "        0.20192803, -0.00698067, -0.06021197, -0.2623176 , -0.48435646,\n",
       "       -0.766966  , -0.8770792 , -0.923558  , -0.6009695 , -0.07422341,\n",
       "        0.09809355,  0.32736987,  0.33863762,  0.39433283, -0.7493665 ,\n",
       "        0.0799524 , -2.7351027 , -2.7308466 , -2.7534893 , -2.7508385 ,\n",
       "       -2.7467911 , -2.7423272 , -2.7289145 , -2.7068622 , -2.7567394 ,\n",
       "       -2.8160787 , -2.8418326 , -2.8508372 , -2.8314052 , -2.7851121 ,\n",
       "       -2.75357   , -2.7324924 , -2.697494  , -2.6377678 , -2.5509682 ,\n",
       "       -2.4614575 , -2.3771005 , -2.282122  , -2.2031343 , -2.1041572 ,\n",
       "       -2.830089  , -2.766547  , -1.5174654 , -1.2457684 ,  1.0128323 ,\n",
       "       -1.2126629 ,  0.6112026 , -1.4055864 ,  0.69214314,  1.7301879 ,\n",
       "       -0.5145969 , -0.28161559, -0.95238805,  1.2274542 ,  1.7814603 ,\n",
       "       -0.21255562, -0.7084541 , -0.7171439 , -1.7174919 ,  2.3824143 ,\n",
       "        1.8131588 ,  0.15757623], dtype=float32)"
      ]
     },
     "execution_count": 297,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# reshape 3d to 2d\n",
    "future_value_2d = future_value.reshape(future_value.shape[1], future_value.shape[2])\n",
    "future_value_2d[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 306,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Predict\n",
    "pred_list = []\n",
    "for row in future_value_2d:\n",
    "    row = row.reshape(1,-1)\n",
    "    pred_list.append(int(ext_grid.predict(row)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 307,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[array([0.]), array([0.]), array([0.]), array([0.]), array([0.])]"
      ]
     },
     "execution_count": 307,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(pred_list)"
   ]
  }
 ],
 "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": 4
}