test/Code/test_nn.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Suggestions from Tutor\n",
    "* Perhaps with SPY or other indices\n",
    "* perform an hourly training/5 min traning, etc. \n",
    "* store pre-trained model \n",
    "* current data for prediction\n",
    "* load model\n",
    "* get ticker name input\n",
    "* predict usining model\n",
    "* show results"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Initial ML Attempt using LSTM RNN\n",
    "\n",
    "_**Binary Keras LSTM model does not output binary predictions**_\n",
    "\n",
    "\"There is no 'binary' in neural networks, but a continuous function within limits.\n",
    "Only with continuous functions a model can train and learn using \"stochastic gradient descent\".\n",
    "For trying to achieve binary results, we use the sigmoid function, which goes from 0 to 1. But initially, your model is not trained, all its \"weights\" are sort of initialised randomly. \n",
    "The result is indeed results tending to mean values, which are 0.5 in sigmoid functions.\n",
    "All you need is to train your model with enough data for enough epochs, \n",
    "so the results will gradually approach (but never hit) 0 or 1 (or whatever targets \"y\" you have in your training data)\"\n",
    "\n",
    "Source: https://stackoverflow.com/questions/46040656/binary-keras-lstm-model-does-not-output-binary-predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 647,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Source: https://github.com/Derrick-Sherrill/DerrickSherrill.com/blob/master/stocks.py\n",
    "\n",
    "import os\n",
    "import pandas as pd\n",
    "from alpha_vantage.timeseries import TimeSeries\n",
    "import time\n",
    "from path import Path\n",
    "%matplotlib inline\n",
    "from datetime import datetime as dt\n",
    "from dateutil.tz import gettz\n",
    "import re\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import hvplot.pandas\n",
    "from sklearn.preprocessing import OneHotEncoder, StandardScaler\n",
    "import tensorflow as tf\n",
    "import keras\n",
    "from newsapi import NewsApiClient\n",
    "import ipywidgets as widgets\n",
    "from IPython.display import display\n",
    "import pandas as pd\n",
    "from datetime import datetime, timedelta\n",
    "from iexfinance.stocks import get_historical_data\n",
    "from pathlib import Path\n",
    "from pathlib import Path\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense\n",
    "\n",
    "#!pip install alpha_vantage\n",
    "# get API from alpha vantage\n",
    "api_key_stocks = os.getenv(\"ALPHA_VANTAGE_API_KEY\")\n",
    "\n",
    "#import nltk\n",
    "#nltk.download()\n",
    "from nltk.corpus import reuters\n",
    "from nltk.sentiment.vader import SentimentIntensityAnalyzer\n",
    "analyzer = SentimentIntensityAnalyzer()\n",
    "\n",
    "# Read your api key environment variable\n",
    "api_key_news = os.getenv(\"NEWS_API_KEY\")\n",
    "# Create a newsapi client\n",
    "newsapi = NewsApiClient(api_key=api_key_news)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 648,
   "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>text</th>\n",
       "      <th>compound</th>\n",
       "      <th>positive</th>\n",
       "      <th>negative</th>\n",
       "      <th>neutral</th>\n",
       "      <th>close</th>\n",
       "      <th>lagged_return_1d</th>\n",
       "      <th>lagged_return_3d</th>\n",
       "      <th>lagged_return_5d</th>\n",
       "      <th>return</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",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2020-03-17</th>\n",
       "      <td>Following several other movies, including Mula...</td>\n",
       "      <td>0.9963</td>\n",
       "      <td>0.107</td>\n",
       "      <td>0.040</td>\n",
       "      <td>0.853</td>\n",
       "      <td>93.53</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.015700</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-03-18</th>\n",
       "      <td>It was supposed to launch alongside WarnerMedi...</td>\n",
       "      <td>0.9715</td>\n",
       "      <td>0.109</td>\n",
       "      <td>0.083</td>\n",
       "      <td>0.808</td>\n",
       "      <td>88.80</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.051896</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-03-19</th>\n",
       "      <td>If your streaming queue is starting to look a ...</td>\n",
       "      <td>-0.8074</td>\n",
       "      <td>0.069</td>\n",
       "      <td>0.070</td>\n",
       "      <td>0.861</td>\n",
       "      <td>94.93</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.066753</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-03-20</th>\n",
       "      <td>Disney said on Friday that it is postponing th...</td>\n",
       "      <td>0.8507</td>\n",
       "      <td>0.045</td>\n",
       "      <td>0.036</td>\n",
       "      <td>0.919</td>\n",
       "      <td>85.98</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.015700</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.099025</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-03-23</th>\n",
       "      <td>If you decide to take advantage, keep in mind ...</td>\n",
       "      <td>0.9500</td>\n",
       "      <td>0.069</td>\n",
       "      <td>0.034</td>\n",
       "      <td>0.897</td>\n",
       "      <td>85.76</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.051896</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.002562</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                         text  compound  \\\n",
       "date                                                                      \n",
       "2020-03-17  Following several other movies, including Mula...    0.9963   \n",
       "2020-03-18  It was supposed to launch alongside WarnerMedi...    0.9715   \n",
       "2020-03-19  If your streaming queue is starting to look a ...   -0.8074   \n",
       "2020-03-20  Disney said on Friday that it is postponing th...    0.8507   \n",
       "2020-03-23  If you decide to take advantage, keep in mind ...    0.9500   \n",
       "\n",
       "            positive  negative  neutral  close  lagged_return_1d  \\\n",
       "date                                                               \n",
       "2020-03-17     0.107     0.040    0.853  93.53               0.0   \n",
       "2020-03-18     0.109     0.083    0.808  88.80               0.0   \n",
       "2020-03-19     0.069     0.070    0.861  94.93               0.0   \n",
       "2020-03-20     0.045     0.036    0.919  85.98               0.0   \n",
       "2020-03-23     0.069     0.034    0.897  85.76               0.0   \n",
       "\n",
       "            lagged_return_3d  lagged_return_5d    return  \n",
       "date                                                      \n",
       "2020-03-17          0.000000               0.0 -0.015700  \n",
       "2020-03-18          0.000000               0.0 -0.051896  \n",
       "2020-03-19          0.000000               0.0  0.066753  \n",
       "2020-03-20         -0.015700               0.0 -0.099025  \n",
       "2020-03-23         -0.051896               0.0 -0.002562  "
      ]
     },
     "execution_count": 648,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Read in data\n",
    "data = Path(\"../Data/lnreturns_disney_nontrading.pickle\")\n",
    "df = pd.read_pickle(data)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 649,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.drop(columns = 'text', inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 650,
   "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>compound</th>\n",
       "      <th>positive</th>\n",
       "      <th>negative</th>\n",
       "      <th>neutral</th>\n",
       "      <th>close</th>\n",
       "      <th>lagged_return_1d</th>\n",
       "      <th>lagged_return_3d</th>\n",
       "      <th>lagged_return_5d</th>\n",
       "      <th>return</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",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2020-03-17</th>\n",
       "      <td>0.9963</td>\n",
       "      <td>0.107</td>\n",
       "      <td>0.040</td>\n",
       "      <td>0.853</td>\n",
       "      <td>93.53</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.015700</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-03-18</th>\n",
       "      <td>0.9715</td>\n",
       "      <td>0.109</td>\n",
       "      <td>0.083</td>\n",
       "      <td>0.808</td>\n",
       "      <td>88.80</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.051896</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-03-19</th>\n",
       "      <td>-0.8074</td>\n",
       "      <td>0.069</td>\n",
       "      <td>0.070</td>\n",
       "      <td>0.861</td>\n",
       "      <td>94.93</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.066753</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-03-20</th>\n",
       "      <td>0.8507</td>\n",
       "      <td>0.045</td>\n",
       "      <td>0.036</td>\n",
       "      <td>0.919</td>\n",
       "      <td>85.98</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.015700</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.099025</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-03-23</th>\n",
       "      <td>0.9500</td>\n",
       "      <td>0.069</td>\n",
       "      <td>0.034</td>\n",
       "      <td>0.897</td>\n",
       "      <td>85.76</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.051896</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.002562</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            compound  positive  negative  neutral  close  lagged_return_1d  \\\n",
       "date                                                                         \n",
       "2020-03-17    0.9963     0.107     0.040    0.853  93.53               0.0   \n",
       "2020-03-18    0.9715     0.109     0.083    0.808  88.80               0.0   \n",
       "2020-03-19   -0.8074     0.069     0.070    0.861  94.93               0.0   \n",
       "2020-03-20    0.8507     0.045     0.036    0.919  85.98               0.0   \n",
       "2020-03-23    0.9500     0.069     0.034    0.897  85.76               0.0   \n",
       "\n",
       "            lagged_return_3d  lagged_return_5d    return  \n",
       "date                                                      \n",
       "2020-03-17          0.000000               0.0 -0.015700  \n",
       "2020-03-18          0.000000               0.0 -0.051896  \n",
       "2020-03-19          0.000000               0.0  0.066753  \n",
       "2020-03-20         -0.015700               0.0 -0.099025  \n",
       "2020-03-23         -0.051896               0.0 -0.002562  "
      ]
     },
     "execution_count": 650,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 651,
   "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>compound</th>\n",
       "      <th>positive</th>\n",
       "      <th>negative</th>\n",
       "      <th>neutral</th>\n",
       "      <th>close</th>\n",
       "      <th>lagged_return_1d</th>\n",
       "      <th>lagged_return_3d</th>\n",
       "      <th>lagged_return_5d</th>\n",
       "      <th>return</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",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2020-04-07</th>\n",
       "      <td>0.9745</td>\n",
       "      <td>0.068</td>\n",
       "      <td>0.035</td>\n",
       "      <td>0.897</td>\n",
       "      <td>101.24</td>\n",
       "      <td>-0.032589</td>\n",
       "      <td>0.021367</td>\n",
       "      <td>-3.206413</td>\n",
       "      <td>0.016533</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-04-08</th>\n",
       "      <td>0.9789</td>\n",
       "      <td>0.056</td>\n",
       "      <td>0.018</td>\n",
       "      <td>0.926</td>\n",
       "      <td>101.07</td>\n",
       "      <td>-0.017544</td>\n",
       "      <td>-0.032384</td>\n",
       "      <td>-1.739130</td>\n",
       "      <td>-0.001681</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-04-09</th>\n",
       "      <td>0.9773</td>\n",
       "      <td>0.093</td>\n",
       "      <td>0.057</td>\n",
       "      <td>0.850</td>\n",
       "      <td>104.50</td>\n",
       "      <td>0.021367</td>\n",
       "      <td>0.058944</td>\n",
       "      <td>2.159713</td>\n",
       "      <td>0.033374</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-04-13</th>\n",
       "      <td>0.9912</td>\n",
       "      <td>0.103</td>\n",
       "      <td>0.047</td>\n",
       "      <td>0.849</td>\n",
       "      <td>103.50</td>\n",
       "      <td>-0.032384</td>\n",
       "      <td>0.016533</td>\n",
       "      <td>-3.186553</td>\n",
       "      <td>-0.009615</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-04-14</th>\n",
       "      <td>0.9932</td>\n",
       "      <td>0.110</td>\n",
       "      <td>0.050</td>\n",
       "      <td>0.840</td>\n",
       "      <td>106.03</td>\n",
       "      <td>0.058944</td>\n",
       "      <td>-0.001681</td>\n",
       "      <td>6.071581</td>\n",
       "      <td>0.024150</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            compound  positive  negative  neutral   close  lagged_return_1d  \\\n",
       "date                                                                          \n",
       "2020-04-07    0.9745     0.068     0.035    0.897  101.24         -0.032589   \n",
       "2020-04-08    0.9789     0.056     0.018    0.926  101.07         -0.017544   \n",
       "2020-04-09    0.9773     0.093     0.057    0.850  104.50          0.021367   \n",
       "2020-04-13    0.9912     0.103     0.047    0.849  103.50         -0.032384   \n",
       "2020-04-14    0.9932     0.110     0.050    0.840  106.03          0.058944   \n",
       "\n",
       "            lagged_return_3d  lagged_return_5d    return  \n",
       "date                                                      \n",
       "2020-04-07          0.021367         -3.206413  0.016533  \n",
       "2020-04-08         -0.032384         -1.739130 -0.001681  \n",
       "2020-04-09          0.058944          2.159713  0.033374  \n",
       "2020-04-13          0.016533         -3.186553 -0.009615  \n",
       "2020-04-14         -0.001681          6.071581  0.024150  "
      ]
     },
     "execution_count": 651,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 652,
   "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>compound</th>\n",
       "      <th>positive</th>\n",
       "      <th>negative</th>\n",
       "      <th>neutral</th>\n",
       "      <th>close</th>\n",
       "      <th>lagged_return_1d</th>\n",
       "      <th>lagged_return_3d</th>\n",
       "      <th>lagged_return_5d</th>\n",
       "      <th>return</th>\n",
       "      <th>Increase</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",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2020-03-17</th>\n",
       "      <td>0.9963</td>\n",
       "      <td>0.107</td>\n",
       "      <td>0.040</td>\n",
       "      <td>0.853</td>\n",
       "      <td>93.53</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.015700</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-03-18</th>\n",
       "      <td>0.9715</td>\n",
       "      <td>0.109</td>\n",
       "      <td>0.083</td>\n",
       "      <td>0.808</td>\n",
       "      <td>88.80</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.051896</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-03-19</th>\n",
       "      <td>-0.8074</td>\n",
       "      <td>0.069</td>\n",
       "      <td>0.070</td>\n",
       "      <td>0.861</td>\n",
       "      <td>94.93</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.066753</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-03-20</th>\n",
       "      <td>0.8507</td>\n",
       "      <td>0.045</td>\n",
       "      <td>0.036</td>\n",
       "      <td>0.919</td>\n",
       "      <td>85.98</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.015700</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.099025</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-03-23</th>\n",
       "      <td>0.9500</td>\n",
       "      <td>0.069</td>\n",
       "      <td>0.034</td>\n",
       "      <td>0.897</td>\n",
       "      <td>85.76</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.051896</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.002562</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            compound  positive  negative  neutral  close  lagged_return_1d  \\\n",
       "date                                                                         \n",
       "2020-03-17    0.9963     0.107     0.040    0.853  93.53               0.0   \n",
       "2020-03-18    0.9715     0.109     0.083    0.808  88.80               0.0   \n",
       "2020-03-19   -0.8074     0.069     0.070    0.861  94.93               0.0   \n",
       "2020-03-20    0.8507     0.045     0.036    0.919  85.98               0.0   \n",
       "2020-03-23    0.9500     0.069     0.034    0.897  85.76               0.0   \n",
       "\n",
       "            lagged_return_3d  lagged_return_5d    return  Increase  \n",
       "date                                                                \n",
       "2020-03-17          0.000000               0.0 -0.015700         0  \n",
       "2020-03-18          0.000000               0.0 -0.051896         0  \n",
       "2020-03-19          0.000000               0.0  0.066753         1  \n",
       "2020-03-20         -0.015700               0.0 -0.099025         0  \n",
       "2020-03-23         -0.051896               0.0 -0.002562         0  "
      ]
     },
     "execution_count": 652,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['Increase'] = (df['return'] >= 0).astype(int)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 653,
   "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>compound</th>\n",
       "      <th>positive</th>\n",
       "      <th>negative</th>\n",
       "      <th>neutral</th>\n",
       "      <th>close</th>\n",
       "      <th>lagged_return_1d</th>\n",
       "      <th>lagged_return_3d</th>\n",
       "      <th>lagged_return_5d</th>\n",
       "      <th>return</th>\n",
       "      <th>Increase</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",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2020-03-17</th>\n",
       "      <td>0.9963</td>\n",
       "      <td>0.107</td>\n",
       "      <td>0.040</td>\n",
       "      <td>0.853</td>\n",
       "      <td>93.53</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.015700</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-03-18</th>\n",
       "      <td>0.9715</td>\n",
       "      <td>0.109</td>\n",
       "      <td>0.083</td>\n",
       "      <td>0.808</td>\n",
       "      <td>88.80</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.051896</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-03-19</th>\n",
       "      <td>-0.8074</td>\n",
       "      <td>0.069</td>\n",
       "      <td>0.070</td>\n",
       "      <td>0.861</td>\n",
       "      <td>94.93</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.066753</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-03-20</th>\n",
       "      <td>0.8507</td>\n",
       "      <td>0.045</td>\n",
       "      <td>0.036</td>\n",
       "      <td>0.919</td>\n",
       "      <td>85.98</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.015700</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.099025</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-03-23</th>\n",
       "      <td>0.9500</td>\n",
       "      <td>0.069</td>\n",
       "      <td>0.034</td>\n",
       "      <td>0.897</td>\n",
       "      <td>85.76</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.051896</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.002562</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            compound  positive  negative  neutral  close  lagged_return_1d  \\\n",
       "date                                                                         \n",
       "2020-03-17    0.9963     0.107     0.040    0.853  93.53               0.0   \n",
       "2020-03-18    0.9715     0.109     0.083    0.808  88.80               0.0   \n",
       "2020-03-19   -0.8074     0.069     0.070    0.861  94.93               0.0   \n",
       "2020-03-20    0.8507     0.045     0.036    0.919  85.98               0.0   \n",
       "2020-03-23    0.9500     0.069     0.034    0.897  85.76               0.0   \n",
       "\n",
       "            lagged_return_3d  lagged_return_5d    return  Increase  \n",
       "date                                                                \n",
       "2020-03-17          0.000000               0.0 -0.015700         0  \n",
       "2020-03-18          0.000000               0.0 -0.051896         0  \n",
       "2020-03-19          0.000000               0.0  0.066753         1  \n",
       "2020-03-20         -0.015700               0.0 -0.099025         0  \n",
       "2020-03-23         -0.051896               0.0 -0.002562         0  "
      ]
     },
     "execution_count": 653,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.Increase = df.Increase.astype('str')\n",
    "df = pd.get_dummies(df, drop_first=True)\n",
    "df.rename(columns={'Increase_1': 'Increase'}, inplace=True)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 654,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "compound            float64\n",
       "positive            float64\n",
       "negative            float64\n",
       "neutral             float64\n",
       "close               float64\n",
       "lagged_return_1d    float64\n",
       "lagged_return_3d    float64\n",
       "lagged_return_5d    float64\n",
       "return              float64\n",
       "Increase              uint8\n",
       "dtype: object"
      ]
     },
     "execution_count": 654,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 655,
   "metadata": {},
   "outputs": [],
   "source": [
    "#df.to_pickle('../Data/dis_dum.pickle')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 656,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Read in data\n",
    "#data = Path(\"../Data/lnreturns_disney_nontrading.pickle\")\n",
    "#df = pd.read_pickle(data)\n",
    "#df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 657,
   "metadata": {},
   "outputs": [],
   "source": [
    "#df.drop(columns = 'text', inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 658,
   "metadata": {},
   "outputs": [],
   "source": [
    "#df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 659,
   "metadata": {},
   "outputs": [],
   "source": [
    "#df.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 660,
   "metadata": {},
   "outputs": [],
   "source": [
    "#df['Increase'] = (df['return'] >= 0).astype(int)\n",
    "#df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 661,
   "metadata": {},
   "outputs": [],
   "source": [
    "#df.Increase = df.Increase.astype('str')\n",
    "#df = pd.get_dummies(df, drop_first=True)\n",
    "#df.rename(columns={'Increase_1': 'Increase'}, inplace=True)\n",
    "#df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 662,
   "metadata": {},
   "outputs": [],
   "source": [
    "# df.dtypes()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 663,
   "metadata": {},
   "outputs": [],
   "source": [
    "#df.to_pickle('../Data/dis_ln_dum.pickle')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Neural Network Sequential Model\n",
    "## _Data Preparation_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 664,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create the features (X) and target (y) sets\n",
    "X = df.iloc[:, 0:4].values\n",
    "y = df[\"Increase\"].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 665,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.9963,  0.107 ,  0.04  ,  0.853 ],\n",
       "       [ 0.9715,  0.109 ,  0.083 ,  0.808 ],\n",
       "       [-0.8074,  0.069 ,  0.07  ,  0.861 ],\n",
       "       [ 0.8507,  0.045 ,  0.036 ,  0.919 ],\n",
       "       [ 0.95  ,  0.069 ,  0.034 ,  0.897 ],\n",
       "       [ 0.9942,  0.093 ,  0.036 ,  0.872 ],\n",
       "       [ 0.9952,  0.115 ,  0.053 ,  0.832 ],\n",
       "       [ 0.3439,  0.066 ,  0.047 ,  0.887 ],\n",
       "       [ 0.9948,  0.12  ,  0.047 ,  0.834 ],\n",
       "       [ 0.722 ,  0.082 ,  0.085 ,  0.834 ],\n",
       "       [ 0.9961,  0.092 ,  0.028 ,  0.88  ],\n",
       "       [ 0.9927,  0.1   ,  0.046 ,  0.855 ],\n",
       "       [ 0.9963,  0.123 ,  0.054 ,  0.823 ],\n",
       "       [ 0.9963,  0.12  ,  0.047 ,  0.833 ],\n",
       "       [ 0.9973,  0.125 ,  0.047 ,  0.828 ],\n",
       "       [ 0.9745,  0.068 ,  0.035 ,  0.897 ],\n",
       "       [ 0.9789,  0.056 ,  0.018 ,  0.926 ],\n",
       "       [ 0.9773,  0.093 ,  0.057 ,  0.85  ],\n",
       "       [ 0.9912,  0.103 ,  0.047 ,  0.849 ],\n",
       "       [ 0.9932,  0.11  ,  0.05  ,  0.84  ]])"
      ]
     },
     "execution_count": 665,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 666,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 0, 1, 0, 0, 1, 1, 1, 0, 1, 0, 0, 1, 0, 1, 1, 0, 1, 0, 1],\n",
       "      dtype=uint8)"
      ]
     },
     "execution_count": 666,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 667,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(20, 1)"
      ]
     },
     "execution_count": 667,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y= y.reshape(-1,1)\n",
    "y.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 668,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(20, 4)"
      ]
     },
     "execution_count": 668,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## _Model Fitting_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 669,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Scale the data\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "scaler = StandardScaler().fit(X)\n",
    "X = scaler.transform(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 670,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Creating training and testing data sets\n",
    "from sklearn.model_selection import train_test_split\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.3, shuffle=False) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 671,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define the model - shallow neural net\n",
    "number_hidden_nodes = 8\n",
    "number_input_features = 4\n",
    "\n",
    "nn = Sequential()\n",
    "# Hidden layer\n",
    "nn.add(\n",
    "    Dense(units=number_hidden_nodes, input_dim=number_input_features, activation=\"sigmoid\") # look like a call option\n",
    ")\n",
    "# Output layer\n",
    "nn.add(Dense(units=1, activation=\"tanh\")) # output activation should somehow match the "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 672,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_60\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_190 (Dense)            (None, 8)                 40        \n",
      "_________________________________________________________________\n",
      "dense_191 (Dense)            (None, 1)                 9         \n",
      "=================================================================\n",
      "Total params: 49\n",
      "Trainable params: 49\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# Compile the model\n",
    "nn.compile(loss=\"mean_squared_error\", optimizer=\"adam\", metrics=[\"mse\"])\n",
    "nn.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 673,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Creating validation data sets\n",
    "from sklearn.model_selection import train_test_split\n",
    "X_train, X_val, y_train, y_val = train_test_split(X_train, y_train, test_size = 0.3, shuffle=False) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 674,
   "metadata": {
    "tags": [
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend"
    ]
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 9 samples, validate on 5 samples\n",
      "Epoch 1/200\n",
      "9/9 [==============================] - 0s 30ms/sample - loss: 0.4886 - mse: 0.4886 - val_loss: 0.4468 - val_mse: 0.4468\n",
      "Epoch 2/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.4603 - mse: 0.4603 - val_loss: 0.4251 - val_mse: 0.4251\n",
      "Epoch 3/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.4348 - mse: 0.4348 - val_loss: 0.4057 - val_mse: 0.4057\n",
      "Epoch 4/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.4122 - mse: 0.4122 - val_loss: 0.3888 - val_mse: 0.3888\n",
      "Epoch 5/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.3922 - mse: 0.3922 - val_loss: 0.3741 - val_mse: 0.3741\n",
      "Epoch 6/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.3746 - mse: 0.3746 - val_loss: 0.3614 - val_mse: 0.3614\n",
      "Epoch 7/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.3593 - mse: 0.3593 - val_loss: 0.3504 - val_mse: 0.3504\n",
      "Epoch 8/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.3459 - mse: 0.3459 - val_loss: 0.3410 - val_mse: 0.3410\n",
      "Epoch 9/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.3342 - mse: 0.3342 - val_loss: 0.3329 - val_mse: 0.3329\n",
      "Epoch 10/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.3239 - mse: 0.3239 - val_loss: 0.3258 - val_mse: 0.3258\n",
      "Epoch 11/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.3149 - mse: 0.3149 - val_loss: 0.3198 - val_mse: 0.3198\n",
      "Epoch 12/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.3070 - mse: 0.3070 - val_loss: 0.3145 - val_mse: 0.3145\n",
      "Epoch 13/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.3000 - mse: 0.3000 - val_loss: 0.3099 - val_mse: 0.3099\n",
      "Epoch 14/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2939 - mse: 0.2939 - val_loss: 0.3058 - val_mse: 0.3058\n",
      "Epoch 15/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2884 - mse: 0.2884 - val_loss: 0.3023 - val_mse: 0.3023\n",
      "Epoch 16/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2835 - mse: 0.2835 - val_loss: 0.2991 - val_mse: 0.2991\n",
      "Epoch 17/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2791 - mse: 0.2791 - val_loss: 0.2964 - val_mse: 0.2964\n",
      "Epoch 18/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2751 - mse: 0.2751 - val_loss: 0.2939 - val_mse: 0.2939\n",
      "Epoch 19/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2716 - mse: 0.2716 - val_loss: 0.2916 - val_mse: 0.2916\n",
      "Epoch 20/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2683 - mse: 0.2683 - val_loss: 0.2896 - val_mse: 0.2896\n",
      "Epoch 21/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2654 - mse: 0.2654 - val_loss: 0.2878 - val_mse: 0.2878\n",
      "Epoch 22/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2626 - mse: 0.2626 - val_loss: 0.2861 - val_mse: 0.2861\n",
      "Epoch 23/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2601 - mse: 0.2601 - val_loss: 0.2845 - val_mse: 0.2845\n",
      "Epoch 24/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2578 - mse: 0.2578 - val_loss: 0.2831 - val_mse: 0.2831\n",
      "Epoch 25/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2557 - mse: 0.2557 - val_loss: 0.2818 - val_mse: 0.2818\n",
      "Epoch 26/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2537 - mse: 0.2537 - val_loss: 0.2805 - val_mse: 0.2805\n",
      "Epoch 27/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2518 - mse: 0.2518 - val_loss: 0.2794 - val_mse: 0.2794\n",
      "Epoch 28/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2501 - mse: 0.2501 - val_loss: 0.2782 - val_mse: 0.2782\n",
      "Epoch 29/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2484 - mse: 0.2484 - val_loss: 0.2772 - val_mse: 0.2772\n",
      "Epoch 30/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2468 - mse: 0.2468 - val_loss: 0.2762 - val_mse: 0.2762\n",
      "Epoch 31/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2453 - mse: 0.2453 - val_loss: 0.2752 - val_mse: 0.2752\n",
      "Epoch 32/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2439 - mse: 0.2439 - val_loss: 0.2743 - val_mse: 0.2743\n",
      "Epoch 33/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2426 - mse: 0.2426 - val_loss: 0.2734 - val_mse: 0.2734\n",
      "Epoch 34/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2413 - mse: 0.2413 - val_loss: 0.2726 - val_mse: 0.2726\n",
      "Epoch 35/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2400 - mse: 0.2400 - val_loss: 0.2718 - val_mse: 0.2718\n",
      "Epoch 36/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2388 - mse: 0.2388 - val_loss: 0.2710 - val_mse: 0.2710\n",
      "Epoch 37/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2376 - mse: 0.2376 - val_loss: 0.2702 - val_mse: 0.2702\n",
      "Epoch 38/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2365 - mse: 0.2365 - val_loss: 0.2694 - val_mse: 0.2694\n",
      "Epoch 39/200\n",
      "9/9 [==============================] - 0s 3ms/sample - loss: 0.2354 - mse: 0.2354 - val_loss: 0.2687 - val_mse: 0.2687\n",
      "Epoch 40/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2343 - mse: 0.2343 - val_loss: 0.2680 - val_mse: 0.2680\n",
      "Epoch 41/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2333 - mse: 0.2333 - val_loss: 0.2673 - val_mse: 0.2673\n",
      "Epoch 42/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2323 - mse: 0.2323 - val_loss: 0.2666 - val_mse: 0.2666\n",
      "Epoch 43/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2313 - mse: 0.2313 - val_loss: 0.2659 - val_mse: 0.2659\n",
      "Epoch 44/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2304 - mse: 0.2304 - val_loss: 0.2653 - val_mse: 0.2653\n",
      "Epoch 45/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2294 - mse: 0.2294 - val_loss: 0.2647 - val_mse: 0.2647\n",
      "Epoch 46/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2285 - mse: 0.2285 - val_loss: 0.2640 - val_mse: 0.2640\n",
      "Epoch 47/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2277 - mse: 0.2277 - val_loss: 0.2634 - val_mse: 0.2634\n",
      "Epoch 48/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2268 - mse: 0.2268 - val_loss: 0.2628 - val_mse: 0.2628\n",
      "Epoch 49/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2259 - mse: 0.2259 - val_loss: 0.2623 - val_mse: 0.2623\n",
      "Epoch 50/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2251 - mse: 0.2251 - val_loss: 0.2617 - val_mse: 0.2617\n",
      "Epoch 51/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2243 - mse: 0.2243 - val_loss: 0.2611 - val_mse: 0.2611\n",
      "Epoch 52/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2235 - mse: 0.2235 - val_loss: 0.2606 - val_mse: 0.2606\n",
      "Epoch 53/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2227 - mse: 0.2227 - val_loss: 0.2601 - val_mse: 0.2601\n",
      "Epoch 54/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2219 - mse: 0.2219 - val_loss: 0.2595 - val_mse: 0.2595\n",
      "Epoch 55/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2212 - mse: 0.2212 - val_loss: 0.2590 - val_mse: 0.2590\n",
      "Epoch 56/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2204 - mse: 0.2204 - val_loss: 0.2585 - val_mse: 0.2585\n",
      "Epoch 57/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2197 - mse: 0.2197 - val_loss: 0.2580 - val_mse: 0.2580\n",
      "Epoch 58/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2190 - mse: 0.2190 - val_loss: 0.2576 - val_mse: 0.2576\n",
      "Epoch 59/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2183 - mse: 0.2183 - val_loss: 0.2571 - val_mse: 0.2571\n",
      "Epoch 60/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2176 - mse: 0.2176 - val_loss: 0.2566 - val_mse: 0.2566\n",
      "Epoch 61/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2169 - mse: 0.2169 - val_loss: 0.2562 - val_mse: 0.2562\n",
      "Epoch 62/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2162 - mse: 0.2162 - val_loss: 0.2558 - val_mse: 0.2558\n",
      "Epoch 63/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2156 - mse: 0.2156 - val_loss: 0.2553 - val_mse: 0.2553\n",
      "Epoch 64/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2149 - mse: 0.2149 - val_loss: 0.2549 - val_mse: 0.2549\n",
      "Epoch 65/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2143 - mse: 0.2143 - val_loss: 0.2545 - val_mse: 0.2545\n",
      "Epoch 66/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2136 - mse: 0.2136 - val_loss: 0.2541 - val_mse: 0.2541\n",
      "Epoch 67/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2130 - mse: 0.2130 - val_loss: 0.2537 - val_mse: 0.2537\n",
      "Epoch 68/200\n",
      "9/9 [==============================] - 0s 3ms/sample - loss: 0.2124 - mse: 0.2124 - val_loss: 0.2534 - val_mse: 0.2534\n",
      "Epoch 69/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2118 - mse: 0.2118 - val_loss: 0.2530 - val_mse: 0.2530\n",
      "Epoch 70/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2112 - mse: 0.2112 - val_loss: 0.2527 - val_mse: 0.2527\n",
      "Epoch 71/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2106 - mse: 0.2106 - val_loss: 0.2523 - val_mse: 0.2523\n",
      "Epoch 72/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2101 - mse: 0.2101 - val_loss: 0.2520 - val_mse: 0.2520\n",
      "Epoch 73/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2095 - mse: 0.2095 - val_loss: 0.2516 - val_mse: 0.2516\n",
      "Epoch 74/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2089 - mse: 0.2089 - val_loss: 0.2513 - val_mse: 0.2513\n",
      "Epoch 75/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2084 - mse: 0.2084 - val_loss: 0.2510 - val_mse: 0.2510\n",
      "Epoch 76/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2079 - mse: 0.2079 - val_loss: 0.2507 - val_mse: 0.2507\n",
      "Epoch 77/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2073 - mse: 0.2073 - val_loss: 0.2504 - val_mse: 0.2504\n",
      "Epoch 78/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2068 - mse: 0.2068 - val_loss: 0.2501 - val_mse: 0.2501\n",
      "Epoch 79/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2063 - mse: 0.2063 - val_loss: 0.2499 - val_mse: 0.2499\n",
      "Epoch 80/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2058 - mse: 0.2058 - val_loss: 0.2496 - val_mse: 0.2496\n",
      "Epoch 81/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2053 - mse: 0.2053 - val_loss: 0.2493 - val_mse: 0.2493\n",
      "Epoch 82/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2048 - mse: 0.2048 - val_loss: 0.2491 - val_mse: 0.2491\n",
      "Epoch 83/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2043 - mse: 0.2043 - val_loss: 0.2488 - val_mse: 0.2488\n",
      "Epoch 84/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2038 - mse: 0.2038 - val_loss: 0.2486 - val_mse: 0.2486\n",
      "Epoch 85/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2033 - mse: 0.2033 - val_loss: 0.2484 - val_mse: 0.2484\n",
      "Epoch 86/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2028 - mse: 0.2028 - val_loss: 0.2482 - val_mse: 0.2482\n",
      "Epoch 87/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2023 - mse: 0.2023 - val_loss: 0.2480 - val_mse: 0.2480\n",
      "Epoch 88/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2019 - mse: 0.2019 - val_loss: 0.2478 - val_mse: 0.2478\n",
      "Epoch 89/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2014 - mse: 0.2014 - val_loss: 0.2476 - val_mse: 0.2476\n",
      "Epoch 90/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2010 - mse: 0.2010 - val_loss: 0.2474 - val_mse: 0.2474\n",
      "Epoch 91/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2005 - mse: 0.2005 - val_loss: 0.2472 - val_mse: 0.2472\n",
      "Epoch 92/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2001 - mse: 0.2001 - val_loss: 0.2470 - val_mse: 0.2470\n",
      "Epoch 93/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1996 - mse: 0.1996 - val_loss: 0.2468 - val_mse: 0.2468\n",
      "Epoch 94/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1992 - mse: 0.1992 - val_loss: 0.2467 - val_mse: 0.2467\n",
      "Epoch 95/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1988 - mse: 0.1988 - val_loss: 0.2465 - val_mse: 0.2465\n",
      "Epoch 96/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1984 - mse: 0.1984 - val_loss: 0.2464 - val_mse: 0.2464\n",
      "Epoch 97/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1979 - mse: 0.1979 - val_loss: 0.2462 - val_mse: 0.2462\n",
      "Epoch 98/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1975 - mse: 0.1975 - val_loss: 0.2461 - val_mse: 0.2461\n",
      "Epoch 99/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1971 - mse: 0.1971 - val_loss: 0.2460 - val_mse: 0.2460\n",
      "Epoch 100/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1967 - mse: 0.1967 - val_loss: 0.2459 - val_mse: 0.2459\n",
      "Epoch 101/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1963 - mse: 0.1963 - val_loss: 0.2457 - val_mse: 0.2457\n",
      "Epoch 102/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1959 - mse: 0.1959 - val_loss: 0.2456 - val_mse: 0.2456\n",
      "Epoch 103/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1955 - mse: 0.1955 - val_loss: 0.2455 - val_mse: 0.2455\n",
      "Epoch 104/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1951 - mse: 0.1951 - val_loss: 0.2454 - val_mse: 0.2454\n",
      "Epoch 105/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1948 - mse: 0.1948 - val_loss: 0.2453 - val_mse: 0.2453\n",
      "Epoch 106/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1944 - mse: 0.1944 - val_loss: 0.2453 - val_mse: 0.2453\n",
      "Epoch 107/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1940 - mse: 0.1940 - val_loss: 0.2452 - val_mse: 0.2452\n",
      "Epoch 108/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1936 - mse: 0.1936 - val_loss: 0.2451 - val_mse: 0.2451\n",
      "Epoch 109/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1933 - mse: 0.1933 - val_loss: 0.2450 - val_mse: 0.2450\n",
      "Epoch 110/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1929 - mse: 0.1929 - val_loss: 0.2450 - val_mse: 0.2450\n",
      "Epoch 111/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1925 - mse: 0.1925 - val_loss: 0.2449 - val_mse: 0.2449\n",
      "Epoch 112/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1922 - mse: 0.1922 - val_loss: 0.2449 - val_mse: 0.2449\n",
      "Epoch 113/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1918 - mse: 0.1918 - val_loss: 0.2448 - val_mse: 0.2448\n",
      "Epoch 114/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1915 - mse: 0.1915 - val_loss: 0.2448 - val_mse: 0.2448\n",
      "Epoch 115/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1911 - mse: 0.1911 - val_loss: 0.2447 - val_mse: 0.2447\n",
      "Epoch 116/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1908 - mse: 0.1908 - val_loss: 0.2447 - val_mse: 0.2447\n",
      "Epoch 117/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1904 - mse: 0.1904 - val_loss: 0.2447 - val_mse: 0.2447\n",
      "Epoch 118/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1901 - mse: 0.1901 - val_loss: 0.2447 - val_mse: 0.2447\n",
      "Epoch 119/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1898 - mse: 0.1898 - val_loss: 0.2446 - val_mse: 0.2446\n",
      "Epoch 120/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1894 - mse: 0.1894 - val_loss: 0.2446 - val_mse: 0.2446\n",
      "Epoch 121/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1891 - mse: 0.1891 - val_loss: 0.2446 - val_mse: 0.2446\n",
      "Epoch 122/200\n",
      "9/9 [==============================] - 0s 3ms/sample - loss: 0.1888 - mse: 0.1888 - val_loss: 0.2446 - val_mse: 0.2446\n",
      "Epoch 123/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1884 - mse: 0.1884 - val_loss: 0.2446 - val_mse: 0.2446\n",
      "Epoch 124/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1881 - mse: 0.1881 - val_loss: 0.2446 - val_mse: 0.2446\n",
      "Epoch 125/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1878 - mse: 0.1878 - val_loss: 0.2446 - val_mse: 0.2446\n",
      "Epoch 126/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1875 - mse: 0.1875 - val_loss: 0.2447 - val_mse: 0.2447\n",
      "Epoch 127/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1872 - mse: 0.1872 - val_loss: 0.2447 - val_mse: 0.2447\n",
      "Epoch 128/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1868 - mse: 0.1868 - val_loss: 0.2447 - val_mse: 0.2447\n",
      "Epoch 129/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1865 - mse: 0.1865 - val_loss: 0.2447 - val_mse: 0.2447\n",
      "Epoch 130/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1862 - mse: 0.1862 - val_loss: 0.2448 - val_mse: 0.2448\n",
      "Epoch 131/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1859 - mse: 0.1859 - val_loss: 0.2448 - val_mse: 0.2448\n",
      "Epoch 132/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1856 - mse: 0.1856 - val_loss: 0.2448 - val_mse: 0.2448\n",
      "Epoch 133/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1853 - mse: 0.1853 - val_loss: 0.2449 - val_mse: 0.2449\n",
      "Epoch 134/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1850 - mse: 0.1850 - val_loss: 0.2449 - val_mse: 0.2449\n",
      "Epoch 135/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1847 - mse: 0.1847 - val_loss: 0.2450 - val_mse: 0.2450\n",
      "Epoch 136/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1844 - mse: 0.1844 - val_loss: 0.2451 - val_mse: 0.2451\n",
      "Epoch 137/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1841 - mse: 0.1841 - val_loss: 0.2451 - val_mse: 0.2451\n",
      "Epoch 138/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1839 - mse: 0.1839 - val_loss: 0.2452 - val_mse: 0.2452\n",
      "Epoch 139/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1836 - mse: 0.1836 - val_loss: 0.2452 - val_mse: 0.2452\n",
      "Epoch 140/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1833 - mse: 0.1833 - val_loss: 0.2453 - val_mse: 0.2453\n",
      "Epoch 141/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1830 - mse: 0.1830 - val_loss: 0.2454 - val_mse: 0.2454\n",
      "Epoch 142/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1827 - mse: 0.1827 - val_loss: 0.2455 - val_mse: 0.2455\n",
      "Epoch 143/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1825 - mse: 0.1825 - val_loss: 0.2455 - val_mse: 0.2455\n",
      "Epoch 144/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1822 - mse: 0.1822 - val_loss: 0.2456 - val_mse: 0.2456\n",
      "Epoch 145/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1819 - mse: 0.1819 - val_loss: 0.2457 - val_mse: 0.2457\n",
      "Epoch 146/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1816 - mse: 0.1816 - val_loss: 0.2458 - val_mse: 0.2458\n",
      "Epoch 147/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1814 - mse: 0.1814 - val_loss: 0.2459 - val_mse: 0.2459\n",
      "Epoch 148/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1811 - mse: 0.1811 - val_loss: 0.2460 - val_mse: 0.2460\n",
      "Epoch 149/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1808 - mse: 0.1808 - val_loss: 0.2461 - val_mse: 0.2461\n",
      "Epoch 150/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1806 - mse: 0.1806 - val_loss: 0.2462 - val_mse: 0.2462\n",
      "Epoch 151/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1803 - mse: 0.1803 - val_loss: 0.2463 - val_mse: 0.2463\n",
      "Epoch 152/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1800 - mse: 0.1800 - val_loss: 0.2464 - val_mse: 0.2464\n",
      "Epoch 153/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1798 - mse: 0.1798 - val_loss: 0.2465 - val_mse: 0.2465\n",
      "Epoch 154/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1795 - mse: 0.1795 - val_loss: 0.2466 - val_mse: 0.2466\n",
      "Epoch 155/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1793 - mse: 0.1793 - val_loss: 0.2468 - val_mse: 0.2468\n",
      "Epoch 156/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1790 - mse: 0.1790 - val_loss: 0.2469 - val_mse: 0.2469\n",
      "Epoch 157/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1788 - mse: 0.1788 - val_loss: 0.2470 - val_mse: 0.2470\n",
      "Epoch 158/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1785 - mse: 0.1785 - val_loss: 0.2471 - val_mse: 0.2471\n",
      "Epoch 159/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1783 - mse: 0.1783 - val_loss: 0.2472 - val_mse: 0.2472\n",
      "Epoch 160/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1780 - mse: 0.1780 - val_loss: 0.2474 - val_mse: 0.2474\n",
      "Epoch 161/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1778 - mse: 0.1778 - val_loss: 0.2475 - val_mse: 0.2475\n",
      "Epoch 162/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1775 - mse: 0.1775 - val_loss: 0.2476 - val_mse: 0.2476\n",
      "Epoch 163/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1773 - mse: 0.1773 - val_loss: 0.2478 - val_mse: 0.2478\n",
      "Epoch 164/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1771 - mse: 0.1771 - val_loss: 0.2479 - val_mse: 0.2479\n",
      "Epoch 165/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1768 - mse: 0.1768 - val_loss: 0.2481 - val_mse: 0.2481\n",
      "Epoch 166/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1766 - mse: 0.1766 - val_loss: 0.2482 - val_mse: 0.2482\n",
      "Epoch 167/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1763 - mse: 0.1763 - val_loss: 0.2483 - val_mse: 0.2483\n",
      "Epoch 168/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1761 - mse: 0.1761 - val_loss: 0.2485 - val_mse: 0.2485\n",
      "Epoch 169/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1759 - mse: 0.1759 - val_loss: 0.2486 - val_mse: 0.2486\n",
      "Epoch 170/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1757 - mse: 0.1757 - val_loss: 0.2488 - val_mse: 0.2488\n",
      "Epoch 171/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1754 - mse: 0.1754 - val_loss: 0.2489 - val_mse: 0.2489\n",
      "Epoch 172/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1752 - mse: 0.1752 - val_loss: 0.2491 - val_mse: 0.2491\n",
      "Epoch 173/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1750 - mse: 0.1750 - val_loss: 0.2493 - val_mse: 0.2493\n",
      "Epoch 174/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1747 - mse: 0.1747 - val_loss: 0.2494 - val_mse: 0.2494\n",
      "Epoch 175/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1745 - mse: 0.1745 - val_loss: 0.2496 - val_mse: 0.2496\n",
      "Epoch 176/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1743 - mse: 0.1743 - val_loss: 0.2497 - val_mse: 0.2497\n",
      "Epoch 177/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1741 - mse: 0.1741 - val_loss: 0.2499 - val_mse: 0.2499\n",
      "Epoch 178/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1739 - mse: 0.1739 - val_loss: 0.2501 - val_mse: 0.2501\n",
      "Epoch 179/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1737 - mse: 0.1737 - val_loss: 0.2502 - val_mse: 0.2502\n",
      "Epoch 180/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1734 - mse: 0.1734 - val_loss: 0.2504 - val_mse: 0.2504\n",
      "Epoch 181/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1732 - mse: 0.1732 - val_loss: 0.2506 - val_mse: 0.2506\n",
      "Epoch 182/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1730 - mse: 0.1730 - val_loss: 0.2508 - val_mse: 0.2508\n",
      "Epoch 183/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1728 - mse: 0.1728 - val_loss: 0.2509 - val_mse: 0.2509\n",
      "Epoch 184/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1726 - mse: 0.1726 - val_loss: 0.2511 - val_mse: 0.2511\n",
      "Epoch 185/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1724 - mse: 0.1724 - val_loss: 0.2513 - val_mse: 0.2513\n",
      "Epoch 186/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1722 - mse: 0.1722 - val_loss: 0.2515 - val_mse: 0.2515\n",
      "Epoch 187/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1720 - mse: 0.1720 - val_loss: 0.2516 - val_mse: 0.2516\n",
      "Epoch 188/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1718 - mse: 0.1718 - val_loss: 0.2518 - val_mse: 0.2518\n",
      "Epoch 189/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1716 - mse: 0.1716 - val_loss: 0.2520 - val_mse: 0.2520\n",
      "Epoch 190/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1714 - mse: 0.1714 - val_loss: 0.2522 - val_mse: 0.2522\n",
      "Epoch 191/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1712 - mse: 0.1712 - val_loss: 0.2524 - val_mse: 0.2524\n",
      "Epoch 192/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1710 - mse: 0.1710 - val_loss: 0.2525 - val_mse: 0.2525\n",
      "Epoch 193/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1708 - mse: 0.1708 - val_loss: 0.2527 - val_mse: 0.2527\n",
      "Epoch 194/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1706 - mse: 0.1706 - val_loss: 0.2529 - val_mse: 0.2529\n",
      "Epoch 195/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1704 - mse: 0.1704 - val_loss: 0.2531 - val_mse: 0.2531\n",
      "Epoch 196/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1702 - mse: 0.1702 - val_loss: 0.2533 - val_mse: 0.2533\n",
      "Epoch 197/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1700 - mse: 0.1700 - val_loss: 0.2535 - val_mse: 0.2535\n",
      "Epoch 198/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1698 - mse: 0.1698 - val_loss: 0.2537 - val_mse: 0.2537\n",
      "Epoch 199/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1696 - mse: 0.1696 - val_loss: 0.2539 - val_mse: 0.2539\n",
      "Epoch 200/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1694 - mse: 0.1694 - val_loss: 0.2541 - val_mse: 0.2541\n"
     ]
    }
   ],
   "source": [
    "# Train the model\n",
    "model_1 = nn.fit(X_train, y_train, shuffle=False, validation_data=(X_val, y_val), batch_size=1, epochs=200)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 675,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot the train and test loss function\n",
    "plt.plot(model_1.history[\"loss\"])\n",
    "plt.title(\"loss_function_nn_disney - 1 hidden layer\")\n",
    "plt.legend([\"loss\"])\n",
    "plt.savefig('../Images/loss_function_nn_disney - 1 hidden layer.png')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 676,
   "metadata": {},
   "outputs": [],
   "source": [
    "import h5py\n",
    "\n",
    "nn.save('../Models/disney_nn_loss_1_layer.h5')  # creates a HDF5 file 'my_model.h5'\n",
    "del nn  # deletes the existing model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Add Another Layer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 677,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define the model - deep neural net\n",
    "number_input_features = 4\n",
    "hidden_nodes_layer1 = 8\n",
    "hidden_nodes_layer2 = 4\n",
    "\n",
    "nn = Sequential()\n",
    "# First hidden layer\n",
    "nn.add(\n",
    "    Dense(units=hidden_nodes_layer1, input_dim=number_input_features, activation=\"sigmoid\")\n",
    ")\n",
    "# Second hidden layer\n",
    "nn.add(Dense(units=hidden_nodes_layer2, activation=\"tanh\"))\n",
    "# Output layer\n",
    "nn.add(Dense(units=1, activation=\"linear\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 678,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_61\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_192 (Dense)            (None, 8)                 40        \n",
      "_________________________________________________________________\n",
      "dense_193 (Dense)            (None, 4)                 36        \n",
      "_________________________________________________________________\n",
      "dense_194 (Dense)            (None, 1)                 5         \n",
      "=================================================================\n",
      "Total params: 81\n",
      "Trainable params: 81\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# Compile model\n",
    "nn.compile(loss=\"mean_squared_error\", optimizer=\"adam\", metrics=[\"mse\"])\n",
    "nn.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 679,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Creating validation data sets\n",
    "#from sklearn.model_selection import train_test_split\n",
    "#X_train, X_val, y_train, y_val = train_test_split(X_train, y_train, test_size = 0.3, shuffle=False) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 680,
   "metadata": {
    "tags": [
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend"
    ]
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 9 samples, validate on 5 samples\n",
      "Epoch 1/200\n",
      "9/9 [==============================] - 0s 33ms/sample - loss: 1.4334 - mse: 1.4334 - val_loss: 1.1242 - val_mse: 1.1242\n",
      "Epoch 2/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 1.2053 - mse: 1.2053 - val_loss: 0.9333 - val_mse: 0.9333\n",
      "Epoch 3/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 1.0099 - mse: 1.0099 - val_loss: 0.7717 - val_mse: 0.7717\n",
      "Epoch 4/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.8444 - mse: 0.8444 - val_loss: 0.6383 - val_mse: 0.6383\n",
      "Epoch 5/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.7072 - mse: 0.7072 - val_loss: 0.5306 - val_mse: 0.5306\n",
      "Epoch 6/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.5954 - mse: 0.5954 - val_loss: 0.4454 - val_mse: 0.4454\n",
      "Epoch 7/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.5060 - mse: 0.5060 - val_loss: 0.3794 - val_mse: 0.3794\n",
      "Epoch 8/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.4356 - mse: 0.4356 - val_loss: 0.3292 - val_mse: 0.3292\n",
      "Epoch 9/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.3809 - mse: 0.3809 - val_loss: 0.2918 - val_mse: 0.2918\n",
      "Epoch 10/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.3389 - mse: 0.3389 - val_loss: 0.2644 - val_mse: 0.2644\n",
      "Epoch 11/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.3071 - mse: 0.3071 - val_loss: 0.2448 - val_mse: 0.2448\n",
      "Epoch 12/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2833 - mse: 0.2833 - val_loss: 0.2312 - val_mse: 0.2312\n",
      "Epoch 13/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2656 - mse: 0.2656 - val_loss: 0.2219 - val_mse: 0.2219\n",
      "Epoch 14/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2526 - mse: 0.2526 - val_loss: 0.2157 - val_mse: 0.2157\n",
      "Epoch 15/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2429 - mse: 0.2429 - val_loss: 0.2118 - val_mse: 0.2118\n",
      "Epoch 16/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2359 - mse: 0.2359 - val_loss: 0.2095 - val_mse: 0.2095\n",
      "Epoch 17/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2306 - mse: 0.2306 - val_loss: 0.2081 - val_mse: 0.2081\n",
      "Epoch 18/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2266 - mse: 0.2266 - val_loss: 0.2075 - val_mse: 0.2075\n",
      "Epoch 19/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2236 - mse: 0.2236 - val_loss: 0.2072 - val_mse: 0.2072\n",
      "Epoch 20/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2212 - mse: 0.2212 - val_loss: 0.2072 - val_mse: 0.2072\n",
      "Epoch 21/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2192 - mse: 0.2192 - val_loss: 0.2073 - val_mse: 0.2073\n",
      "Epoch 22/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2175 - mse: 0.2175 - val_loss: 0.2074 - val_mse: 0.2074\n",
      "Epoch 23/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2160 - mse: 0.2160 - val_loss: 0.2076 - val_mse: 0.2076\n",
      "Epoch 24/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2146 - mse: 0.2146 - val_loss: 0.2077 - val_mse: 0.2077\n",
      "Epoch 25/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2133 - mse: 0.2133 - val_loss: 0.2078 - val_mse: 0.2078\n",
      "Epoch 26/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2120 - mse: 0.2120 - val_loss: 0.2078 - val_mse: 0.2078\n",
      "Epoch 27/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2108 - mse: 0.2108 - val_loss: 0.2079 - val_mse: 0.2079\n",
      "Epoch 28/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2097 - mse: 0.2097 - val_loss: 0.2079 - val_mse: 0.2079\n",
      "Epoch 29/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2085 - mse: 0.2085 - val_loss: 0.2079 - val_mse: 0.2079\n",
      "Epoch 30/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2074 - mse: 0.2074 - val_loss: 0.2079 - val_mse: 0.2079\n",
      "Epoch 31/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2062 - mse: 0.2062 - val_loss: 0.2078 - val_mse: 0.2078\n",
      "Epoch 32/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2051 - mse: 0.2051 - val_loss: 0.2078 - val_mse: 0.2078\n",
      "Epoch 33/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2040 - mse: 0.2040 - val_loss: 0.2078 - val_mse: 0.2078\n",
      "Epoch 34/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2029 - mse: 0.2029 - val_loss: 0.2078 - val_mse: 0.2078\n",
      "Epoch 35/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2019 - mse: 0.2019 - val_loss: 0.2078 - val_mse: 0.2078\n",
      "Epoch 36/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.2008 - mse: 0.2008 - val_loss: 0.2078 - val_mse: 0.2078\n",
      "Epoch 37/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1998 - mse: 0.1998 - val_loss: 0.2078 - val_mse: 0.2078\n",
      "Epoch 38/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1987 - mse: 0.1987 - val_loss: 0.2079 - val_mse: 0.2079\n",
      "Epoch 39/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1977 - mse: 0.1977 - val_loss: 0.2079 - val_mse: 0.2079\n",
      "Epoch 40/200\n",
      "9/9 [==============================] - 0s 3ms/sample - loss: 0.1967 - mse: 0.1967 - val_loss: 0.2080 - val_mse: 0.2080\n",
      "Epoch 41/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1958 - mse: 0.1958 - val_loss: 0.2081 - val_mse: 0.2081\n",
      "Epoch 42/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1948 - mse: 0.1948 - val_loss: 0.2082 - val_mse: 0.2082\n",
      "Epoch 43/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1939 - mse: 0.1939 - val_loss: 0.2083 - val_mse: 0.2083\n",
      "Epoch 44/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1929 - mse: 0.1929 - val_loss: 0.2085 - val_mse: 0.2085\n",
      "Epoch 45/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1920 - mse: 0.1920 - val_loss: 0.2086 - val_mse: 0.2086\n",
      "Epoch 46/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1911 - mse: 0.1911 - val_loss: 0.2088 - val_mse: 0.2088\n",
      "Epoch 47/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1902 - mse: 0.1902 - val_loss: 0.2090 - val_mse: 0.2090\n",
      "Epoch 48/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1894 - mse: 0.1894 - val_loss: 0.2092 - val_mse: 0.2092\n",
      "Epoch 49/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1885 - mse: 0.1885 - val_loss: 0.2095 - val_mse: 0.2095\n",
      "Epoch 50/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1877 - mse: 0.1877 - val_loss: 0.2097 - val_mse: 0.2097\n",
      "Epoch 51/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1869 - mse: 0.1869 - val_loss: 0.2100 - val_mse: 0.2100\n",
      "Epoch 52/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1861 - mse: 0.1861 - val_loss: 0.2102 - val_mse: 0.2102\n",
      "Epoch 53/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1853 - mse: 0.1853 - val_loss: 0.2105 - val_mse: 0.2105\n",
      "Epoch 54/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1846 - mse: 0.1846 - val_loss: 0.2109 - val_mse: 0.2109\n",
      "Epoch 55/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1838 - mse: 0.1838 - val_loss: 0.2112 - val_mse: 0.2112\n",
      "Epoch 56/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1831 - mse: 0.1831 - val_loss: 0.2115 - val_mse: 0.2115\n",
      "Epoch 57/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1824 - mse: 0.1824 - val_loss: 0.2119 - val_mse: 0.2119\n",
      "Epoch 58/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1817 - mse: 0.1817 - val_loss: 0.2122 - val_mse: 0.2122\n",
      "Epoch 59/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1810 - mse: 0.1810 - val_loss: 0.2126 - val_mse: 0.2126\n",
      "Epoch 60/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1803 - mse: 0.1803 - val_loss: 0.2130 - val_mse: 0.2130\n",
      "Epoch 61/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1796 - mse: 0.1796 - val_loss: 0.2134 - val_mse: 0.2134\n",
      "Epoch 62/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1790 - mse: 0.1790 - val_loss: 0.2138 - val_mse: 0.2138\n",
      "Epoch 63/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1784 - mse: 0.1784 - val_loss: 0.2143 - val_mse: 0.2143\n",
      "Epoch 64/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1777 - mse: 0.1777 - val_loss: 0.2147 - val_mse: 0.2147\n",
      "Epoch 65/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1771 - mse: 0.1771 - val_loss: 0.2152 - val_mse: 0.2152\n",
      "Epoch 66/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1766 - mse: 0.1766 - val_loss: 0.2156 - val_mse: 0.2156\n",
      "Epoch 67/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1760 - mse: 0.1760 - val_loss: 0.2161 - val_mse: 0.2161\n",
      "Epoch 68/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1754 - mse: 0.1754 - val_loss: 0.2166 - val_mse: 0.2166\n",
      "Epoch 69/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1749 - mse: 0.1749 - val_loss: 0.2171 - val_mse: 0.2171\n",
      "Epoch 70/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1743 - mse: 0.1743 - val_loss: 0.2176 - val_mse: 0.2176\n",
      "Epoch 71/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1738 - mse: 0.1738 - val_loss: 0.2181 - val_mse: 0.2181\n",
      "Epoch 72/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1733 - mse: 0.1733 - val_loss: 0.2186 - val_mse: 0.2186\n",
      "Epoch 73/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1728 - mse: 0.1728 - val_loss: 0.2192 - val_mse: 0.2192\n",
      "Epoch 74/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1723 - mse: 0.1723 - val_loss: 0.2197 - val_mse: 0.2197\n",
      "Epoch 75/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1718 - mse: 0.1718 - val_loss: 0.2203 - val_mse: 0.2203\n",
      "Epoch 76/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1713 - mse: 0.1713 - val_loss: 0.2208 - val_mse: 0.2208\n",
      "Epoch 77/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1709 - mse: 0.1709 - val_loss: 0.2214 - val_mse: 0.2214\n",
      "Epoch 78/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1704 - mse: 0.1704 - val_loss: 0.2220 - val_mse: 0.2220\n",
      "Epoch 79/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1700 - mse: 0.1700 - val_loss: 0.2226 - val_mse: 0.2226\n",
      "Epoch 80/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1696 - mse: 0.1696 - val_loss: 0.2231 - val_mse: 0.2231\n",
      "Epoch 81/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1692 - mse: 0.1692 - val_loss: 0.2237 - val_mse: 0.2237\n",
      "Epoch 82/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1687 - mse: 0.1687 - val_loss: 0.2243 - val_mse: 0.2243\n",
      "Epoch 83/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1683 - mse: 0.1683 - val_loss: 0.2249 - val_mse: 0.2249\n",
      "Epoch 84/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1680 - mse: 0.1680 - val_loss: 0.2256 - val_mse: 0.2256\n",
      "Epoch 85/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1676 - mse: 0.1676 - val_loss: 0.2262 - val_mse: 0.2262\n",
      "Epoch 86/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1672 - mse: 0.1672 - val_loss: 0.2268 - val_mse: 0.2268\n",
      "Epoch 87/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1668 - mse: 0.1668 - val_loss: 0.2274 - val_mse: 0.2274\n",
      "Epoch 88/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1665 - mse: 0.1665 - val_loss: 0.2281 - val_mse: 0.2281\n",
      "Epoch 89/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1662 - mse: 0.1662 - val_loss: 0.2287 - val_mse: 0.2287\n",
      "Epoch 90/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1658 - mse: 0.1658 - val_loss: 0.2293 - val_mse: 0.2293\n",
      "Epoch 91/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1655 - mse: 0.1655 - val_loss: 0.2300 - val_mse: 0.2300\n",
      "Epoch 92/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1652 - mse: 0.1652 - val_loss: 0.2306 - val_mse: 0.2306\n",
      "Epoch 93/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1649 - mse: 0.1649 - val_loss: 0.2313 - val_mse: 0.2313\n",
      "Epoch 94/200\n",
      "9/9 [==============================] - 0s 3ms/sample - loss: 0.1646 - mse: 0.1646 - val_loss: 0.2319 - val_mse: 0.2319\n",
      "Epoch 95/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1643 - mse: 0.1643 - val_loss: 0.2326 - val_mse: 0.2326\n",
      "Epoch 96/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1640 - mse: 0.1640 - val_loss: 0.2332 - val_mse: 0.2332\n",
      "Epoch 97/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1637 - mse: 0.1637 - val_loss: 0.2339 - val_mse: 0.2339\n",
      "Epoch 98/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1634 - mse: 0.1634 - val_loss: 0.2346 - val_mse: 0.2346\n",
      "Epoch 99/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1631 - mse: 0.1631 - val_loss: 0.2352 - val_mse: 0.2352\n",
      "Epoch 100/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1629 - mse: 0.1629 - val_loss: 0.2359 - val_mse: 0.2359\n",
      "Epoch 101/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1626 - mse: 0.1626 - val_loss: 0.2366 - val_mse: 0.2366\n",
      "Epoch 102/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1624 - mse: 0.1624 - val_loss: 0.2372 - val_mse: 0.2372\n",
      "Epoch 103/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1621 - mse: 0.1621 - val_loss: 0.2379 - val_mse: 0.2379\n",
      "Epoch 104/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1619 - mse: 0.1619 - val_loss: 0.2386 - val_mse: 0.2386\n",
      "Epoch 105/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1616 - mse: 0.1616 - val_loss: 0.2393 - val_mse: 0.2393\n",
      "Epoch 106/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1614 - mse: 0.1614 - val_loss: 0.2399 - val_mse: 0.2399\n",
      "Epoch 107/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1612 - mse: 0.1612 - val_loss: 0.2406 - val_mse: 0.2406\n",
      "Epoch 108/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1610 - mse: 0.1610 - val_loss: 0.2413 - val_mse: 0.2413\n",
      "Epoch 109/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1608 - mse: 0.1608 - val_loss: 0.2420 - val_mse: 0.2420\n",
      "Epoch 110/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1606 - mse: 0.1606 - val_loss: 0.2426 - val_mse: 0.2426\n",
      "Epoch 111/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1603 - mse: 0.1603 - val_loss: 0.2433 - val_mse: 0.2433\n",
      "Epoch 112/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1602 - mse: 0.1602 - val_loss: 0.2440 - val_mse: 0.2440\n",
      "Epoch 113/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1600 - mse: 0.1600 - val_loss: 0.2447 - val_mse: 0.2447\n",
      "Epoch 114/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1598 - mse: 0.1598 - val_loss: 0.2453 - val_mse: 0.2453\n",
      "Epoch 115/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1596 - mse: 0.1596 - val_loss: 0.2460 - val_mse: 0.2460\n",
      "Epoch 116/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1594 - mse: 0.1594 - val_loss: 0.2467 - val_mse: 0.2467\n",
      "Epoch 117/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1592 - mse: 0.1592 - val_loss: 0.2473 - val_mse: 0.2473\n",
      "Epoch 118/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1591 - mse: 0.1591 - val_loss: 0.2480 - val_mse: 0.2480\n",
      "Epoch 119/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1589 - mse: 0.1589 - val_loss: 0.2487 - val_mse: 0.2487\n",
      "Epoch 120/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1587 - mse: 0.1587 - val_loss: 0.2493 - val_mse: 0.2493\n",
      "Epoch 121/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1586 - mse: 0.1586 - val_loss: 0.2500 - val_mse: 0.2500\n",
      "Epoch 122/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1584 - mse: 0.1584 - val_loss: 0.2507 - val_mse: 0.2507\n",
      "Epoch 123/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1582 - mse: 0.1582 - val_loss: 0.2513 - val_mse: 0.2513\n",
      "Epoch 124/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1581 - mse: 0.1581 - val_loss: 0.2520 - val_mse: 0.2520\n",
      "Epoch 125/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1580 - mse: 0.1580 - val_loss: 0.2526 - val_mse: 0.2526\n",
      "Epoch 126/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1578 - mse: 0.1578 - val_loss: 0.2533 - val_mse: 0.2533\n",
      "Epoch 127/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1577 - mse: 0.1577 - val_loss: 0.2539 - val_mse: 0.2539\n",
      "Epoch 128/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1575 - mse: 0.1575 - val_loss: 0.2546 - val_mse: 0.2546\n",
      "Epoch 129/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1574 - mse: 0.1574 - val_loss: 0.2552 - val_mse: 0.2552\n",
      "Epoch 130/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1573 - mse: 0.1573 - val_loss: 0.2559 - val_mse: 0.2559\n",
      "Epoch 131/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1571 - mse: 0.1571 - val_loss: 0.2565 - val_mse: 0.2565\n",
      "Epoch 132/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1570 - mse: 0.1570 - val_loss: 0.2571 - val_mse: 0.2571\n",
      "Epoch 133/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1569 - mse: 0.1569 - val_loss: 0.2578 - val_mse: 0.2578\n",
      "Epoch 134/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1568 - mse: 0.1568 - val_loss: 0.2584 - val_mse: 0.2584\n",
      "Epoch 135/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1566 - mse: 0.1566 - val_loss: 0.2590 - val_mse: 0.2590\n",
      "Epoch 136/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1565 - mse: 0.1565 - val_loss: 0.2596 - val_mse: 0.2596\n",
      "Epoch 137/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1564 - mse: 0.1564 - val_loss: 0.2603 - val_mse: 0.2603\n",
      "Epoch 138/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1563 - mse: 0.1563 - val_loss: 0.2609 - val_mse: 0.2609\n",
      "Epoch 139/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1562 - mse: 0.1562 - val_loss: 0.2615 - val_mse: 0.2615\n",
      "Epoch 140/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1561 - mse: 0.1561 - val_loss: 0.2621 - val_mse: 0.2621\n",
      "Epoch 141/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1560 - mse: 0.1560 - val_loss: 0.2627 - val_mse: 0.2627\n",
      "Epoch 142/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1559 - mse: 0.1559 - val_loss: 0.2633 - val_mse: 0.2633\n",
      "Epoch 143/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1558 - mse: 0.1558 - val_loss: 0.2639 - val_mse: 0.2639\n",
      "Epoch 144/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1557 - mse: 0.1557 - val_loss: 0.2645 - val_mse: 0.2645\n",
      "Epoch 145/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1556 - mse: 0.1556 - val_loss: 0.2650 - val_mse: 0.2650\n",
      "Epoch 146/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1555 - mse: 0.1555 - val_loss: 0.2656 - val_mse: 0.2656\n",
      "Epoch 147/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1554 - mse: 0.1554 - val_loss: 0.2662 - val_mse: 0.2662\n",
      "Epoch 148/200\n",
      "9/9 [==============================] - 0s 3ms/sample - loss: 0.1553 - mse: 0.1553 - val_loss: 0.2668 - val_mse: 0.2668\n",
      "Epoch 149/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1552 - mse: 0.1552 - val_loss: 0.2673 - val_mse: 0.2673\n",
      "Epoch 150/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1551 - mse: 0.1551 - val_loss: 0.2679 - val_mse: 0.2679\n",
      "Epoch 151/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1550 - mse: 0.1550 - val_loss: 0.2684 - val_mse: 0.2684\n",
      "Epoch 152/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1550 - mse: 0.1550 - val_loss: 0.2690 - val_mse: 0.2690\n",
      "Epoch 153/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1549 - mse: 0.1549 - val_loss: 0.2695 - val_mse: 0.2695\n",
      "Epoch 154/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1548 - mse: 0.1548 - val_loss: 0.2701 - val_mse: 0.2701\n",
      "Epoch 155/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1547 - mse: 0.1547 - val_loss: 0.2706 - val_mse: 0.2706\n",
      "Epoch 156/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1546 - mse: 0.1546 - val_loss: 0.2712 - val_mse: 0.2712\n",
      "Epoch 157/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1546 - mse: 0.1546 - val_loss: 0.2717 - val_mse: 0.2717\n",
      "Epoch 158/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1545 - mse: 0.1545 - val_loss: 0.2722 - val_mse: 0.2722\n",
      "Epoch 159/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1544 - mse: 0.1544 - val_loss: 0.2727 - val_mse: 0.2727\n",
      "Epoch 160/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1543 - mse: 0.1543 - val_loss: 0.2732 - val_mse: 0.2732\n",
      "Epoch 161/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1543 - mse: 0.1543 - val_loss: 0.2737 - val_mse: 0.2737\n",
      "Epoch 162/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1542 - mse: 0.1542 - val_loss: 0.2742 - val_mse: 0.2742\n",
      "Epoch 163/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1541 - mse: 0.1541 - val_loss: 0.2747 - val_mse: 0.2747\n",
      "Epoch 164/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1541 - mse: 0.1541 - val_loss: 0.2752 - val_mse: 0.2752\n",
      "Epoch 165/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1540 - mse: 0.1540 - val_loss: 0.2757 - val_mse: 0.2757\n",
      "Epoch 166/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1539 - mse: 0.1539 - val_loss: 0.2762 - val_mse: 0.2762\n",
      "Epoch 167/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1539 - mse: 0.1539 - val_loss: 0.2767 - val_mse: 0.2767\n",
      "Epoch 168/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1538 - mse: 0.1538 - val_loss: 0.2771 - val_mse: 0.2771\n",
      "Epoch 169/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1537 - mse: 0.1537 - val_loss: 0.2776 - val_mse: 0.2776\n",
      "Epoch 170/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1537 - mse: 0.1537 - val_loss: 0.2781 - val_mse: 0.2781\n",
      "Epoch 171/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1536 - mse: 0.1536 - val_loss: 0.2785 - val_mse: 0.2785\n",
      "Epoch 172/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1536 - mse: 0.1536 - val_loss: 0.2790 - val_mse: 0.2790\n",
      "Epoch 173/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1535 - mse: 0.1535 - val_loss: 0.2794 - val_mse: 0.2794\n",
      "Epoch 174/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1534 - mse: 0.1534 - val_loss: 0.2798 - val_mse: 0.2798\n",
      "Epoch 175/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1534 - mse: 0.1534 - val_loss: 0.2803 - val_mse: 0.2803\n",
      "Epoch 176/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1533 - mse: 0.1533 - val_loss: 0.2807 - val_mse: 0.2807\n",
      "Epoch 177/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1533 - mse: 0.1533 - val_loss: 0.2811 - val_mse: 0.2811\n",
      "Epoch 178/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1532 - mse: 0.1532 - val_loss: 0.2815 - val_mse: 0.2815\n",
      "Epoch 179/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1532 - mse: 0.1532 - val_loss: 0.2820 - val_mse: 0.2820\n",
      "Epoch 180/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1531 - mse: 0.1531 - val_loss: 0.2824 - val_mse: 0.2824\n",
      "Epoch 181/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1530 - mse: 0.1530 - val_loss: 0.2828 - val_mse: 0.2828\n",
      "Epoch 182/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1530 - mse: 0.1530 - val_loss: 0.2832 - val_mse: 0.2832\n",
      "Epoch 183/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1529 - mse: 0.1529 - val_loss: 0.2835 - val_mse: 0.2835\n",
      "Epoch 184/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1529 - mse: 0.1529 - val_loss: 0.2839 - val_mse: 0.2839\n",
      "Epoch 185/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1528 - mse: 0.1528 - val_loss: 0.2843 - val_mse: 0.2843\n",
      "Epoch 186/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1528 - mse: 0.1528 - val_loss: 0.2847 - val_mse: 0.2847\n",
      "Epoch 187/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1527 - mse: 0.1527 - val_loss: 0.2851 - val_mse: 0.2851\n",
      "Epoch 188/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1527 - mse: 0.1527 - val_loss: 0.2854 - val_mse: 0.2854\n",
      "Epoch 189/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1526 - mse: 0.1526 - val_loss: 0.2858 - val_mse: 0.2858\n",
      "Epoch 190/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1526 - mse: 0.1526 - val_loss: 0.2861 - val_mse: 0.2861\n",
      "Epoch 191/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1525 - mse: 0.1525 - val_loss: 0.2865 - val_mse: 0.2865\n",
      "Epoch 192/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1525 - mse: 0.1525 - val_loss: 0.2868 - val_mse: 0.2868\n",
      "Epoch 193/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1525 - mse: 0.1525 - val_loss: 0.2872 - val_mse: 0.2872\n",
      "Epoch 194/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1524 - mse: 0.1524 - val_loss: 0.2875 - val_mse: 0.2875\n",
      "Epoch 195/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1524 - mse: 0.1524 - val_loss: 0.2878 - val_mse: 0.2878\n",
      "Epoch 196/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1523 - mse: 0.1523 - val_loss: 0.2882 - val_mse: 0.2882\n",
      "Epoch 197/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1523 - mse: 0.1523 - val_loss: 0.2885 - val_mse: 0.2885\n",
      "Epoch 198/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1522 - mse: 0.1522 - val_loss: 0.2888 - val_mse: 0.2888\n",
      "Epoch 199/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1522 - mse: 0.1522 - val_loss: 0.2891 - val_mse: 0.2891\n",
      "Epoch 200/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1521 - mse: 0.1521 - val_loss: 0.2894 - val_mse: 0.2894\n"
     ]
    }
   ],
   "source": [
    "# Fit the model\n",
    "model_2 = nn.fit(X_train, y_train, shuffle=False, validation_data=(X_val, y_val), batch_size=1, epochs=200)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 681,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAEICAYAAABPgw/pAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nO3deZwcdZ3/8denj7mP3CQhhCRcJkDkSMIpBDkMiMAu/oD8cAUB8VhWF3VXdt0FRPZyfwoirghyuIoExHVlMXgAsiiEIwiEIwECBDLkmiQkmWRmMt09n98fVdPpmXTP9Exmpqe738/Hox9VXfXtqk+qJ++q/lZ1tbk7IiJS/CKFLkBERAaHAl1EpEQo0EVESoQCXUSkRCjQRURKhAJdRKREKNALyMxWmdkpw7xOM7M7zex9M3tmmNf9kJldNJzrHApm5ma2fzh+i5n9Y6FrKhQz+62ZXTjYbWVgTNehF46ZrQIuc/eHh3GdHwLuAQ5y9x1DuJ5rgf3d/RNDtY5CMTMHDnD3lYWupT/M7CHgQ+HTSsCBjvD5T9z9swUpTAZNrNAFyLDbF1g1lGEuI5O7n941bmZ3AU3u/g+52ptZzN2Tw1GbDA51uYwQZlZpZjea2ZrwcaOZVYbzxpnZg2a2xcw2m9kfzCwSzvuqmb1nZi1m9pqZndzLOi4FfggcY2bbzezrZnaxmf2xR7vMLoW7zOx7ZvarcB1Pm9l+GW0PNrPfhXWtN7O/N7MFwN8D54freTFs+5iZXRaOR8zsH8zsHTPbYGb/aWaN4bxpYQ0Xmdm7ZrbRzL6Wxza81szuC5fVYmavmNmcjPmrzOwrZrbMzLaa2b1mVpXHcv/GzNaG78slPebdZWbXh+O9vU+9rtvMzjSzF8LXPmlmszPW/fMe6/yumd3YV939ZWanhHX+vZmtA24zs7FmttjMmsNuuv8xs70zXvNHM7s4HL/MzP7XzG4I/x1vmdlpA2y7X9i+xYKumu+HOyHphQJ95PgacDRwGPBBYB7QdfT0ZaAJGA/sRRCWbmYHAVcAc929HvgIsCrXCtz9duCzwBJ3r3P3a/KsbSHwdWA0sBL4JwAzqwceBn4NTAb2Bx5x918D/wzcG67ng1mWeXH4OAmYAdQBN/doczxwEHAycLWZzcyj1rOARcAo4IEsyzwPWABMB2aHNeQU7py+ApwKHAD0ds4j6/vU17rN7AjgDuAzwFjgB8AD4Q79J8ACMxsVto0B5wM/7q3uPTCF4L2YCnyeICNuC5/vCySA7/Ty+mOBlwj+HTcAtw+w7T3AE+G864GS67obCgr0keNC4Dp33+DuzQQB+hfhvAQwCdjX3RPu/gcPTn6kCPpCZ5lZ3N1XufubQ1Dbf7n7M+HH77sJdjoAZwLr3P1b7t7u7i3u/nSey7wQ+La7v+Xu24G/Ay4IA6vL1929zd1fBF4k2NH15Y/uvtjdUwSh1/M1N7n7GnffDPxPxr8ll/OAO9395bCb6tpe2uZ6n/pa96eBH7j70+6ecvcfATuBo919LfA48H/CtguAje7+XB91D1QSuNbdO8Jt3+zuvwjHtxHsqE/s5fVvuvsd4fb/ETDFzMb1p62ZzSB437rqeBz41eD9E0uXAn3kmAy8k/H8nXAawL8THBn/NvxoehVAeFLurwlCZoOZLTKzyQy+dRnjrQRHcAD7AAPdgWT798YIjmz7Wm9ver6mqsdOor/LnAys7lFnLlnfpzzWvS/w5bDrYYuZbSHYtl3v5Y/YdYT6CXIcnZvZhWEX13YLToAOxHp37zpRipnVmtkPw66vbcCjQK6Aht3/jZB7G+dqOxnY5O5tGfMz3wPJQYE+cqwh+I/dZWo4jfDI98vuPgP4GPAlC/vK3f2n7n58+FoH/q2f690B1HQ9MbOJ/XjtamC/HPP6unwq2783Cazvx/qHw1qCcO0yNVfD3t6nPqwG/sndR2U8atz9nnD+fwOzzewQgk9Fd+dY/91hF1dd5gnQfur5vv0tQRfRPHdvAD48wOX2x1pgbI/zG/vkaiy7KNBHjnuAfzCz8eFH1KsJ+k+7Tpjtb2YGbCPoakmZ2UFm9uGwr7UdaAvn9ceLwMFmdlj4H+jafrz2QWCimf21BSd1683sqHDeemBa10nBHP/eK81supnVsavPfaRdVXEfcLGZzTKzGiDneYdc71Me67gN+KyZHWWBWjP7aHiOAndvB+4Hfgo84+7v7uk/qh/qCY6e3zezsQR/l0Mq7DZ8CbjGzCrM7Hjgo0O93lKgQB85rgeWAssI/pj/FE6D4GTcw8B2YAnwH+7+GEH/+b8CGwk+vk4gOBGXN3d/HbguXP4bwB97f0W317YQnCz8WLj+NwhOcgL8LBxuMrM/ZXn5HQRdB48DbxPskP6qP7UPB3d/CLiRoKthZTjMJdf71Nc6lhL0o98MvB+u5+IezX4EHMrQnQzN5dtAI7AJeBIYaFdOfy0ETgjXew1wL8F5BemFvlgkUgTMbCqwApgYnpwsK+Glmy+4+zcKXctIpiN0kREu7Lb6ErCoXMLczOaF3XERMzuD4NzBLwtd10inQC9BFtwzZXuWR7+6Y0aiofi3hV+kybbM4epe6K22WoL++FPppf++BE0m6I5rIbhG/dPuvqywJY186nIRESkROkIXESkRBbs517hx43zatGmFWr2ISFF67rnnNrr7+GzzChbo06ZNY+nSpYVavYhIUTKznN9WVpeLiEiJUKCLiJQIBbqISInQLxaJlKhEIkFTUxPt7e2FLkUGoKqqiilTphCPx/N+jQJdpEQ1NTVRX1/PtGnTCO4XJsXC3dm0aRNNTU1Mnz4979epy0WkRLW3tzN27FiFeREyM8aOHdvvT1cKdJESpjAvXgN574ov0Ne/Co98A3ZsLHQlIiIjSvEF+qY34A//D1rW9d1WRArmkksuYcKECRxyyCE521x88cXcf//9u01fs2YNH//4x7O+Zv78+Vm/lHjXXXdxxRVXDLzgDNOmTWPjxuI7aCy+QI/XBsOOHYWtQ0R6dfHFF/PrX/96QK+dPHly1qAvFalUf39YLD99BrqZ3WFmG8zs5T7azTWzlJll360Olorw5y8TCnSRkeyEE05gzJgxfbZ7/PHHOfbYY5kxY0Y6xFetWpU+sm9ra+OCCy5g9uzZnH/++bS17frt6DvvvJMDDzyQE088kSeeeCI9vbm5mXPPPZe5c+cyd+7c9Lxrr72WSy65hPnz5zNjxgxuuummPus755xzOPLIIzn44IO59dZbAbj99tu58sor021uu+02vvSlLwHwk5/8hHnz5nHYYYfxmc98Jh3edXV1XH311Rx11FEsWbKkz/UORD6XLd5F8NNY/5mrgZlFCX6c+DeDU1Yv4mGgd7T23k5E0r7+P6/w6prB/W2MWZMbuOZjB+/xctauXcsf//hHVqxYwVlnnbVbV8v3v/99ampqWLZsGcuWLeOII45Iv+6aa67hueeeo7GxkZNOOonDDz8cgC9+8YtceeWVHH/88bz77rt85CMfYfny5QCsWLGC3//+97S0tHDQQQfxuc99rtdrve+44w7GjBlDW1sbc+fO5dxzz03vYL75zW8Sj8e58847+cEPfsDy5cu59957eeKJJ4jH43z+85/n7rvv5pOf/CQ7duzgkEMO4brrrtvjbZZLn4Hu7o+b2bQ+mv0V8HNg7iDU1LuKsMsloUAXKQXnnHMOkUiEWbNmsX79+t3mP/7443zhC18AYPbs2cyePRuAp59+mvnz5zN+fHDjwfPPP5/XX38dgIcffphXX301vYxt27bR0tICwEc/+lEqKyuprKxkwoQJrF+/nilTpuSs76abbuIXv/gFAKtXr+aNN97g6KOP5sMf/jAPPvggM2fOJJFIcOihh3LzzTfz3HPPMXduEIVtbW1MmDABgGg0yrnnnrtH26ove/zFIjPbG/gz4MP0EehmdjlwOcDUqVMHtsIK9aGL9NdgHEkPlcrKyvR4rh/cyXUJX67pnZ2dLFmyhOrq6l7XF41GSSaTOWt77LHHePjhh1myZAk1NTXMnz8/fW34ZZddxj//8z/zgQ98gE996lPp+i+66CL+5V/+ZbdlVVVVEY1Gc65rMAzGSdEbga+6e5+9/O5+q7vPcfc5XXvVfuvqctERukhZOOGEE7j77rsBePnll1m2LPgluqOOOorHHnuMTZs2kUgk+NnPfpZ+zWmnncbNN9+cfv7CCy8MaN1bt25l9OjR1NTUsGLFCp566qn0vKOOOorVq1fz05/+lIULFwJw8sknc//997NhwwYANm/ezDvv5Lzb7aAbjECfAywys1XAx4H/MLNzBmG52aWP0BXoIiPZwoULOeaYY3jttdeYMmUKt99++4CW87nPfY7t27en+6znzZsHwKRJk7j22ms55phjOOWUU9J96xB0kyxdupTZs2cza9YsbrnllgGte8GCBSSTSWbPns0//uM/cvTRR3ebf95553HccccxevRoAGbNmsX111/PaaedxuzZszn11FNZu3btgNY9EHn9pmjYh/6gu+e+oDRod1fYrs/rjebMmeMD/oGL68bBsVfAKdcO7PUiZWD58uXMnDmz0GWUtDPPPJMrr7ySk08+eUiWn+09NLPn3H1Otvb5XLZ4D7AEOMjMmszsUjP7rJl9dlAqHoiKGh2hi0jBbNmyhQMPPJDq6uohC/OByOcql4X5LszdL96javIVr9V16CJSMKNGjUpfUTOSFN83RSE8Qlegi4hkKs5Aj6vLRUSkp+IM9Io6XbYoItJDkQa6ulxERHoqzkCP1+gIXWQEW716NSeddBIzZ87k4IMP5jvf+U7Wdrp97uAqzt8UrahVH7rICBaLxfjWt77FEUccQUtLC0ceeSSnnnoqs2bNyuv15XD73KG4DUARH6Gry0VkpJo0aVL6m5v19fXMnDmT9957L2vbcr197lVXXcWsWbOYPXs2X/nKV/LbsH0o0iN0XeUi0i8PXQXrXhrcZU48FE7/1z6brVq1iueff56jjjoq6/xyvH3u5s2bufTSS1mxYgVmxpYtW/La5H0pzkCP10KyDTo7IVKcHzJEysH27ds599xzufHGG2loaMjaphxvn9vQ0EBVVRWXXXYZH/3oRznzzDPz3KK9K85Ar8i442JlXWFrESkGeRxJD7ZEIsG5557LhRdeyJ//+Z/nbFeOt8+NxWI888wzPPLIIyxatIibb76ZRx99NGcd+SrOw1vdE11kRHN3Lr30UmbOnJnuWx6oUrx97vbt29m6dStnnHEGN95444Dr66k4A73rh6J1YlRkRHriiSf48Y9/zKOPPsphhx3GYYcdxuLFiwe0rFK8fW5LSwtnnnkms2fP5sQTT+SGG24YUH095XX73KGwR7fPffWXcN8n4bNPwMRe7+grUrZ0+9yhV3S3zx2R4vpdUREpnKK9fe6I1HVSVH3oIlIAun3uYNLviorkpVBdqrLnBvLeFWeg63dFRfpUVVXFpk2bFOpFyN3ZtGkTVVVV/XpdcXa5pI/Q1eUiksuUKVNoamqiubm50KXIAFRVVfX6hadsijPQdYQu0qd4PM706dMLXYYMo+LuctERuohIWnEGejQOkbiuchERyVCcgQ6646KISA99BrqZ3WFmG8zs5RzzLzSzZeHjSTP74OCXmUW8Vl0uIiIZ8jlCvwtY0Mv8t4ET3X028A3g1kGoq286QhcR6abPq1zc/XEzm9bL/Ccznj4F9O86m4GqqFUfuohIhsHuQ78UeCjXTDO73MyWmtnSPb42tqJOgS4ikmHQAt3MTiII9K/mauPut7r7HHef0/UrIwNWUQcdLXu2DBGREjIogW5ms4EfAme7+6bBWGafKutg5/ZhWZWISDHY40A3s6nAfwF/4e7Dd/uxijroUKCLiHTp86Somd0DzAfGmVkTcA0QB3D3W4CrgbHAf4S/75fMdfP1QVVZryN0EZEM+VzlsrCP+ZcBlw1aRfmqqAuuQ+/shEjxfj9KRGSwFG8SVtYFQ3W7iIgAxRzoFQp0EZFMxRvolfXBUP3oIiJAMQd6+ghd16KLiEAxB3pXH7qO0EVEgGIOdPWhi4h0U7yBrj50EZFuijfQ1YcuItJN8Qa6+tBFRLop3kCP14BF1IcuIhIq3kA3C7pddIQuIgIUc6CD7okuIpKhuANd90QXEUkr7kDXPdFFRNKKO9B1hC4iklbcgV5RryN0EZFQcQd6ZR3s1ElREREo9kBXH7qISFpxB7r60EVE0oo70CvqIbUTUolCVyIiUnBFHui1wVD96CIiRR7o+qFoEZG0PgPdzO4wsw1m9nKO+WZmN5nZSjNbZmZHDH6ZOVTojosiIl3yOUK/C1jQy/zTgQPCx+XA9/e8rDxVNgRDdbmIiPQd6O7+OLC5lyZnA//pgaeAUWY2abAK7FVVV6BvG5bViYiMZIPRh743sDrjeVM4bTdmdrmZLTWzpc3NzXu+5q4j9Pate74sEZEiNxiBblmmebaG7n6ru89x9znjx4/f8zXrCF1EJG0wAr0J2Cfj+RRgzSAst29VjcGwXYEuIjIYgf4A8Mnwapejga3uvnYQltu3eA1YVF0uIiJArK8GZnYPMB8YZ2ZNwDVAHMDdbwEWA2cAK4FW4FNDVWyW4oJuF3W5iIj0HejuvrCP+Q785aBV1F9VjepyERGh2L8pCsGVLupyEREpgUCvalSXi4gIpRLo6nIRESmBQFeXi4gIUAqBrqtcRESAkgj0xuDmXJ2dha5ERKSgij/QKxsA11G6iJS94g903c9FRAQoiUDX/VxERKAUAl230BURAUoh0LuO0NXlIiJlrnQCXV0uIlLmij/Q1eUiIgKUQqCnr3JRoItIeSv+QI9VQqxKXS4iUvaKP9BB93MREaFUAl230BURKZVAb1CXi4iUvRIJ9FHQvqXQVYiIFFRpBHr1aGh7v9BViIgUVOkEeuvmQlchIlJQpRHoNWOCq1w6U4WuRESkYPIKdDNbYGavmdlKM7sqy/ypZvZ7M3vezJaZ2RmDX2ovqkcDrksXRaSs9RnoZhYFvgecDswCFprZrB7N/gG4z90PBy4A/mOwC+1V9ZhgqH50ESlj+RyhzwNWuvtb7t4BLALO7tHGgfA7+DQCawavxDxUjw6G6kcXkTKWT6DvDazOeN4UTst0LfAJM2sCFgN/lW1BZna5mS01s6XNzc0DKDeHGh2hi4jkE+iWZZr3eL4QuMvdpwBnAD82s92W7e63uvscd58zfvz4/lebS9cRepuO0EWkfOUT6E3APhnPp7B7l8qlwH0A7r4EqALGDUaBeUkHuo7QRaR85RPozwIHmNl0M6sgOOn5QI827wInA5jZTIJAH8Q+lT5UNQKmPnQRKWt9Brq7J4ErgN8AywmuZnnFzK4zs7PCZl8GPm1mLwL3ABe7e89umaETiUL1KHW5iEhZi+XTyN0XE5zszJx2dcb4q8Bxg1taP+nr/yJS5krjm6IQXIuuLhcRKWMlFOg6QheR8lY6gV4zRn3oIlLWSifQq0dDm+6JLiLlq4QCfUzwM3SpRKErEREpiBIK9K4vF+koXUTKU+kEevp+LupHF5HyVDqBXj0qGOpKFxEpU6UT6DVjg2HrpsLWISJSIKUT6LXh3Ru3byhsHSIiBVJ6gb5jY2HrEBEpkNIJ9FhlcNfFHTpCF5HyVDqBDsFRurpcRKRMlVigT1CXi4iUrdIK9Lrx6nIRkbJVWoFeO0FdLiJStkos0MdD+xZIdhS6EhGRYVdagV4XXrrYqn50ESk/pRXotROCobpdRKQMlVigd325qLmwdYiIFEBpBXqdAl1EyldpBbq6XESkjOUV6Ga2wMxeM7OVZnZVjjbnmdmrZvaKmf10cMvMU2UdxGt0hC4iZSnWVwMziwLfA04FmoBnzewBd381o80BwN8Bx7n7+2Y2YagK7lPtOB2hi0hZyucIfR6w0t3fcvcOYBFwdo82nwa+5+7vA7h74RK1doKO0EWkLOUT6HsDqzOeN4XTMh0IHGhmT5jZU2a2INuCzOxyM1tqZkubm4codOv0bVERKU/5BLplmeY9nseAA4D5wELgh2Y2arcXud/q7nPcfc748eP7W2t+GiZDy5qhWbaIyAiWT6A3AftkPJ8C9EzMJuCX7p5w97eB1wgCfvg1TA5+V7SjtSCrFxEplHwC/VngADObbmYVwAXAAz3a/DdwEoCZjSPognlrMAvNW8OUYLhNR+kiUl76DHR3TwJXAL8BlgP3ufsrZnadmZ0VNvsNsMnMXgV+D/yNuxfm15obJgfDbU0FWb2ISKH0edkigLsvBhb3mHZ1xrgDXwofhZUOdB2hi0h5Ka1vigI0hBfgbHuvsHWIiAyz0gv0eBXUjIWtCnQRKS+lF+gQdLuoy0VEykyJBvoUdbmISNkp0UCfrEAXkbJTmoHeuLe+XCQiZac0Az19pYv60UWkfJRooHddi65uFxEpH6UZ6I3h1/+3ru69nYhICSnRQJ8KkRhsfrvQlYiIDJuiC/QnVm7kY9/9I1tbE7kbRWMwaipsLsz9wURECqHoAr2uMsZL723l16+s7b3hmBkKdBEpK0UX6LOnNLLv2BoeeLGPK1jGzAi6XLznb3GIiJSmogt0M+PsD07myTc3sWFbe+6GY2bAzq3Qunn4ihMRKaCiC3SAsw6bjDs8uKyXbpcxM4Khul1EpEwUZaDvP6Gegyc38PM/9fIjFgp0ESkzRRnoAOfN2YdX1mzj5fe2Zm8waipYRIEuImWjaAP9nMP2piIW4b6lOb48FKsMvmCkQBeRMlG0gd5YE+eMQybyi+ffoz2Ryt5Ily6KSBkp2kAHOH/uVFrak/zyhRz3bBm7P2x8XZcuikhZKOpAP3rGGD4wsZ47n1iFZwvtvQ6Gndt0TxcRKQtFHehmxiXHTWfFuhaWvLVp9wZ7HRIM1708vIWJiBRAXoFuZgvM7DUzW2lmV/XS7uNm5mY2Z/BK7N1Zh01mTG0FP/xDlhtxTZgZDNe/MlzliIgUTJ+BbmZR4HvA6cAsYKGZzcrSrh74AvD0YBfZm6p4lE8dO41HV2zY/RLGynoYPR3W6whdREpfPkfo84CV7v6Wu3cAi4Czs7T7BvBNoJfv4w+Ni46bRkNVjO888sbuM/c6WIEuImUhn0DfG8g8q9gUTkszs8OBfdz9wd4WZGaXm9lSM1va3Nzc72JzaaiKc8nx0/ndq+t3P0qfeChselO/LyoiJS+fQLcs09KXlJhZBLgB+HJfC3L3W919jrvPGT9+fP5V5uGS46czpraCf/rV8u5XvOx1cFDuhuWDuj4RkZEmn0BvAvbJeD4FyLx3bT1wCPCYma0CjgYeGM4ToxAcpV95ygEseWsTv3t1/a4Zex0cDNctG85yRESGXT6B/ixwgJlNN7MK4ALgga6Z7r7V3ce5+zR3nwY8BZzl7kuHpOJeLJw3lf0n1HH9r5bT1hF+e3T0dKgZB6uH9VytiMiw6zPQ3T0JXAH8BlgO3Ofur5jZdWZ21lAX2B+xaIRvnH0I725u5YaHXw8mmsHUo+HdJYUtTkRkiMXyaeTui4HFPaZdnaPt/D0va+CO2W8sC+dN5Yd/eIvTD5nI4VNHw77HwooHYdtaaJhUyPJERIZMUX9TNJe/O+MDTGqs5guLnmdbewKmHhPM0FG6iJSwkgz0hqo4Ny08nDVb2vnq/cvwiYdCvFaBLiIlrSQDHeDIfUdz1YIP8NDL67jh0bdhn7nwjgJdREpXyQY6wGUfms55c6Zw0yNvsCz+QVj/ErSsK3RZIiJDoqQD3cy4/pxD+dAB4/jKS+GXW19b3PuLRESKVEkHOkBFLMKtfzGHMfseytude/HeU/cXuiQRkSFR8oEOUF0R5c5PHcWKUScwvvkpvv0/S0l16leMRKS0lEWgQxDqp/7Zp6iwFKuW/IKL73yGdVuH/caQIiJDpmwCHSC279HQMIW/n7SUZ1dt5tQb/pf7lq7O/vN1IiJFpqwCnUgU5lzMxE1P8fAnJzNzUgN/e/8yPnnHM6xYt63Q1YmI7JHyCnSAIy6CSJwpK+9h0aeP5utnHcwLq7dw+nf+wBcXPc+qjTsKXaGIyICUX6DXTYBZZ8MLdxNpf5+Ljp3GH/72JD5zwn785pV1fPhbj/GZHy/lyZUb1RUjIkXFChVac+bM8aVLh/0Ou4H1r8Itx8G8z8Dp/5qevGFbO3c8sYp7n32X91sT7De+lj87fG/OnD2ZaeNqC1OriEgGM3vO3bP+3kR5BjrAA1+AF+6Gzz8N4/bvNqs9keLBZWtZ9My7LH3nfQAO3buRkz4wgRMOGMcH9xlFPFp+H25EpPAU6Nm0rIfvHhn8otHFD0I0nrXZmi1t/GrZWha/vJYXV2+h06G+MsYx+43l2P3GcvjU0cyc1EBFTAEvIkNPgZ7LS/fDzy+FY/8KTru+z+ZbWxM8+eZGHn9jI4+/3sx7W9qA4Nuoh0xu4LB9RnPw5AYOmljP/hPqqIpHh/pfICJlprdAz+sHLkrWoR+Hd56EJ78LtePhuC/22ryxJs7ph07i9EMn4e6s3drOC6u38Py77/P8u1u4++l32JnsBCBiMG1sLQfuVc+BE+vZb3wtU8fUsO/YWkbXxDHL9tvbIiIDV96BDnD6v0H7Fvjd1bB9A5x8DcQq+nyZmTF5VDWTR1VzxqHBryAlU528s7mV19a18Nq6Fl5fHwx/++o6Mu80UF8VY9+xNew7ppapY2vYd0wNk0dVM6mxikmjqqmr1NsiIv1X3l0uXTpT8NBX4dnbYOJsWPAvMO34QVt8eyLF6s2tvLOplXc2t/LOph28s6mVdze3snpzK8ke95Wpr4oxubGaSaOqmNQYBn1jFRMaqhhfV8n4+krG1FYQjegoX6TcqA89X8sfhMVfgZa1MPVYOPIiOOA0qBkzZKtMpjpZu7U9fLQFwy1trOl6vqWdTTs6dntdxGBMbRDu4+sr00E/rq4iPW1sbSWja+OMrqnQVTkiJUKB3h+JNlh6BzxzG7z/NmCw9xEw46RguNchMGoqDGMfeHsixfpt7TS37KS5ZScbtwfD5q5hy042bu+guWUnHanOrMuor4oxuqaC0bUVjKmJM7q2gtE1FYxJD+OMyng+qiaunYDICLTHgW5mC81v75IAAA0cSURBVIDvAFHgh+7+rz3mfwm4DEgCzcAl7v5Ob8scsYHepbMT3lsKbz4KKx8Jxj0My8pGGDsDRk+HMdO7D+snQaQwQejubGtL0ry9nQ0tO3l/R4LNrR28v6OD98Ph5tZEMNzRwZbWDnZ0pHIur6YiSkNVnMbqOA3VsWBYFaehOnxUhdPC6Znt6ipjOvErMgT2KNDNLAq8DpwKNAHPAgvd/dWMNicBT7t7q5l9Dpjv7uf3ttwRH+g9dewIvmG6bhmsfwU2vxUcwW9ZDZ4RipEY1E+GxinQuDc07B2Mdw0bp0D16GE9wu9NeyLFltYEm7tCP70DSLCtLcHWtgTb2sNhWzI93tKe7HW5ESMd9HWVseBRFaM2HK+vilFbEUyrq4xSVxnvNl5bGaU+HMb0SUEkbU8vW5wHrHT3t8KFLQLOBtKB7u6/z2j/FPCJgZc7QlXUBj80vc/c7tNTCdi6Gja/De+vgq1NwWPbe7D6Gdi2BjoT3V8Tqw7CvnEKNITBXz8R6iZC/V7BUX7tBIgO/dUuVfEoExujTGys6tfrUp3O9vZkRtjvCv9tbcluO4IdO5O0tCfZ0NLO9uYk23em2L4zQXsie/fQ7jVGgsCvjAY7hYpgx1BTEQ0fMaorotTEo9RkTK+OB/NqKsN28bBdOC+ik8pSYvJJjL2B1RnPm4Cjeml/KfDQnhRVVKJxGDMjeGTT2Qk7NsDW92BbUzDc2rRr/M1Hwh+u7vlJyYJr4+v3CoM+45H5vG6vnN9yHUrRiNFYE6exJs4+A1xGMtXJjp0ptnck2d6eZPvORBD27clgJ7AzGG7vemRM39DSTmtHiraOFDt2JmlLpEik+nc+KAj8aDrkayoydgYVMWriwbyqeJSqeCQYxiLpaZWxXdOr4z3aheMV0Yi6nmTY5BPo2f4as/7PMbNPAHOAE3PMvxy4HGDq1Kl5lljkIpFd4cuR2dukErCjObi6pmV9MNy+Pgj6lnWwfV3Q1bOjeVc/fqbq0VAzLtgB1I7NGB8XPDKfV48ZliP/fMSiERprIjTWDM4OKZHq3BXyHUnaOlK0dqRoDcd3dKRo60iG01K0JcKdQVe7RDB/7dZEt9e1JzvpSOb3aaInM6gKg78r9Cu7gj/WtcMIxivTO4ZIuMOIUBmLUBELx8MdRGU4ryKcXxnb1bYyFk230yeQ8pPP/+wm6HYQNgVY07ORmZ0CfA040d13ZluQu98K3ApBH3q/qy1V0Tg0TA4evUklg1Dfvq578O9oDh+bYOMbsONJaN1M9v2uBTuA2nHBsHo0VI3aNV49Kvu0qlEjZkeQSzwaobE6QmP14H9i6ex02pMp2hOdtCdStCeCHUJ7opOdiRTtyRRtHeG8Hu3aw3ZtGeM7k8GOp6Ul0aNtMN7zuwkDEY8albFoRvB37QR23yGkx+MRKqLBTqHnvIpYsKOoiEWIRyPEo5aeFs+YvquNEe96jXYwwyKf/6HPAgeY2XTgPeAC4P9mNjCzw4EfAAvcfcOgVymBaAwaJgWPvnSmglBv3Qg7NgaB37opDP6NwfS294M+/vWvBt+W3dnHrzZV1GcEfCNU1u/+qMgyrbJh13i8esScEO6PSMTCLpnhWV8iFXwq6Eh2sjMZ7AAyx3d2jSc66UgFO5WuaR0ZbdLj6XbhssLxlvZklvbB8/52YfUlGrEw/G1X+Kd3Dl07DMu9c4hGuu1A4hnzuk1P73iMWCRCLBqsNxaNEIsEbWORYHmxcLnxSMZ41Iq2m6zPQHf3pJldAfyG4LLFO9z9FTO7Dljq7g8A/w7UAT8LN8S77n7WENYtfYlEoW588MhXKgHtW6FtSxD27eGw63nmtJ0twcngnS3Bo33b7id/s7FoRrjXQEUNxGuDoO8ar6gJnqfHazLaZhmPV0OsEmJVEK0oyh1GT12BVVtZuBo6O73bTmBnspNEqpNEyoOdTaozveNJhONBG08/T7dLOh2pVPfXhq/rSHXSkfRuO7EdO5N0pJyOcMfSbVnh+CB8iMkpGrEg/MPQj0UjxCPBJ46unUF6h9BjZxCLBDuVeMTSr60IdybBuDFv+liOP2DcoNed12dod18MLO4x7eqM8VMGuS4phGh8V7/7QCR3hgG/bVfQpx9ZpnXsCL7IlWgNPjlsCccTrdDRCsm2gdURq4Jo5a6QTw8rejyvzNEunBaJB9skGg92FJFYxnhv82LBMBrvMR4v2HcUBiISMaoi0fCuocN/4r0vqU7P2Il037l0JIOdUTK1aweT7AzGkxk7nGSnp3dSyfB5RzJom0x1LcPTr02Ez3u+vmsnlMjSNtnZGS7T08v83HwvXKCL5KUrCAe6Q+ipszMI9Y5WSITh39GaEfo7gmGyPdiZpIc7ezzPGKY6gp3Jjubcr8l+zn9wWHT3sI9Eg4dFg2mRcGiRHs+jOdr29TyPZVmkx8PC12Sbl8cj0tv8aPd1dHtddPcasr0OI2oRomZUAUQjEDOozHxNHLBdy8kcL+CnOHdnqL6gr0CXkSsSCa7/r6gF+tF1tCfcg66nZDt0JoPxVEfQnZQKH7nGUx0De42ngp1XZzIcTwbnQDpTuz9PdUAi2Xubbs+TwZVR3Z7n/nZw2ekZ9H2OWzgeyWM8e3uzCHbkRXDMXw76P0eBLpLJLOyaGaazn4XgHoZ8RuB3e4TzPZVlXji/M9e8Ph6dqYzl93ykeqy/5+u62nj3fwfeYzxcxm7jnmN6P9pnXV/mOHm08eAy4iGgQBcpN2a7ulwo4R1XGSqeMzQiItIrBbqISIlQoIuIlAgFuohIiVCgi4iUCAW6iEiJUKCLiJQIBbqISInI60eih2TFZs1Arz8k3YtxwMZBLGcwjdTaVFf/jNS6YOTWprr6Z6B17evuWb9qWrBA3xNmtjTXj6QW2kitTXX1z0itC0Zubaqrf4aiLnW5iIiUCAW6iEiJKNZAv7XQBfRipNamuvpnpNYFI7c21dU/g15XUfahi4jI7or1CF1ERHpQoIuIlIiiC3QzW2Bmr5nZSjO7qoB17GNmvzez5Wb2ipl9MZx+rZm9Z2YvhI8zClDbKjN7KVz/0nDaGDP7nZm9EQ5HF6CugzK2ywtmts3M/roQ28zM7jCzDWb2csa0rNvIAjeFf3PLzOyIYa7r381sRbjuX5jZqHD6NDNry9hutwxzXTnfNzP7u3B7vWZmHxmqunqp7d6MulaZ2Qvh9OHcZrkyYuj+zoIfLC2OBxAF3gRmEPzUyovArALVMgk4IhyvB14HZgHXAl8p8HZaBYzrMe2bwFXh+FXAv42A93IdsG8hthlwAnAE8HJf2wg4A3gIMOBo4Olhrus0IBaO/1tGXdMy2xVge2V938L/By8ClcD08P9sdDhr6zH/W8DVBdhmuTJiyP7Oiu0IfR6w0t3fcvcOYBFwdiEKcfe17v6ncLwFWA7sXYha8nQ28KNw/EfAOQWsBeBk4E13H+i3hfeIuz8ObO4xOdc2Ohv4Tw88BYwys0nDVZe7/9bdk+HTp4ApQ7Hu/tbVi7OBRe6+093fBlYS/N8d9trMzIDzgHuGav259JIRQ/Z3VmyBvjewOuN5EyMgRM1sGnA48HQ46YrwI9MdhejaIPip2t+a2XNmdnk4bS93XwvBHxowoQB1ZbqA7v/JCr3NIPc2Gkl/d5cQHMV1mW5mz5vZ/5rZhwpQT7b3bSRtrw8B6939jYxpw77NemTEkP2dFVugW5ZpBb3u0szqgJ8Df+3u24DvA/sBhwFrCT7uDbfj3P0I4HTgL83shALUkJOZVQBnAT8LJ42EbdabEfF3Z2ZfA5LA3eGktcBUdz8c+BLwUzNrGMaScr1vI2J7hRbS/cBh2LdZlozI2TTLtH5tt2IL9CZgn4znU4A1BaoFM4sTvFF3u/t/Abj7endPuXsncBtD+FEzF3dfEw43AL8Ia1jf9fEtHG4Y7roynA78yd3Xw8jYZqFc26jgf3dmdhFwJnChhx2uYZfGpnD8OYK+6gOHq6Ze3reCby8AM4sBfw7c2zVtuLdZtoxgCP/Oii3QnwUOMLPp4VHeBcADhSgk7Ju7HVju7t/OmJ7Z5/VnwMs9XzvEddWaWX3XOMEJtZcJttNFYbOLgF8OZ109dDtqKvQ2y5BrGz0AfDK8CuFoYGvXR+bhYGYLgK8CZ7l7a8b08WYWDcdnAAcAbw1jXbnetweAC8ys0symh3U9M1x1ZTgFWOHuTV0ThnOb5coIhvLvbDjO9g7ymeMzCM4Wvwl8rYB1HE/wcWgZ8EL4OAP4MfBSOP0BYNIw1zWD4AqDF4FXurYRMBZ4BHgjHI4p0HarATYBjRnThn2bEexQ1gIJgiOjS3NtI4KPwt8L/+ZeAuYMc10rCfpWu/7Obgnbnhu+xy8CfwI+Nsx15XzfgK+F2+s14PThfi/D6XcBn+3Rdji3Wa6MGLK/M331X0SkRBRbl4uIiOSgQBcRKREKdBGREqFAFxEpEQp0EZESoUAXESkRCnQRkRLx/wF57Pv2JzzmBQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(model_1.history[\"loss\"])\n",
    "plt.plot(model_2.history[\"loss\"])\n",
    "plt.title(\"loss_function_nn_disney - Training\")\n",
    "plt.legend([\"1 hidden layer\", \"2 hidden layers\"])\n",
    "plt.savefig('../Images/loss_function_nn_disney - Training.png')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 682,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Train vs test for shallow net\n",
    "plt.plot(model_1.history[\"loss\"])\n",
    "plt.plot(model_1.history[\"val_loss\"])\n",
    "plt.title(\"loss_function_nn_disney - Training Vs. Validation - 1 hidden layer\")\n",
    "plt.legend([\"train\", \"test\"])\n",
    "plt.savefig('../Images/loss_function_nn_disney - Training Vs. Validation - 1 hidden layer.png')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 683,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Train vs test for deep net\n",
    "plt.plot(model_2.history[\"loss\"])\n",
    "plt.plot(model_2.history[\"val_loss\"])\n",
    "plt.title(\"../Data/loss_function_nn_disney - Training Vs. Validation - 2 hidden layers\")\n",
    "plt.savefig('../Images/loss_function_nn_disney - Training Vs. Validation - 2 hidden layers.png')\n",
    "plt.legend([\"train\", \"test\"])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 684,
   "metadata": {},
   "outputs": [],
   "source": [
    "import h5py\n",
    "\n",
    "nn.save('../Models/disney_nn_loss_2_layers.h5')  # creates a HDF5 file 'my_model.h5'\n",
    "del nn  # deletes the existing model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 706,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define the model - deep neural net\n",
    "number_input_features = 4\n",
    "hidden_nodes_layer1 = 12\n",
    "hidden_nodes_layer2 = 8\n",
    "hidden_nodes_layer3 = 3\n",
    "\n",
    "\n",
    "nn = Sequential()\n",
    "# First hidden layer\n",
    "nn.add(\n",
    "    Dense(units=hidden_nodes_layer1, input_dim=number_input_features, activation=\"relu\")\n",
    ")\n",
    "# Second hidden layer\n",
    "nn.add(Dense(units=hidden_nodes_layer2, activation=\"sigmoid\"))\n",
    "\n",
    "# Third hidden layer\n",
    "nn.add(Dense(units=hidden_nodes_layer3, activation=\"tanh\"))\n",
    "\n",
    "# Output layer\n",
    "nn.add(Dense(units=1, activation=\"linear\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 719,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_64\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_203 (Dense)            (None, 12)                60        \n",
      "_________________________________________________________________\n",
      "dense_204 (Dense)            (None, 8)                 104       \n",
      "_________________________________________________________________\n",
      "dense_205 (Dense)            (None, 3)                 27        \n",
      "_________________________________________________________________\n",
      "dense_206 (Dense)            (None, 1)                 4         \n",
      "=================================================================\n",
      "Total params: 195\n",
      "Trainable params: 195\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# Compile model\n",
    "nn.compile(loss=\"mean_squared_error\", optimizer=\"adam\", metrics=[\"mse\"])\n",
    "nn.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 689,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Creating validation data sets\n",
    "#from sklearn.model_selection import train_test_split\n",
    "#X_train, X_val, y_train, y_val = train_test_split(X_train, y_train, test_size = 0.3, shuffle=False) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 721,
   "metadata": {
    "tags": [
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend",
     "outputPrepend"
    ]
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 9 samples, validate on 5 samples\n",
      "Epoch 1/200\n",
      "9/9 [==============================] - 0s 30ms/sample - loss: 0.3889 - mse: 0.3889 - val_loss: 0.4965 - val_mse: 0.4965\n",
      "Epoch 2/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.3147 - mse: 0.3147 - val_loss: 0.4229 - val_mse: 0.4229\n",
      "Epoch 3/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.2660 - mse: 0.2660 - val_loss: 0.3716 - val_mse: 0.3716\n",
      "Epoch 4/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.2368 - mse: 0.2368 - val_loss: 0.3379 - val_mse: 0.3379\n",
      "Epoch 5/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.2207 - mse: 0.2207 - val_loss: 0.3164 - val_mse: 0.3164\n",
      "Epoch 6/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.2122 - mse: 0.2122 - val_loss: 0.3026 - val_mse: 0.3026\n",
      "Epoch 7/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.2077 - mse: 0.2077 - val_loss: 0.2933 - val_mse: 0.2933\n",
      "Epoch 8/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.2049 - mse: 0.2049 - val_loss: 0.2865 - val_mse: 0.2865\n",
      "Epoch 9/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.2029 - mse: 0.2029 - val_loss: 0.2814 - val_mse: 0.2814\n",
      "Epoch 10/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.2011 - mse: 0.2011 - val_loss: 0.2772 - val_mse: 0.2772\n",
      "Epoch 11/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1995 - mse: 0.1995 - val_loss: 0.2736 - val_mse: 0.2736\n",
      "Epoch 12/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1978 - mse: 0.1978 - val_loss: 0.2705 - val_mse: 0.2705\n",
      "Epoch 13/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1963 - mse: 0.1963 - val_loss: 0.2677 - val_mse: 0.2677\n",
      "Epoch 14/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1948 - mse: 0.1948 - val_loss: 0.2653 - val_mse: 0.2653\n",
      "Epoch 15/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1934 - mse: 0.1934 - val_loss: 0.2631 - val_mse: 0.2631\n",
      "Epoch 16/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1920 - mse: 0.1920 - val_loss: 0.2610 - val_mse: 0.2610\n",
      "Epoch 17/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1907 - mse: 0.1907 - val_loss: 0.2592 - val_mse: 0.2592\n",
      "Epoch 18/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1895 - mse: 0.1895 - val_loss: 0.2573 - val_mse: 0.2573\n",
      "Epoch 19/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1883 - mse: 0.1883 - val_loss: 0.2555 - val_mse: 0.2555\n",
      "Epoch 20/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1871 - mse: 0.1871 - val_loss: 0.2539 - val_mse: 0.2539\n",
      "Epoch 21/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1859 - mse: 0.1859 - val_loss: 0.2527 - val_mse: 0.2527\n",
      "Epoch 22/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1848 - mse: 0.1848 - val_loss: 0.2514 - val_mse: 0.2514\n",
      "Epoch 23/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1838 - mse: 0.1838 - val_loss: 0.2499 - val_mse: 0.2499\n",
      "Epoch 24/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1828 - mse: 0.1828 - val_loss: 0.2486 - val_mse: 0.2486\n",
      "Epoch 25/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1818 - mse: 0.1818 - val_loss: 0.2474 - val_mse: 0.2474\n",
      "Epoch 26/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1808 - mse: 0.1808 - val_loss: 0.2464 - val_mse: 0.2464\n",
      "Epoch 27/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1799 - mse: 0.1799 - val_loss: 0.2454 - val_mse: 0.2454\n",
      "Epoch 28/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1790 - mse: 0.1790 - val_loss: 0.2444 - val_mse: 0.2444\n",
      "Epoch 29/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1781 - mse: 0.1781 - val_loss: 0.2435 - val_mse: 0.2435\n",
      "Epoch 30/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1772 - mse: 0.1772 - val_loss: 0.2427 - val_mse: 0.2427\n",
      "Epoch 31/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1763 - mse: 0.1763 - val_loss: 0.2417 - val_mse: 0.2417\n",
      "Epoch 32/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1754 - mse: 0.1754 - val_loss: 0.2410 - val_mse: 0.2410\n",
      "Epoch 33/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1746 - mse: 0.1746 - val_loss: 0.2404 - val_mse: 0.2404\n",
      "Epoch 34/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1737 - mse: 0.1737 - val_loss: 0.2398 - val_mse: 0.2398\n",
      "Epoch 35/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1729 - mse: 0.1729 - val_loss: 0.2392 - val_mse: 0.2392\n",
      "Epoch 36/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1721 - mse: 0.1721 - val_loss: 0.2384 - val_mse: 0.2384\n",
      "Epoch 37/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1714 - mse: 0.1714 - val_loss: 0.2379 - val_mse: 0.2379\n",
      "Epoch 38/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1705 - mse: 0.1705 - val_loss: 0.2375 - val_mse: 0.2375\n",
      "Epoch 39/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1698 - mse: 0.1698 - val_loss: 0.2371 - val_mse: 0.2371\n",
      "Epoch 40/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1690 - mse: 0.1690 - val_loss: 0.2366 - val_mse: 0.2366\n",
      "Epoch 41/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1682 - mse: 0.1682 - val_loss: 0.2361 - val_mse: 0.2361\n",
      "Epoch 42/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1675 - mse: 0.1675 - val_loss: 0.2357 - val_mse: 0.2357\n",
      "Epoch 43/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1667 - mse: 0.1667 - val_loss: 0.2354 - val_mse: 0.2354\n",
      "Epoch 44/200\n",
      "9/9 [==============================] - 0s 3ms/sample - loss: 0.1660 - mse: 0.1660 - val_loss: 0.2352 - val_mse: 0.2352\n",
      "Epoch 45/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1653 - mse: 0.1653 - val_loss: 0.2349 - val_mse: 0.2349\n",
      "Epoch 46/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1646 - mse: 0.1646 - val_loss: 0.2345 - val_mse: 0.2345\n",
      "Epoch 47/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1639 - mse: 0.1639 - val_loss: 0.2343 - val_mse: 0.2343\n",
      "Epoch 48/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1632 - mse: 0.1632 - val_loss: 0.2342 - val_mse: 0.2342\n",
      "Epoch 49/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1625 - mse: 0.1625 - val_loss: 0.2340 - val_mse: 0.2340\n",
      "Epoch 50/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1618 - mse: 0.1618 - val_loss: 0.2337 - val_mse: 0.2337\n",
      "Epoch 51/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1611 - mse: 0.1611 - val_loss: 0.2335 - val_mse: 0.2335\n",
      "Epoch 52/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1604 - mse: 0.1604 - val_loss: 0.2334 - val_mse: 0.2334\n",
      "Epoch 53/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1598 - mse: 0.1598 - val_loss: 0.2334 - val_mse: 0.2334\n",
      "Epoch 54/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1591 - mse: 0.1591 - val_loss: 0.2334 - val_mse: 0.2334\n",
      "Epoch 55/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1584 - mse: 0.1584 - val_loss: 0.2333 - val_mse: 0.2333\n",
      "Epoch 56/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1577 - mse: 0.1577 - val_loss: 0.2331 - val_mse: 0.2331\n",
      "Epoch 57/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1571 - mse: 0.1571 - val_loss: 0.2330 - val_mse: 0.2330\n",
      "Epoch 58/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1565 - mse: 0.1565 - val_loss: 0.2330 - val_mse: 0.2330\n",
      "Epoch 59/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1558 - mse: 0.1558 - val_loss: 0.2331 - val_mse: 0.2331\n",
      "Epoch 60/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1551 - mse: 0.1551 - val_loss: 0.2331 - val_mse: 0.2331\n",
      "Epoch 61/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1545 - mse: 0.1545 - val_loss: 0.2331 - val_mse: 0.2331\n",
      "Epoch 62/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1538 - mse: 0.1538 - val_loss: 0.2331 - val_mse: 0.2331\n",
      "Epoch 63/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1533 - mse: 0.1533 - val_loss: 0.2330 - val_mse: 0.2330\n",
      "Epoch 64/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1526 - mse: 0.1526 - val_loss: 0.2332 - val_mse: 0.2332\n",
      "Epoch 65/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1520 - mse: 0.1520 - val_loss: 0.2334 - val_mse: 0.2334\n",
      "Epoch 66/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1514 - mse: 0.1514 - val_loss: 0.2335 - val_mse: 0.2335\n",
      "Epoch 67/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1508 - mse: 0.1508 - val_loss: 0.2335 - val_mse: 0.2335\n",
      "Epoch 68/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1502 - mse: 0.1502 - val_loss: 0.2336 - val_mse: 0.2336\n",
      "Epoch 69/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1496 - mse: 0.1496 - val_loss: 0.2338 - val_mse: 0.2338\n",
      "Epoch 70/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1490 - mse: 0.1490 - val_loss: 0.2340 - val_mse: 0.2340\n",
      "Epoch 71/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1484 - mse: 0.1484 - val_loss: 0.2341 - val_mse: 0.2341\n",
      "Epoch 72/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1478 - mse: 0.1478 - val_loss: 0.2341 - val_mse: 0.2341\n",
      "Epoch 73/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1473 - mse: 0.1473 - val_loss: 0.2343 - val_mse: 0.2343\n",
      "Epoch 74/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1466 - mse: 0.1466 - val_loss: 0.2346 - val_mse: 0.2346\n",
      "Epoch 75/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1461 - mse: 0.1461 - val_loss: 0.2348 - val_mse: 0.2348\n",
      "Epoch 76/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1455 - mse: 0.1455 - val_loss: 0.2348 - val_mse: 0.2348\n",
      "Epoch 77/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1449 - mse: 0.1449 - val_loss: 0.2349 - val_mse: 0.2349\n",
      "Epoch 78/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1443 - mse: 0.1443 - val_loss: 0.2351 - val_mse: 0.2351\n",
      "Epoch 79/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1438 - mse: 0.1438 - val_loss: 0.2352 - val_mse: 0.2352\n",
      "Epoch 80/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1431 - mse: 0.1431 - val_loss: 0.2353 - val_mse: 0.2353\n",
      "Epoch 81/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1426 - mse: 0.1426 - val_loss: 0.2354 - val_mse: 0.2354\n",
      "Epoch 82/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1420 - mse: 0.1420 - val_loss: 0.2356 - val_mse: 0.2356\n",
      "Epoch 83/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1414 - mse: 0.1414 - val_loss: 0.2358 - val_mse: 0.2358\n",
      "Epoch 84/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1408 - mse: 0.1408 - val_loss: 0.2359 - val_mse: 0.2359\n",
      "Epoch 85/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1403 - mse: 0.1403 - val_loss: 0.2358 - val_mse: 0.2358\n",
      "Epoch 86/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1397 - mse: 0.1397 - val_loss: 0.2358 - val_mse: 0.2358\n",
      "Epoch 87/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1391 - mse: 0.1391 - val_loss: 0.2359 - val_mse: 0.2359\n",
      "Epoch 88/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1386 - mse: 0.1386 - val_loss: 0.2359 - val_mse: 0.2359\n",
      "Epoch 89/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1380 - mse: 0.1380 - val_loss: 0.2357 - val_mse: 0.2357\n",
      "Epoch 90/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1375 - mse: 0.1375 - val_loss: 0.2356 - val_mse: 0.2356\n",
      "Epoch 91/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1369 - mse: 0.1369 - val_loss: 0.2358 - val_mse: 0.2358\n",
      "Epoch 92/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1364 - mse: 0.1364 - val_loss: 0.2358 - val_mse: 0.2358\n",
      "Epoch 93/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1358 - mse: 0.1358 - val_loss: 0.2355 - val_mse: 0.2355\n",
      "Epoch 94/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1354 - mse: 0.1354 - val_loss: 0.2358 - val_mse: 0.2358\n",
      "Epoch 95/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1349 - mse: 0.1349 - val_loss: 0.2366 - val_mse: 0.2366\n",
      "Epoch 96/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1343 - mse: 0.1343 - val_loss: 0.2374 - val_mse: 0.2374\n",
      "Epoch 97/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1339 - mse: 0.1339 - val_loss: 0.2377 - val_mse: 0.2377\n",
      "Epoch 98/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1334 - mse: 0.1334 - val_loss: 0.2378 - val_mse: 0.2378\n",
      "Epoch 99/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1330 - mse: 0.1330 - val_loss: 0.2383 - val_mse: 0.2383\n",
      "Epoch 100/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1324 - mse: 0.1324 - val_loss: 0.2389 - val_mse: 0.2389\n",
      "Epoch 101/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1319 - mse: 0.1319 - val_loss: 0.2393 - val_mse: 0.2393\n",
      "Epoch 102/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1315 - mse: 0.1315 - val_loss: 0.2393 - val_mse: 0.2393\n",
      "Epoch 103/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1311 - mse: 0.1311 - val_loss: 0.2396 - val_mse: 0.2396\n",
      "Epoch 104/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1305 - mse: 0.1305 - val_loss: 0.2401 - val_mse: 0.2401\n",
      "Epoch 105/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1301 - mse: 0.1301 - val_loss: 0.2404 - val_mse: 0.2404\n",
      "Epoch 106/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1296 - mse: 0.1296 - val_loss: 0.2408 - val_mse: 0.2408\n",
      "Epoch 107/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1291 - mse: 0.1291 - val_loss: 0.2412 - val_mse: 0.2412\n",
      "Epoch 108/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1287 - mse: 0.1287 - val_loss: 0.2413 - val_mse: 0.2413\n",
      "Epoch 109/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1282 - mse: 0.1282 - val_loss: 0.2418 - val_mse: 0.2418\n",
      "Epoch 110/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1277 - mse: 0.1277 - val_loss: 0.2422 - val_mse: 0.2422\n",
      "Epoch 111/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1273 - mse: 0.1273 - val_loss: 0.2424 - val_mse: 0.2424\n",
      "Epoch 112/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1268 - mse: 0.1268 - val_loss: 0.2426 - val_mse: 0.2426\n",
      "Epoch 113/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1264 - mse: 0.1264 - val_loss: 0.2429 - val_mse: 0.2429\n",
      "Epoch 114/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1259 - mse: 0.1259 - val_loss: 0.2432 - val_mse: 0.2432\n",
      "Epoch 115/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1254 - mse: 0.1254 - val_loss: 0.2435 - val_mse: 0.2435\n",
      "Epoch 116/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1250 - mse: 0.1250 - val_loss: 0.2437 - val_mse: 0.2437\n",
      "Epoch 117/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1245 - mse: 0.1245 - val_loss: 0.2438 - val_mse: 0.2438\n",
      "Epoch 118/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1241 - mse: 0.1241 - val_loss: 0.2440 - val_mse: 0.2440\n",
      "Epoch 119/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1236 - mse: 0.1236 - val_loss: 0.2442 - val_mse: 0.2442\n",
      "Epoch 120/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1232 - mse: 0.1232 - val_loss: 0.2442 - val_mse: 0.2442\n",
      "Epoch 121/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1228 - mse: 0.1228 - val_loss: 0.2443 - val_mse: 0.2443\n",
      "Epoch 122/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1224 - mse: 0.1224 - val_loss: 0.2445 - val_mse: 0.2445\n",
      "Epoch 123/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1219 - mse: 0.1219 - val_loss: 0.2447 - val_mse: 0.2447\n",
      "Epoch 124/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1215 - mse: 0.1215 - val_loss: 0.2446 - val_mse: 0.2446\n",
      "Epoch 125/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1211 - mse: 0.1211 - val_loss: 0.2447 - val_mse: 0.2447\n",
      "Epoch 126/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1207 - mse: 0.1207 - val_loss: 0.2448 - val_mse: 0.2448\n",
      "Epoch 127/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1203 - mse: 0.1203 - val_loss: 0.2449 - val_mse: 0.2449\n",
      "Epoch 128/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1199 - mse: 0.1199 - val_loss: 0.2452 - val_mse: 0.2452\n",
      "Epoch 129/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1195 - mse: 0.1195 - val_loss: 0.2454 - val_mse: 0.2454\n",
      "Epoch 130/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1190 - mse: 0.1190 - val_loss: 0.2453 - val_mse: 0.2453\n",
      "Epoch 131/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1187 - mse: 0.1187 - val_loss: 0.2451 - val_mse: 0.2451\n",
      "Epoch 132/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1183 - mse: 0.1183 - val_loss: 0.2451 - val_mse: 0.2451\n",
      "Epoch 133/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1178 - mse: 0.1178 - val_loss: 0.2452 - val_mse: 0.2452\n",
      "Epoch 134/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1174 - mse: 0.1174 - val_loss: 0.2455 - val_mse: 0.2455\n",
      "Epoch 135/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1170 - mse: 0.1170 - val_loss: 0.2456 - val_mse: 0.2456\n",
      "Epoch 136/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1166 - mse: 0.1166 - val_loss: 0.2454 - val_mse: 0.2454\n",
      "Epoch 137/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1162 - mse: 0.1162 - val_loss: 0.2454 - val_mse: 0.2454\n",
      "Epoch 138/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1157 - mse: 0.1157 - val_loss: 0.2454 - val_mse: 0.2454\n",
      "Epoch 139/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1153 - mse: 0.1153 - val_loss: 0.2454 - val_mse: 0.2454\n",
      "Epoch 140/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1149 - mse: 0.1149 - val_loss: 0.2453 - val_mse: 0.2453\n",
      "Epoch 141/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1145 - mse: 0.1145 - val_loss: 0.2452 - val_mse: 0.2452\n",
      "Epoch 142/200\n",
      "9/9 [==============================] - 0s 3ms/sample - loss: 0.1141 - mse: 0.1141 - val_loss: 0.2452 - val_mse: 0.2452\n",
      "Epoch 143/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1137 - mse: 0.1137 - val_loss: 0.2453 - val_mse: 0.2453\n",
      "Epoch 144/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1133 - mse: 0.1133 - val_loss: 0.2454 - val_mse: 0.2454\n",
      "Epoch 145/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1129 - mse: 0.1129 - val_loss: 0.2454 - val_mse: 0.2454\n",
      "Epoch 146/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1125 - mse: 0.1125 - val_loss: 0.2452 - val_mse: 0.2452\n",
      "Epoch 147/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1121 - mse: 0.1121 - val_loss: 0.2452 - val_mse: 0.2452\n",
      "Epoch 148/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1116 - mse: 0.1116 - val_loss: 0.2451 - val_mse: 0.2451\n",
      "Epoch 149/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1113 - mse: 0.1113 - val_loss: 0.2449 - val_mse: 0.2449\n",
      "Epoch 150/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1108 - mse: 0.1108 - val_loss: 0.2447 - val_mse: 0.2447\n",
      "Epoch 151/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1104 - mse: 0.1104 - val_loss: 0.2447 - val_mse: 0.2447\n",
      "Epoch 152/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1100 - mse: 0.1100 - val_loss: 0.2448 - val_mse: 0.2448\n",
      "Epoch 153/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1096 - mse: 0.1096 - val_loss: 0.2448 - val_mse: 0.2448\n",
      "Epoch 154/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1091 - mse: 0.1091 - val_loss: 0.2445 - val_mse: 0.2445\n",
      "Epoch 155/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1088 - mse: 0.1088 - val_loss: 0.2442 - val_mse: 0.2442\n",
      "Epoch 156/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1084 - mse: 0.1084 - val_loss: 0.2441 - val_mse: 0.2441\n",
      "Epoch 157/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1080 - mse: 0.1080 - val_loss: 0.2441 - val_mse: 0.2441\n",
      "Epoch 158/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1075 - mse: 0.1075 - val_loss: 0.2440 - val_mse: 0.2440\n",
      "Epoch 159/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1071 - mse: 0.1071 - val_loss: 0.2439 - val_mse: 0.2439\n",
      "Epoch 160/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1067 - mse: 0.1067 - val_loss: 0.2437 - val_mse: 0.2437\n",
      "Epoch 161/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1063 - mse: 0.1063 - val_loss: 0.2435 - val_mse: 0.2435\n",
      "Epoch 162/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1059 - mse: 0.1059 - val_loss: 0.2433 - val_mse: 0.2433\n",
      "Epoch 163/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1055 - mse: 0.1055 - val_loss: 0.2433 - val_mse: 0.2433\n",
      "Epoch 164/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1051 - mse: 0.1051 - val_loss: 0.2434 - val_mse: 0.2434\n",
      "Epoch 165/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1047 - mse: 0.1047 - val_loss: 0.2433 - val_mse: 0.2433\n",
      "Epoch 166/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1042 - mse: 0.1042 - val_loss: 0.2430 - val_mse: 0.2430\n",
      "Epoch 167/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1039 - mse: 0.1039 - val_loss: 0.2428 - val_mse: 0.2428\n",
      "Epoch 168/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1034 - mse: 0.1034 - val_loss: 0.2425 - val_mse: 0.2425\n",
      "Epoch 169/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1030 - mse: 0.1030 - val_loss: 0.2423 - val_mse: 0.2423\n",
      "Epoch 170/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1026 - mse: 0.1026 - val_loss: 0.2423 - val_mse: 0.2423\n",
      "Epoch 171/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1021 - mse: 0.1021 - val_loss: 0.2424 - val_mse: 0.2424\n",
      "Epoch 172/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1017 - mse: 0.1017 - val_loss: 0.2421 - val_mse: 0.2421\n",
      "Epoch 173/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1013 - mse: 0.1013 - val_loss: 0.2417 - val_mse: 0.2417\n",
      "Epoch 174/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1009 - mse: 0.1009 - val_loss: 0.2414 - val_mse: 0.2414\n",
      "Epoch 175/200\n",
      "9/9 [==============================] - 0s 1ms/sample - loss: 0.1004 - mse: 0.1004 - val_loss: 0.2414 - val_mse: 0.2414\n",
      "Epoch 176/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.1000 - mse: 0.1000 - val_loss: 0.2414 - val_mse: 0.2414\n",
      "Epoch 177/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.0996 - mse: 0.0996 - val_loss: 0.2411 - val_mse: 0.2411\n",
      "Epoch 178/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.0992 - mse: 0.0992 - val_loss: 0.2407 - val_mse: 0.2407\n",
      "Epoch 179/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.0987 - mse: 0.0987 - val_loss: 0.2407 - val_mse: 0.2407\n",
      "Epoch 180/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.0983 - mse: 0.0983 - val_loss: 0.2406 - val_mse: 0.2406\n",
      "Epoch 181/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.0980 - mse: 0.0980 - val_loss: 0.2399 - val_mse: 0.2399\n",
      "Epoch 182/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.0977 - mse: 0.0977 - val_loss: 0.2393 - val_mse: 0.2393\n",
      "Epoch 183/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.0973 - mse: 0.0973 - val_loss: 0.2391 - val_mse: 0.2391\n",
      "Epoch 184/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.0969 - mse: 0.0969 - val_loss: 0.2391 - val_mse: 0.2391\n",
      "Epoch 185/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.0962 - mse: 0.0962 - val_loss: 0.2387 - val_mse: 0.2387\n",
      "Epoch 186/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.0956 - mse: 0.0956 - val_loss: 0.2377 - val_mse: 0.2377\n",
      "Epoch 187/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.0949 - mse: 0.0949 - val_loss: 0.2363 - val_mse: 0.2363\n",
      "Epoch 188/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.0942 - mse: 0.0942 - val_loss: 0.2352 - val_mse: 0.2352\n",
      "Epoch 189/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.0934 - mse: 0.0934 - val_loss: 0.2343 - val_mse: 0.2343\n",
      "Epoch 190/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.0927 - mse: 0.0927 - val_loss: 0.2331 - val_mse: 0.2331\n",
      "Epoch 191/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.0920 - mse: 0.0920 - val_loss: 0.2317 - val_mse: 0.2317\n",
      "Epoch 192/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.0913 - mse: 0.0913 - val_loss: 0.2308 - val_mse: 0.2308\n",
      "Epoch 193/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.0905 - mse: 0.0905 - val_loss: 0.2301 - val_mse: 0.2301\n",
      "Epoch 194/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.0897 - mse: 0.0897 - val_loss: 0.2289 - val_mse: 0.2289\n",
      "Epoch 195/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.0895 - mse: 0.0895 - val_loss: 0.2278 - val_mse: 0.2278\n",
      "Epoch 196/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.0891 - mse: 0.0891 - val_loss: 0.2276 - val_mse: 0.2276\n",
      "Epoch 197/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.0886 - mse: 0.0886 - val_loss: 0.2276 - val_mse: 0.2276\n",
      "Epoch 198/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.0882 - mse: 0.0882 - val_loss: 0.2274 - val_mse: 0.2274\n",
      "Epoch 199/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.0878 - mse: 0.0878 - val_loss: 0.2269 - val_mse: 0.2269\n",
      "Epoch 200/200\n",
      "9/9 [==============================] - 0s 2ms/sample - loss: 0.0875 - mse: 0.0875 - val_loss: 0.2266 - val_mse: 0.2266\n"
     ]
    }
   ],
   "source": [
    "# Fit the model\n",
    "model_3 = nn.fit(X_train, y_train, validation_data=(X_val, y_val), batch_size=1, shuffle=False, epochs=200)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 716,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAEICAYAAABPgw/pAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nO3deZwU5Z348c+3j+me+x4EBhhURA4BOTWiYhREY/DAVYlJNGo0GnOYzW5M8ouiMZvNsRvjamJMPLIJxiOr0aiJxCt4E1BBFBAkAzMcA8x9Tx/P74+qbnqG7pmeYaZ7uuf7fr36VVVPPVX1neqeb1U/Vf2UGGNQSimV+hzJDkAppdTg0ISulFJpQhO6UkqlCU3oSimVJjShK6VUmtCErpRSaWLEJ3QRqRSRsxK8TRGRB0WkXkTWJnjbfxGRKxK5zaEgIkZEjrXH7xWR7yU7plQQ7+ddRCrsfexKRFwDJSJXiMhfBrtuqhrWb1YaWwgsBsqNMa1DtRERWQkca4z5bKjMGHPOUG0vWYwxX0p2DNGIyI+BFUA+UA/cZ4z5QXKjSh4RuRz4lT3pBDxAW2i+MSanv+s0xvwW+O1g101VI/4MPUkmAJVDmczVsHA/cLwxJg/4BPAZEbkoyTEljTFmlTEmx07c5wB7QtPRkvlw/3YwHGlCjyAiHhG5U0T22K87RcRjzysRkWdEpEFE6kTkVRFx2PO+JSK7RaRZRLaKyJm9bONq4DfAySLSIiK3iciVIvJaj3qRTQoPicg9IvKsvY23ReSYiLrTRORvdlw1IvIdEVkKfAe41N7OBrvuKyJyjT3uEJH/JyI7RWS/iPyviOTb80Jfua8QkV0iclBEvhvHPlwpIo/Z62oWkQ9EZG7E/EoR+aaIbBSRRhF5VES8caz330Rkr/2+XNVj3kMicoc93tv71Ou2ReQ8EXnPXvYNEZkRse3/67HN/xGRO3uL2RiztcdBOwgc28ffGdrvXxCRKrtZ7ksiMs+Ou0FE7o6oH/M9tOd/zp5X2/P9s5e9WUQ+tuc/JiJFMeK6UkR22O/pP8U62x50IlJt7+/3sc/e7b8vtO0PRGRZRP1rROQVe9xl77vrRGS7ve/uGmBdp1j//7X2tr8iIsP/Z/XGmBH9AiqBs+zx24G3gDKgFHgD+L4974fAvYDbfp0KCDAZqALG2PUqgGP62OaVwGuxpu0yg9VcAvAQUAfMx2omWwU8Ys/LBfYC/wp47ekF9ryVwO97rPcV4Bp7/CpgO3A0kAM8Afwu4u8wwK+BTGAm0AlM6eNvWwl0AOdifa3+IfBWj/29FhgDFAGbgS/1sc6lQA0wHcgGHo6yf+7o7X3qa9vAbGA/sMCO+wq7vgcYDbQCBXZdl113Thyfr5uBFjveHVjNbL3VD+33e+33c4m9P/+E9bkca2/79Djew6n2tk+z/47/Bvwc+rx/HevzXm7P/xXwhx5xuOx93gRMtueNBqYd4f/dIqA6Snk1sN6OKdMuu8TepgP4jP03jbLnXQO8EvG+GOAprGauCqz/m7MGUPdGYJO9v4uAlwGT7HzV10vP0Lu7HLjdGLPfGHMAuA34nD3Ph/WhmmCM8RljXjXWOx/A+meYKiJuY0ylMebjIYjtCWPMWmOMHyuhz7LLzwP2GWP+yxjTYYxpNsa8Hec6Lwf+2xizwxjTAnwbuEy6f9W9zRjTbozZAGzASux9ec0Y85wxJgD8Lsoydxlj9hhj6oA/R/wtsVwCPGiM2WSsM96VvdSN9T71te0vAr8yxrxtjAkYq721EzjJGLMXWAP8i113KXDQGLO+j7gxxvwn1kF2Nta+aOxrGdv37fdzNdbB5A/253I38Cpwol2vt/fwYuAZY8waY0wn8D2sbwkh1wHfNcZU2/NXAhdL9KaOIDBdRDKNMXuNMR/E+XcMxM/tmNoBjDGP2dsMGmMexjrQzu1l+R8aYxqNMZVYJzC9fb5i1b0E+JkxZrf9WfnRkfxBiaIJvbsxwM6I6Z12GcBPsM6EVttfwW4GMMZsxzrTWQnsF5FHRGQMg29fxHgb1tkYwDhgoAeQaH+vCxgVx3Z703MZb48k0d91jsH6FhQZZyxR36c4tj0B+Fe7SaNBRBqw9m3ovfwtELq4/Fms5BwXY3kXaMc6SYhHTcR4e5TpUNy9vYfd9pt9MKyNqDsBeDLi792MdYIS+f6HlrsU+BKwV6ymv+OjBS1W817oNT7Ov7WnyPc61NyzISLO44GSXpbvz+crVt2en7luMQ1XmtC724P1IQ8Zb5dhn/n+qzHmaODTwDfEbis3xjxsjFloL2vo/9G8FcgKTYjIUf1Ytgo4Jsa8vtr8ov29fronj+FgL1ZyDYmZKHp7n/pQBfzAGFMQ8coyxvzBnv8nYIaITMf6VrRqAH+Hi9jv1UD19h52228ikgUUR9StAs7p8Td77W8B3RhjnjfGLMb69rMFqynuMCbiIqcxZtcA/6bw51ZEjgZ+CVwPFBtjCuztywDXHa+9WM0+IeNiVRxONKF39wfg/4lIqYiUALcAv4fwBbNjRUSw2hMDQEBEJovIJ8W6eNqBdfYU6Od2NwDTRGSWfZFuZT+WfQY4SkS+LtZF3VwRWWDPqwEqQhcFY/y9N4nIRBHJAf4DeNRu1hlOHgOuFJGpdlK6NVbFWO9THNv4NfAlEVkglmwR+ZSI5AIYYzqAP2K136/tK1nZFxyvE5FCe33zgS8DL8bzB/dDb+/hH4HzRGShiGRgXSOK/CzcC/xARCbYMZeKyPlR/pZRIrJMRLKxmqFa6P9nfKBysBL8ASsUuQbrDH2oPQZ8XUTGiEgh8G8J2OYR04Te3R3AOmAj8D7wjl0GMAl4AevD/CbwC2PMK1jt5/8JHMT6+laGdXdJ3IwxH2H9s70AbANe632Jbss2Y93T/ml7+9uAM+zZj9vDWhF5J8riD2A1HawB/ol1QPpKf2JPBGPMX4A7gZewmlNe6qV6rPepr22sw2pHvxvrnvHtWBerI/0WOIH4m1suxGoOa8Y6Mfgf+zWYYr6Hdjv3l7EOQnux/q7qiGV/DjyN1TzVjHWBdAGHc2BddN+DdeHwdOCGQf47ojLGbATuwrqYvRcrmcd7jehI/BKrTf19rIu0zwJdCdjuEQld/VdK9cFuE94CHGWMaUp2PCpxROTTwJ3GmMFuMhtUeoauVBzsZqtvYN0uqsk8zdlNbkvt+9HLsZpfn0x2XH3RhD5ExOozpSXKq1/NMcPRUPxtYv0YKto6k973ht123ITVtHVrj3nRYm4RkVN7Wd/lMZYZylsBVf8I8AOs20zXYzXDxnuHUtJok4tSSqUJPUNXSqk0kbTOb0pKSkxFRUWyNq+UUilp/fr1B40xpdHmJS2hV1RUsG7dumRtXimlUpKIxPyltDa5KKVUmtCErpRSaUITulJKpQl9IohSacrn81FdXU1HR0eyQ1ED4PV6KS8vx+12x72MJnSl0lR1dTW5ublUVFRg9VWmUoUxhtraWqqrq5k4cWLcy2mTi1JpqqOjg+LiYk3mKUhEKC4u7ve3K03oSqUxTeapayDvXeol9JoP4cXvQ2tt33WVUmoESb2EXrsdXv0pNO9NdiRKqV5cddVVlJWVMX369Jh1rrzySv74xz8eVr5nzx4uvvjiqMssWrQo6o8SH3roIW688caBBxyhoqKCgwcPDsq6Ein1EnqG/aQ2X1ty41BK9erKK6/kr3/964CWHTNmTNREny4CgaF54FOfCV1EHhCR/SKyqY9680QkICLRD6uDJcN+hmtXy5BuRil1ZE477TSKior6rLdmzRo+8YlPcPTRR4eTeGVlZfjMvr29ncsuu4wZM2Zw6aWX0t7eHl72wQcf5LjjjuP000/n9ddfD5cfOHCA5cuXM2/ePObNmxeet3LlSq666ioWLVrE0UcfzV133dVnfBdccAFz5sxh2rRp3HfffQDcf//93HTTTeE6v/71r/nGN74BwO9//3vmz5/PrFmzuO6668LJOycnh1tuuYUFCxbw5ptv9rndgYjntsWHsB7L9b+xKoiIE+vByM8PTli9yMi2hl2tQ74ppdLFbX/+gA/3DO5zOaaOyePWT0874vXs3buX1157jS1btrBs2bLDmlp++ctfkpWVxcaNG9m4cSOzZ88OL3frrbeyfv168vPzOeOMMzjxxBMB+NrXvsZNN93EwoUL2bVrF2effTabN28GYMuWLbz88ss0NzczefJkrr/++l7v9X7ggQcoKiqivb2defPmsXz58vAB5sc//jFut5sHH3yQX/3qV2zevJlHH32U119/HbfbzQ033MCqVav4/Oc/T2trK9OnT+f2228/4n0WS58J3RizRkQq+qj2FeD/gHmDEFPv3HaTS5c2uSiVDi644AIcDgdTp06lpqbmsPlr1qzhq1/9KgAzZsxgxowZALz99tssWrSI0lKr48FLL72Ujz76CIAXXniBDz/8MLyOpqYmmpubAfjUpz6Fx+PB4/FQVlZGTU0N5eXlMeO76667ePJJ62FFVVVVbNu2jZNOOolPfvKTPPPMM0yZMgWfz8cJJ5zA3Xffzfr165k3z0qF7e3tlJWVAeB0Olm+fPkR7au+HPEPi0RkLNbDcD9JHwldRK4FrgUYP378wDaoTS5K9dtgnEkPFY/HEx6P9cCdWLfwxSoPBoO8+eabZGZm9ro9p9OJ3++PGdsrr7zCCy+8wJtvvklWVhaLFi0K3xt+zTXX8B//8R8cf/zxfOELXwjHf8UVV/DDH/7wsHV5vV6cTmfMbQ2GwbgoeifwLWNMn638xpj7jDFzjTFzQ0fVftMmF6VGlNNOO41Vq1YBsGnTJjZu3AjAggULeOWVV6itrcXn8/H444+Hl1myZAl33313ePq9994b0LYbGxspLCwkKyuLLVu28NZbb4XnLViwgKqqKh5++GFWrFgBwJlnnskf//hH9u/fD0BdXR07d8bs7XbQDUZCnws8IiKVwMXAL0TkgkFYb3RuvctFqVSwYsUKTj75ZLZu3Up5eTn333//gNZz/fXX09LSEm6znj9/PgCjR49m5cqVnHzyyZx11lnhtnWwmknWrVvHjBkzmDp1Kvfee++Atr106VL8fj8zZszge9/7HieddFK3+ZdccgmnnHIKhYWFAEydOpU77riDJUuWMGPGDBYvXszevYm7xTquZ4rabejPGGNi31Bq1XvIrtfn/UZz5841A37AxQ9Gw7yrYckdA1teqRFg8+bNTJkyJdlhpLXzzjuPm266iTPPPHNI1h/tPRSR9caYudHqx3Pb4h+AN4HJIlItIleLyJdE5EuDEvFAZGRrk4tSKmkaGho47rjjyMzMHLJkPhDx3OWyIt6VGWOuPKJo4uXO0rtclFJJU1BQEL6jZjhJvV+KgnWni97lopRS3aRoQtcmF6WU6ilFE3qW3uWilFI9pGhCz9EzdKWU6iFFE3q2tqErNYxVVVVxxhlnMGXKFKZNm8bPf/7zqPW0+9zBlZrPFNW7XJQa1lwuF//1X//F7NmzaW5uZs6cOSxevJipU6fGtfxI6D53KLoBSOEzdG1yUWq4Gj16dPiXm7m5uUyZMoXdu3dHrTtSu8+9+eabmTp1KjNmzOCb3/xmfDu2D6l5hp6RDb5WCAbBkZrHJKUS6i83w773B3edR50A5/xnn9UqKyt59913WbBgQdT5I7H73Lq6Oq6++mq2bNmCiNDQ0BDXLu9L6iZ0AH/7oXGl1LDT0tLC8uXLufPOO8nLy4taZyR2n5uXl4fX6+Waa67hU5/6FOedd16ce7R3qZ3Qu1o1oSsVjzjOpAebz+dj+fLlXH755Vx00UUx643E7nNdLhdr167lxRdf5JFHHuHuu+/mpZdeihlHvFKzvcIdSuh6p4tSw5ExhquvvpopU6aE25YHKh27z21paaGxsZFzzz2XO++8c8Dx9ZSaCT18hq53uig1HL3++uv87ne/46WXXmLWrFnMmjWL5557bkDrSsfuc5ubmznvvPOYMWMGp59+Oj/72c8GFF9PcXWfOxSOqPvc7S/C7y+Cq1bD+OgXWpQa6bT73KGXct3nDksZ2uSilEqelO0+d1gKJXTtz0UplQTafe5g0ueKKqXUYVIzoetdLkopdZjUTOh6l4tSSh0mNRO6O8saapOLUkqFpWZCdzjsHhe1yUWp4aijo4P58+czc+ZMpk2bxq233hq1XqyucNetWxf+uX9Psbq2XblyJT/96U+PLHBbTk7OoKwn0fq8y0VEHgDOA/YbY6ZHmX858C17sgW43hizYVCjjCYjW+9yUWqY8ng8vPTSS+Tk5ODz+Vi4cCHnnHPOYT/MiWXu3LnMnRv1VuuUZ4zBGINjCDoWjGeNDwFLe5n/T+B0Y8wM4PvAfYMQV9+0C12lhi0RCZ/l+nw+fD5fzH5XHn/8cebPn89xxx3Hq6++Clh9qIQ6rKqtrWXJkiWceOKJXHfddd36e/nBD37A5MmTOeuss9i6dWu4/OOPP2bp0qXMmTOHU089lS1btgDWAzW++tWvHtZdbywtLS2ceeaZzJ49mxNOOIGnnnoKgO9973vdHtrx3e9+N9wV709+8hPmzZvHjBkzwt9MKisrmTJlCjfccAOzZ8+mqqqKK6+8kunTp3PCCScM2i9F+zxDN8asEZGKXua/ETH5FhC727JBcLD9IB/WfshcdxZZmtCVisuP1v6ILXVbBnWdxxcdz7fmfyvm/EAgwJw5c9i+fTtf/vKXY3af6/f7Wbt2Lc899xy33XYbL7zwQrf5t912GwsXLuSWW27h2WefDfdJvn79eh555BHeffdd/H4/s2fPZs6cOQBce+213HvvvUyaNIm3336bG264Idz5VV/d9Ubyer08+eST5OXlcfDgQU466SSWLVvG1VdfzUUXXcTXvvY1gsEgjzzyCGvXrmX16tVs27aNtWvXYoxh2bJlrFmzhvHjx7N161YefPBBfvGLX7B+/Xp2797Npk2bAIZt97lXA38Z5HV2s75mPd/8+zd5wlPCpM7modyUUuoIOJ1O3nvvPRoaGrjwwgvZtGlT+KEVkUI9Mc6ZM4fKysrD5q9Zs4YnnngCsLq+DfWb8uqrr3LhhReSlWXdJLFs2TLAOqt+4403+Jd/+ZfwOjo7O8PjfXXXG8kYw3e+8x3WrFmDw+Fg9+7d1NTUUFFRQXFxMe+++y41NTWceOKJFBcXs3r1alavXh3ul72lpYVt27Yxfvx4JkyYEG5yOvroo9mxYwdf+cpX+NSnPsWSJUvi2qd9GbSELiJnYCX0hb3UuRa4FmD8+PED2k6+Jx+AxgwvdOhFUaXi0duZ9FArKChg0aJF/PWvf42a0EPd2fbWlW1/us8NBoMUFBTE7MEwnu56Q1atWsWBAwdYv349brebioqKbt3nPvTQQ+zbt4+rrroqvL5vf/vbXHfddd3WU1lZSXb2oa6+CwsL2bBhA88//zz33HMPjz32GA888ECvscRjUFrlRWQG8BvgfGNMbax6xpj7jDFzjTFzQ53S91eBpwCARlcG6Bm6UsPSgQMHws0I7e3tvPDCCxx//PEDWldk97l/+ctfqK+vD5c/+eSTtLe309zczJ///GfAenjExIkTw93pGmPYsGFg92k0NjZSVlaG2+3m5Zdf7tYV7oUXXshf//pX/vGPf3D22WcDcPbZZ/PAAw/Q0mKdbO7evTvclW6kgwcPEgwGWb58Od///vd55513BhRfT0d8hi4i44EngM8ZY4a8c4NQQm9wujShKzVM7d27lyuuuIJAIEAwGOSSSy4Z8FN5br31VlasWMHs2bM5/fTTw9/uZ8+ezaWXXsqsWbOYMGECp556aniZVatWcf3113PHHXfg8/m47LLLmDlzZr+3ffnll/PpT3+auXPnMmvWrG4HpYyMDM444wwKCgrCD65YsmQJmzdv5uSTTwas2x9///vfH/ZA6N27d/OFL3yBYDAIEPWBGAPRZ/e5IvIHYBFQAtQAtwJuAGPMvSLyG2A5EDp0+WN17RhpoN3ntvvbmb9qPl/PmcLVW9+A71T3ex1KjQTafe7QCgaDzJ49m8cff5xJkyYNyTb6231uPHe5rOhj/jXANf0J8khkujLxOD00ioGuZn1QtFIq4T788EPOO+88LrzwwiFL5gORkt3n5mfk0yD2N4uuFvBGf/isUkoNhalTp7Jjx45kh3GYlDy1zffm02jsq+H683+lYkrWE8nUkRvIe5eSCb3AU0CD8VkTemFUqai8Xi+1tbWa1FOQMYba2lq8Xm+/lkvJJpcCTwE7GndZE516hq5UNOXl5VRXV3PgwIFkh6IGwOv1Ul7evx/ep2RCz/fk0+BvtyY6m5IbjFLDlNvtZuLEickOQyVQSja55Gfk0+hvxYA2uSillC0lE3qBpwC/CdAqohdFlVLKlpIJPdyfi9OhZ+hKKWVLyYQe/vm/w6Ft6EopZUvJhB4+Q3d79C4XpZSypWRCD5+he7K0yUUppWwpmdBDZ+gNbq8mdKWUsqV0Qm90efQuF6WUsqXkD4tcDhe57lwag0bP0JVSypaSZ+gAeZ48Gpx6l4tSSoWkbEIv9BTS4EDvclFKKVvqJnRvIXVok4tSSoWkdEKvN369KKqUUraUTehF3iLqTRfG1wYBf7LDUUqppEvZhF7oLaTTBGgXsZ4tqpRSI1zqJnRPIQB1TodeGFVKKeJI6CLygIjsF5FNMeaLiNwlIttFZKOIzB78MA9X5C0CoN7h1AujSilFfGfoDwFLe5l/DjDJfl0L/PLIw+pbodc6Q6/Xe9GVUgqII6EbY9YAdb1UOR/4X2N5CygQkdGDFWAsoYRe53RChyZ0pZQajDb0sUBVxHS1XXYYEblWRNaJyLojfXBtuMnF6YCOxiNal1JKpYPBSOgSpcxEq2iMuc8YM9cYM7e0tPSINprlysLjyLDa0DsajmhdSimVDgYjoVcD4yKmy4E9g7DeXokIhd5CavUMXSmlgMFJ6E8Dn7fvdjkJaDTG7B2E9fap0FtEvcutCV0ppYij+1wR+QOwCCgRkWrgVsANYIy5F3gOOBfYDrQBXxiqYHsq0oSulFJhfSZ0Y8yKPuYb4MuDFlE/FHoLqXQ6NaErpRQp/EtRsHtcFDShK6UUKZ7Qi7xFtIuhQ+9yUUqp1E7oof5c6jv1DF0ppVI7oYd+LerTvlyUUiqlE3pJZgkAtf52MFF/y6SUUiNGeiR0hwFfe5KjUUqp5ErphF6cWQzAQb11USmlUjuhe5wecp1eO6HrnS5KqZEtpRM6QElGHge1PxellEr9hF7sKeSgS5tclFIq5RN6SWYptdqGrpRSaZDQs8v0oqhSSpEGCb04ZwytDgdtbUf2BCSllEp1KZ/QS7JHAVCrCV0pNcKlfkIP/bioozbJkSilVHKlTUI/2Kn3oSulRrb0SehdTUmORCmlkivlE3qhpxAHcNDXkuxQlFIqqVI+oTsdTgrFzcFAR7JDUUqppEr5hA5Q4syk1nQlOwyllEqquBK6iCwVka0isl1Ebo4yf7yIvCwi74rIRhE5d/BDja3Enct+B+DTs3Sl1MjVZ0IXESdwD3AOMBVYISJTe1T7f8BjxpgTgcuAXwx2oL05ylPIfqcL2usSuVmllBpW4jlDnw9sN8bsMMZ0AY8A5/eoY4A8ezwf2DN4IfatLKuUWqcDX0tNIjerlFLDSjwJfSxQFTFdbZdFWgl8VkSqgeeAr0RbkYhcKyLrRGTdgQOD98vOUdljMCIcaKgctHUqpVSqiSehS5Syng/wXAE8ZIwpB84Fficih63bGHOfMWauMWZuaWlp/6ONYVTeeAD2N+0ctHUqpVSqiSehVwPjIqbLObxJ5WrgMQBjzJuAFygZjADjMapgIgD7WhLa0qOUUsNKPAn9H8AkEZkoIhlYFz2f7lFnF3AmgIhMwUroCestq6xwEgA12kGXUmoE6zOhG2P8wI3A88BmrLtZPhCR20VkmV3tX4EvisgG4A/AlcaYns0yQyYvq5jMoGG/dtCllBrBXPFUMsY8h3WxM7LslojxD4FTBje0+IkIo4xQ49P+XJRSI1da/FIUoExc1Phbkx2GUkolTdok9FGOTPbrz/+VUiNY+iR0dy77CRI0wWSHopRSSZE+Cd1TiF+grkN//q+UGpnSJqGX2Q+6qGneneRIlFIqOdImoR+VPRqAfXXbkxyJUkolR9ok9LH51s//dzfuSHIkSimVHGmT0PNyy8kJBtndvCvZoSilVFKkTUKX7FLG+Pzsadmb7FCUUiop0iahkzOKsX4/1R0Hkx2JUkolRfok9KwixgaC7O5qJIHdyCil1LCRPgnd4WSsw0u78dPQ2ZDsaJRSKuHSJ6EDY93WU/B2t+i96EqpkSetEvqYTOspSJrQlVIjUVol9LE5YwBN6EqpkSmtEnpO7hjyA0F2N1cnOxSllEq4tEro1q2LPnbrw6KVUiNQeiX07FLKfX6qm6qSHYlSSiVceiX0nFFM8PupbqvBF/AlOxqllEqotEvoFT4/AYJUtehZulJqZEmzhF5Ghc86M9/ZqO3oSqmRJa6ELiJLRWSriGwXkZtj1LlERD4UkQ9E5OHBDTNO3nwqAtafVNlUmZQQlFIqWVx9VRARJ3APsBioBv4hIk8bYz6MqDMJ+DZwijGmXkTKhirgPoIlL7uMInFrQldKjTjxnKHPB7YbY3YYY7qAR4Dze9T5InCPMaYewBizf3DD7IecUiqMk8rGyqSFoJRSyRBPQh8LRF5hrLbLIh0HHCcir4vIWyKyNNqKRORaEVknIusOHDgwsIj7kjOKiT6/nqErpUaceBK6RCnr2T+tC5gELAJWAL8RkYLDFjLmPmPMXGPM3NLS0v7GGp+8sUxob6auo47Gzsah2YZSSg1D8ST0amBcxHQ5sCdKnaeMMT5jzD+BrVgJPvHyy6lobwH0wqhSamSJJ6H/A5gkIhNFJAO4DHi6R50/AWcAiEgJVhNMcp7WnF/OxC7r1sUdDfrAaKXUyNFnQjfG+IEbgeeBzcBjxpgPROR2EVlmV3seqBWRD4GXgX8zxtQOVdC9yh/HOL8fr8PNR/UfJSUEpZRKhj5vWwQwxjwHPNej7JaIcQN8w34lV/5YnMAkT7EmdKXUiJJevxQFyDkKxCCu2gAAABumSURBVMlxjky21m/V54sqpUaM9EvoThfkjeE4PzR2NrK/LXm3xCulVCKlX0IHyC9ncnsrAFvrtyY5GKWUSoy0TejHNVpn5tqOrpQaKdIzoeeNJbdpL2Ozx7C1Ts/QlVIjQ3om9PxyCPo4Lq+CLXVbkh2NUkolRJomdOuHrdMzj6KyqZKmrqYkB6SUUkMvTRN6OQAznLkAvH/g/WRGo5RSCZGeCb2wAoATunw4xMGGAxuSG49SSiVAyiX0j2qaueWpTXT6A7EreXIgdzTZ9bs4tuBYTehKqREh5RL6noZ2/vfNnby8pY/+1IuPhbqPmVk6k/cPvE/QBBMToFJKJUnKJfSFx5ZQmuvhiXeqe69YdDTUbmdm6Uyafc3a86JSKu2lXEJ3OR1cMGsML2/dT11rV+yKxcdCWy0zcysAePfAu4kJUCmlkiTlEjrARbPL8QUMz2zs+ZyNCMXHAjChq4uyzDLe3vt2gqJTSqnkSMmEPmV0HlNG5/HoP6pi96ZoJ3Sp28FJY07irb1vEQj2ciFVKaVSXEomdIDPLBjPB3ua2FAd47mhhRUgDqjdzsljTqaxs1F/NaqUSmspm9AvmDWGrAwnq97aGb2CKwMKxkPtdk4afRIAb+59M4ERKqVUYqVsQs/1urngxLH8eeMeGtt80SsVHwu12ynJLGFy4WTe2PNGYoNUSqkEStmEDvC5kybQ4Qvy8Npd0SuUTIaDH0HAzyfGfoJ3979LY2eMJhqllEpxKZ3Qp4zO49RJJTzw+j+j/3J09Azwd0DtNhaPX4w/6OeVqlcSHqdSSiVCXAldRJaKyFYR2S4iN/dS72IRMSIyd/BC7N11px3DgeZOnno3yi2MR82whvveZ3rJdEZnj2b1ztWJCk0ppRKqz4QuIk7gHuAcYCqwQkSmRqmXC3wVSOgN36ccW8y0MXn88u8f4w/0+Hl/ySRwemDvBkSExRMW88aeN2juak5kiEoplRDxnKHPB7YbY3YYY7qAR4Dzo9T7PvBjoGMQ4+uTiPC1Myfxz4OtPPHO7u4znW4YNRX2bQRgScUS/EE/L1e9nMgQlVIqIeJJ6GOBqojparssTEROBMYZY57pbUUicq2IrBORdQcO9NG5Vj8snjqKmeMK+PmL2w5vSz9qBux7H4zhhJITGJc7jie2PTFo21ZKqeEinoQuUcrCP88UEQfwM+Bf+1qRMeY+Y8xcY8zc0tLS+KPsK0AR/m3JZHY3tPPg65XdZ46eAe310FiNQxxcfNzFrK9Zr511KaXSTjwJvRoYFzFdDkRegcwFpgOviEglcBLwdCIvjAIsnFTCWVPK+J8Xt1HTFNHqE74wajW7nH/M+bgcLh7/6PFEhqeUUkMunoT+D2CSiEwUkQzgMuDp0ExjTKMxpsQYU2GMqQDeApYZY9YNScS9+N55U/EFDXc8u/lQ4ajp4HBD1VoAijOLWTx+MU99/BStvtZEh6iUUkOmz4RujPEDNwLPA5uBx4wxH4jI7SKybKgD7I8Jxdl8edGx/HnDHv66aZ9VmJEFY2fDztfD9T4/7fM0dzXzyJZHkhSpUkoNvrjuQzfGPGeMOc4Yc4wx5gd22S3GmKej1F2UjLPzkBvOOIZpY/L47pPvc7Cl0yqsWAi734HOFgCml0xn4diF/PaD39Lma0tWqEopNahS+pei0bidDv77klk0d/q56dH3CAQNTDgFTACqDt0if92M66jvrOfhLQ8nMVqllBo8aZfQASYflcv3z5/Gq9sO8rO/fQTjFoA4ofK1cJ1ZZbNYNG4R9228j32t+5IYrVJKDY60TOgAl84bz6Vzx3H3y9t5/P16qx09IqEDfGvetwiaID9d99MkRamUUoMnbRM6wPcvmM7CY0u4+Yn3+WfuHNi9HtrqwvPLc8u5+oSreb7yeVZXah8vSqnUltYJPcPl4Jefnc20MXnctHGc1Y6+5dluda454RqmF09n5Zsr2duyN0mRKqXUkUvrhA7WgzB+d/UCGD2LKlPK7jcf7Tbf7XDz49N+TNAE+drLX9O7XpRSKSvtEzpAfqabVV88iU35iyjd/wZ3/PENOnyH+nwZlzeOH5/2Y7bWb+Xf1/w7vmCMJyAppdQwNiISOkC2x8Xi5V8kQwLUv/MUn/6f1/hwT1N4/mnlp/Ht+d/m79V/51trvoUvoEldKZVaRkxCB3CNmwfFx3Lr6DdpaPdxwT2v8/MXttHeZZ2tX3b8Zfz7vH/nbzv/xo0v3aj9piulUsqISug4HLDgS+TVbuCFf8lk8dRR/OyFjzjjp6/wxDvVBIOGz039HLd/4nbW7l3LZ579DFvqtiQ7aqWUisvISugAM1eAJ5/8jfdzz+Wzeey6kynL8/CNxzaw5M41/GHtLs6pWMZ9S+6jxdfCimdWcO+Ge7VdXSk17I28hO7JgTmfhw/+BDUfMn9iEX+64RTuWnEiHpeDbz/xPqf850u8+G4uPz35dyyuWMw9793D5577HO/tfy/Z0SulVExijOm71hCYO3euWbcuSX14tdXBXSdaD7/4/NMg1jM8jDG8taOO+1/7Jy9v3U8gaJg2Jo8px1TydtNvaOiq46zxZ/HV2V9lYv7E5MSulBrRRGS9MSbq8yZGZkIHWPtreO6bsPx+OOHiw2YfbOnkmQ17ePLd3WyobgTppLT8bXw5L2Ho4vTyxVw/62qmFE9JQvBKqZFKE3o0wQDcvwQObIUvvgilk2NW3dfYwUtb9vPi5hre3rkTX+5LuAvWIs5OihzTOHXUMj496UxOGFtMVoYrgX+EUmqk0YQeS2M13LcIvPlw5bOQe1Sfi/gDQTbvbebVj6t4bucT7PStBlcjQX82/qYTOUpO44SyyRxblsMxpdZrYkk2mRnOof97lFJpTxN6b3a9Bb+7CLJL4LNPQMmx/VrcF/Dx3Pa/8+jW/2NT/ZsYAjh942mrn46vaTrGVwTA2IJMji7NZlxRFuWFmZQXWsNxhVmU5GQgEu1Z3Eop1Z0m9L5Ur4dVF4O/E5bcDnO+AI7+n1HXttfyzI5neHbHs2yus55rWp41mTHuBTjbZ7KvNovq+nbqWru6LedxOSgvzGRMQSaj8ryMyvPYQ294ujTHg8s58m5KUkp1pwk9Ho3V8NSNsONlKJ4En/gKTLsQvHkDWl1VUxWrd65m9c7VfFj7IQBTiqZw+rjTmVd2CrkykT0NHVTXt1NV10ZVfRv7mjqpaezgQEun9aSlCCJQkuOxkn2ul5IcD0U5GRRnZ1Cck0Fxtoei7AyrPDuDDJcmf6XSkSb0eBkDm/8Mf/8R1GwClxfGnwQTT7deo6aB29vv1VY3V/O3nX/jxV0vsvHARgyGIm8RC8cu5NTyU/nEmE+Ql3HowBEIGmpbO9nf1ElNUwf7mjqoaepkf1MHNfZ4bWsntS1d+IPR379cj4vinAyKsjMozvFQnG2NF2ZlkJ/pJj/LTUF4mEFBlhuvW9v5lRruNKH3lzFQvQ4+eAJ2vAL7rTNsxAnFx0DZVCg6GgonQGEFFEyA/HJwuvtcdX1HPa/tfo1Xd7/K67tfp6mrCac4mVYyjQVHLWDB6AXMLJ2J19X3gcMYQ1OHn9qWTupauzjY0kVdaxe1LZ3UtnZR29pFnZ34rfGuw878I3lcDvIz3RTYST4vPG4N8zPd5GW6yfO6yfG6yPW6yPW6yfG4yPG4cDr0OoBSQ+2IE7qILAV+DjiB3xhj/rPH/G8A1wB+4ABwlTFmZ2/rHNYJvaeW/bDzdaj5AGo+tBJ8YxUE/YfqiBPyx0L+OMgdDXljIG9s92FOWbe2eX/Qz8YDG3lt92u8ve9tPjj4AQETIMORwayyWSwYvYD5R81nWsk03I6+DxZ9McbQ0umnoc1HY7v1Co03tHfR2HaorKG9i4Y2H03tPhrafbR1Bfpcf3aGk1yvm1yvy074bnI9ocTvIsdzaF6ePZ3jdZHjcZKV4SI7w0WWx4lbrxUoFdMRJXQRcQIfAYuBauAfwApjzIcRdc4A3jbGtInI9cAiY8ylva03pRJ6NAE/NO+B+kqo32kPK6Fpj1XetAcC3S9+Ik7r1si8Md0Tfe5oyD2KFk8u69t38/bBTazdt5at9VsByHRlMr1kOrNKZzGzdCYzS2dS4C1I6J/b5Q/aB4Eumjv8tHT6ae7w09zhs4d+uzxiutOa32JPt/v6PigAZDgdZHmcVoLPcJLlcZGdYSf9cPKPKI823+Mky+3Cm+Eg0+0k0+3Ui8oqLRxpQj8ZWGmMOdue/jaAMeaHMeqfCNxtjDmlt/WmfELvizHQVgtNu63k3u0VUeZrPXxZVybklFGXU8raTC/vumBDsJUtXXUEsN6vipxyZo2azcyyE5lWPI1jCo4hw5mR4D+yf3yBIK32gaDJTvQtnX7augK0dvpp7QrQFhp2+WnttIcxyuP51hDJ7RS8dnLPzLCGUacjDgJeuzxUx+M6VDfTrhtah9ftxONy6IFDDaneEno8P2scC1RFTFcDC3qpfzXwlxiBXAtcCzB+/Pg4Np3CRKx727NLYPTM6HWMgc4mK7G31FhNOy011qu5hqKWGpY27GdpSw2019EmwgeeDDZ4PGxobeOVxp386eOnAXAZOMbhZbK7gCmZRzE5dzzHF04iN3csZBUfemVkh/uuSTS300FBVgYFWYNz4AkGDe2+AK1dfto67aF9cAgNO/xBOroCtPvsV1eADnvcGlrzDzR3Hja/3RdgIJeYnA7B43LgcTnCSd7jcuJxRylzOexy56Fhz2XdUcpirM/tFP1NwwgWT0KP9umI+jEXkc8Cc4HTo803xtwH3AfWGXqcMaYvEetXqt58KOujTxh/J1mtB5jXUsO8lv3QvA/TXMOupn+yuXUPW7rq2OJr5bXAXp7u2geN70E1jPX5mdLVxaQuH8d0dXFMUJiQUYg7q8hK8JmF4C2wYsi0h96IYbgsP66LvonkcAjZHhfZHhfkDv76jTF0+oMRB4Ag7V0RB4OIA0WHL0CnL0in36rX6Q/Q6Q9GLWvrClDf1nWozBe06tr1joSIdeD0OB1kuKyXOzTecxgxHqrjCS8jZDidEfUkYtw6cISmPXFsw+nQA00ixJPQq4FxEdPlwJ6elUTkLOC7wOnGmM7BCU+FuTzWnTT55eEiASbYr6URVQ+07mdLzTts2f8em+u2sKWpkhc7a8NHYScwnnaOCezm6PpKjuno4JjWRiZ0deDt7ZTUnd09wXvzISPH6pI4Iwc8ufFNuzOT9i2hP0SsJhqv20mirlgYY+gKBGMeDKKW+YN0+gLhYVfA0OUP0hUIWEN/EF/AOjh1BYJ0+a2DUGO7z64X7D70B/EFgjFviR0IEcJJ3uNy4HI4cLsEt8M6GLicgttpHUisaesgYtVz4HZIt3oZLgcuR4xlnI7u5Q6x12GVWfWibVPscrvM4cCRYnduxdOG7sK6KHomsBvrouhnjDEfRNQ5EfgjsNQYsy2eDad9G/ow0+5vZ2fTTj5u+JiPGz5mR+MOPm74mKrmKgLmUFv0qMwSxmeOYpynkHHuPMY7MhmPm3FByO5qg44GaG+Ajkbr1dUCnS3W0NcWXzDigIxcq/nHYyd4d9ahoct7eJk7M0ZZlvXbAHcWODOsZV0Z4PRYB8EUOHAMV4GgwWcfXHwRyT7WASBUdlh9e7ozoswfsNbtCxp8/iD+YJCugMEfsOr67PmH6gXx+Y1Vzz5A+YPWcCg5HYLLIT0OANY3Dped9F2hg4hDcEYceFwRB5DQvNDB4tRJpSyeOmpAMR1RG7oxxi8iNwLPY53cPWCM+UBEbgfWGWOeBn4C5ACP21+rdhljlg0oWjUkMl2ZHF90PMcXHd+tvCvQZSX6xo/Z2biTXc27qGqu4u/1m6ntqO1Wt8hbxPjc8YwvrmBMzhjGZI/hqOyjGJNjDT3i6p7gO1ugqzliujnK/Bbwd1gHg85m6zqCrw187YeGPe8W6g9nKLnbyd6ZYSV6l+dQ0o92IAiVOzOspiaHG5wue+gGhyuivOe0y1rusGV6WYfDZd3SOowOQE6H4HQ4h/UPzowx+O0Dz2EHgYiy0DcOn//wg0hcB5Qe6wwErfX4Q3VC40FDW5ffjulQmT8YWpc1XpSdMeCE3hv9YZGKqdXXSlVzFbuarCRf1VzFruZd7Graxf62/Zgel1KKvcWMzh7N6JzRjM4eHU70ZZlllGaVUpxZPLD76QN+8LdHJPmOiKQfmfg7rf54/J09xrtilHeCv6v35QJd3X9vMNTEYd3e6nAeGkaOi9NO/hH1HC573BEx33moLDy/t3WGxh0xXtLLPAcgfdeJe/4A1xOKgYj1IPZVwJ5l0scy0n082vyY66SP+WIfzAfW1faR3uWiRqhsd3bUs3qwepmsaathb+te9rbuZU/LHva17mNPyx621W9jTfUaOgPdL6UIQqG3kLKsMkoySyjLKqM0s9R6ZR0aFnuLcUdegHW6wJlrtcEngzFWUg/4IOizDjBBX/fpQFfseeHpyHX0mA4GrJcJWOXBAJigXe63ywOHhpHj0cqCfmt5fxeY9j7WGTw0H2NNh189p3u81MCc8nVYfNugr1YTuhoQt9NNeW455bnlUecbY6jvrGdv614OtB1gf9t+DrYf7DbcUreFuo46glESQ25GLsXeYoq8RRRnWsNor3xPPnmevEH5JW1MIlbTyDC7y2dYOCzhxzoA9HFgCNU57IAS53rCB6PQOrAPOJFlJkpZsMd808f8WOukH8sEoXzekLwdmtDVkBCRcNKlOHY9f9BPXUcdB9oOcKDdSvx1HXXdXjsadrCuYx0NnQ2HNfOE5LhzrOSekUeBp4B8T/6hV4Y1DJXnefLIz7DqujVJHxkRq8mG4dvOPpJoQldJ5XK4KMsqoyyrrM+6/qCfhs4G6jrqqG2vpa6jjqauJho6G2jqbKKxs5GGzgYauxrZ27rXKu9qivoNICTDkUFORg457hyy3dnkZuR2G+a4c8LzQ+M9y7Pd2bgc+q+kkk8/hSpluBwuSjJLKMksgcL4lgmaIC2+Fho7G7u9GjobaPG10OJrobWrlWZfM62+Vlq6WtjdsptWXyvNXVZZ5G2dsWS6Msl0ZeJ1eq2hyxt+Zbmy8DoPTUfWDU17XV4ynZndp0Prc2eS4dCnWqm+aUJXac0hDvIy8sjLyGNc7ri+F+jBGENHoIOWLiv5h4aRCT9U3hHooN3fTru/nQ5/Bx2BDho7GqkJ1BxW3tu3hmgE6XYw8Dg9eJwevC6vNXR68bg84YNEb/M9Ts+h8R7LhMqG9JqEGjKa0JXqhYiEk2gppYOyTmMMvqDvsCQfOR2rPLKsw99BZ6CTNn8b9R31dAY66Qh00Om3h4HOfh84Qpzi7HZAiHZwCI2Hy/s6ePSo43Udqud2uPUbyCDQhK5UgokIGc4MMpwZ5Hvyh2w71o9u/OHkHjoAdASsA0Nk4g/N61kvsk7keFNbU7hu5Hy/Gdg9+6FvIH0dPGJ9o4h2IIm1vtB4Oh5ANKErlaZEBLfTjdvpJncoei+Lwh/0Rz0oRDtARB5E2v3tMQ8eHf4O6n31UdfbFRz4r4jDyT7i+kbPbxyx5oWaveKpk8iDhyZ0pdSgcTlcuBwust3ZCdle0AS7HQBifbvo7RtI6IASKm/3t9Pqaw1/+4gcxnOBvCdBDiV/O+lffNzFXDHtikHfH5rQlVIpyyGO8DWORPAFfd0Sf+ig0e5v79YEFXm9o1t9u05xZi8/zjgCmtCVUipObocbd0bimrD6S5+VpZRSaUITulJKpQlN6EoplSY0oSulVJrQhK6UUmlCE7pSSqUJTehKKZUmNKErpVSaSNpDokXkALBzgIuXAAcHMZzBNFxj07j6Z7jGBcM3No2rfwYa1wRjTNSuP5OW0I+EiKyL9dTrZBuusWlc/TNc44LhG5vG1T9DEZc2uSilVJrQhK6UUmkiVRP6fckOoBfDNTaNq3+Ga1wwfGPTuPpn0ONKyTZ0pZRSh0vVM3SllFI9aEJXSqk0kXIJXUSWishWEdkuIjcnMY5xIvKyiGwWkQ9E5Gt2+UoR2S0i79mvc5MQW6WIvG9vf51dViQifxORbfawMAlxTY7YL++JSJOIfD0Z+0xEHhCR/SKyKaIs6j4Sy132Z26jiMxOcFw/EZEt9rafFJECu7xCRNoj9tu9CY4r5vsmIt+299dWETl7qOLqJbZHI+KqFJH37PJE7rNYOWLoPmfGmJR5AU7gY+BoIAPYAExNUiyjgdn2eC7wETAVWAl8M8n7qRIo6VH2Y+Bme/xm4EfD4L3cB0xIxj4DTgNmA5v62kfAucBfAAFOAt5OcFxLAJc9/qOIuCoi6yVhf0V93+z/gw2AB5ho/886Exlbj/n/BdyShH0WK0cM2ecs1c7Q5wPbjTE7jDFdwCPA+ckIxBiz1xjzjj3eDGwGxiYjljidD/zWHv8tcEESYwE4E/jYGDPQXwsfEWPMGqCuR3GsfXQ+8L/G8hZQICKjExWXMWa1McZvT74FlA/FtvsbVy/OBx4xxnQaY/4JbMf63014bCIiwCXAH4Zq+7H0kiOG7HOWagl9LFAVMV3NMEiiIlIBnAi8bRfdaH9leiAZTRuAAVaLyHoRudYuG2WM2QvWBw0oS0JckS6j+z9ZsvcZxN5Hw+lzdxXWWVzIRBF5V0T+LiKnJiGeaO/bcNpfpwI1xphtEWUJ32c9csSQfc5SLaFLlLKk3ncpIjnA/wFfN8Y0Ab8EjgFmAXuxvu4l2inGmNnAOcCXReS0JMQQk4hkAMuAx+2i4bDPejMsPnci8l3AD6yyi/YC440xJwLfAB4WkbwEhhTrfRsW+8u2gu4nDgnfZ1FyRMyqUcr6td9SLaFXA+MipsuBPUmKBRFxY71Rq4wxTwAYY2qMMQFjTBD4NUP4VTMWY8wee7gfeNKOoSb09c0e7k90XBHOAd4xxtTA8Nhntlj7KOmfOxG5AjgPuNzYDa52k0atPb4eq636uETF1Mv7lvT9BSAiLuAi4NFQWaL3WbQcwRB+zlItof8DmCQiE+2zvMuAp5MRiN02dz+w2Rjz3xHlkW1eFwKbei47xHFli0huaBzrgtomrP10hV3tCuCpRMbVQ7ezpmTvswix9tHTwOftuxBOAhpDX5kTQUSWAt8Clhlj2iLKS0XEaY8fDUwCdiQwrljv29PAZSLiEZGJdlxrExVXhLOALcaY6lBBIvdZrBzBUH7OEnG1d5CvHJ+LdbX4Y+C7SYxjIdbXoY3Ae/brXOB3wPt2+dPA6ATHdTTWHQYbgA9C+wgoBl4EttnDoiTttyygFsiPKEv4PsM6oOwFfFhnRlfH2kdYX4XvsT9z7wNzExzXdqy21dDn7F677nL7Pd4AvAN8OsFxxXzfgO/a+2srcE6i30u7/CHgSz3qJnKfxcoRQ/Y505/+K6VUmki1JhellFIxaEJXSqk0oQldKaXShCZ0pZRKE5rQlVIqTWhCV0qpNKEJXSml0sT/B6F+FNCrPhqzAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(model_1.history[\"loss\"])\n",
    "plt.plot(model_2.history[\"loss\"])\n",
    "plt.plot(model_3.history[\"loss\"])\n",
    "\n",
    "plt.title(\"loss_function_nn_disney_3_models - Training\")\n",
    "plt.legend([\"1 hidden layer\", \"2 hidden layers\", \"3 hidden layers\"])\n",
    "plt.savefig('../Images/loss_function_nn_disney - Training - 3 hidden layers.png')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 717,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Train vs test for deep net\n",
    "plt.plot(model_3.history[\"loss\"])\n",
    "plt.plot(model_3.history[\"val_loss\"])\n",
    "plt.title(\"Data/loss_function_nn_disney - Training Vs. Validation - 3 hidden layers\")\n",
    "plt.legend([\"train\", \"test\"])\n",
    "plt.savefig('../Images/loss_function_nn_disney - Training Vs. Validation - 3 hidden layer.png')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 722,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " Evaluate on test data for 5-minute rolling positive prices sentiments:\n",
      "\n",
      " For unshuffled valuation, test loss on 3-layer neural network model is Loss: 0.2029.\n"
     ]
    }
   ],
   "source": [
    "# Evaluate the model on the test data using `evaluate`\n",
    "print(f'\\n Evaluate on test data for 5-minute rolling positive prices sentiments:')\n",
    "val_loss, val_mse = nn.evaluate(X, y, verbose=0)\n",
    "print(f'\\n For unshuffled valuation, test loss on 3-layer neural network model is Loss: {val_loss:,.4f}.')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 723,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Make some predictions\n",
    "# Make predictions using the testing data X_test\n",
    "predicted = nn.predict(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 726,
   "metadata": {},
   "outputs": [],
   "source": [
    "real = y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 727,
   "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>Real</th>\n",
       "      <th>Predicted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0.442079</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0.057312</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>1.028403</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>-0.005891</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0.100117</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>1</td>\n",
       "      <td>0.620261</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>1</td>\n",
       "      <td>0.453141</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>1</td>\n",
       "      <td>0.953010</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0</td>\n",
       "      <td>0.337146</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1</td>\n",
       "      <td>0.721791</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0</td>\n",
       "      <td>0.423501</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0</td>\n",
       "      <td>0.567401</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>1</td>\n",
       "      <td>0.339238</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0</td>\n",
       "      <td>0.343152</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>1</td>\n",
       "      <td>0.290719</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>1</td>\n",
       "      <td>0.085541</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0</td>\n",
       "      <td>-0.304741</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>1</td>\n",
       "      <td>0.620823</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>0</td>\n",
       "      <td>0.568212</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>1</td>\n",
       "      <td>0.492780</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Real  Predicted\n",
       "0      0   0.442079\n",
       "1      0   0.057312\n",
       "2      1   1.028403\n",
       "3      0  -0.005891\n",
       "4      0   0.100117\n",
       "5      1   0.620261\n",
       "6      1   0.453141\n",
       "7      1   0.953010\n",
       "8      0   0.337146\n",
       "9      1   0.721791\n",
       "10     0   0.423501\n",
       "11     0   0.567401\n",
       "12     1   0.339238\n",
       "13     0   0.343152\n",
       "14     1   0.290719\n",
       "15     1   0.085541\n",
       "16     0  -0.304741\n",
       "17     1   0.620823\n",
       "18     0   0.568212\n",
       "19     1   0.492780"
      ]
     },
     "execution_count": 727,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Create a DataFrame of Real and Predicted values\n",
    "price_moves = pd.DataFrame({\n",
    "    \"Real\": real.ravel(),\n",
    "    \"Predicted\": predicted.ravel()\n",
    "})\n",
    "price_moves"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 728,
   "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>Real</th>\n",
       "      <th>Predicted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Real  Predicted\n",
       "0      0          0\n",
       "1      0          0\n",
       "2      1          1\n",
       "3      0          0\n",
       "4      0          0\n",
       "5      1          1\n",
       "6      1          0\n",
       "7      1          1\n",
       "8      0          0\n",
       "9      1          1\n",
       "10     0          0\n",
       "11     0          1\n",
       "12     1          0\n",
       "13     0          0\n",
       "14     1          0\n",
       "15     1          0\n",
       "16     0          0\n",
       "17     1          1\n",
       "18     0          1\n",
       "19     1          0"
      ]
     },
     "execution_count": 728,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Transform Gender column\n",
    "def changePredicted(predicted):\n",
    "    if predicted >= 0.5:\n",
    "        return 1\n",
    "    else:\n",
    "        return 0\n",
    "\n",
    "price_moves[\"Predicted\"] = price_moves[\"Predicted\"].apply(changePredicted)\n",
    "price_moves"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 699,
   "metadata": {},
   "outputs": [
    {
     "name": "stdin",
     "output_type": "stream",
     "text": [
      "Enter ticker : DIS\n"
     ]
    }
   ],
   "source": [
    "#TODO: To be replaced by dropdown widgets from maui_stock_data.ipynb\n",
    "tickers = ['SPY','DJIA','AAPL', 'GOOG', 'FB', 'NFLX','DIS', 'MSFT', 'IBM']\n",
    "ticker = input (\"Enter ticker :\") "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 700,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot the real vs predicted values as a line chart\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "plt.figure()\n",
    "price_moves.plot(title=f\"Real Vs. Predicted {ticker} Move on Negative Sentiments\")\n",
    "plt.savefig(f'../Images/{ticker}_real_vs_predicted_nn.png')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 701,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXsAAAEWCAYAAACHVDePAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nO3dd5hc5XX48e/Z2dnei1ZlVUEUSQYhFoxNL6YFkJ0QgwwGV4Id7OSH7RjXYGwnxE7c7RDiAHEJGMfBYEwMDqbYpkpYICQhJITKqm7vfc/vj/ed3dnRzO6sNLuzM3M+zzPP3DZ3zm3nvve9TVQVY4wx6S0r2QEYY4yZepbsjTEmA1iyN8aYDGDJ3hhjMoAle2OMyQCW7I0xJgNYsk8xInKOiNQnO45EEJEdInJBsuOYLiKyQEQ6RSQQx7DTvpwn858icquI/GSc/h8RkQN+eisTF6U5XFOa7NNhYxaRe0REReTUsG5Hi8iMvEHBx7pBRLLCun1FRO6J8/dPisiHpizAwyQiq0VkvYi0i0ijiDwuIouSHdd4Itd/Vd2lqkWqOpSAcatPptlh3bJF5GCy100RCQLfAC7009uUgHHuEJEeEekQkVYReUZEboxYz+8Rka+EtX9QRF7zvzkgIr8WkeIjjSWOWI9oG5wqVrKPTzPwlQmHOkLiJGKZzAWuTsB4pkR4gopz+KOBHwGfAEqBxcAPgOHER5dSWoFLwtovBVqSFEu4GiAP2DjZH06wDVyuqsXAQuB24NPAf8QYz9nAPwBr/G+OB+6fbDxHYMZtg0lL9iLyYRHZJiLNIvKQiMz13UVEvulLKG0i8oqIrPD9LhWRTX5PvUdEPhlj3Fki8nkR2enH8yMRKfX9Fvk97/UissuXEj83Qbj/CZzgV6Bo/1cqIv8hIvt8XF8JHapHHu6G/X+2b39SRL4qIn8EuoElIvJ+Ednsp3O7iPzV5OYuXwO+FCupishpvmTUKiIvi8g5vvtXgTOB7/nD7++JyJdE5Lu+f1BEukTka749X0R6RaTct18hIhv9eJ8UkePD/nOHiHxaRF4BuiJjE5HjRORNEYm2gawE3lTVx9XpUNVfqOou/9ssEblFRN4QkSYRuV9EKsLG/V6/LjSJyOckrMQdpTQ4pipDROaKyC9EpMHH9/Gwfrf6//qRX1YbRaTO9/sxsAD4lZ+Xfxdl2R/pcv4xcF1Y+3W4nWL4fJ3rt69mv719OKxfvp/+FhHZBJwS5bdRpz0WETkG2OJbW0Xkd77720XkRXHb9Isi8vaw3xyyDYz3H6rapqoPAVcB14vPDxFOAZ5V1T/53zSr6n+qakeMuMebTzGX8zgOdxs8V0Q2hA33fyLyQlj7H0Tknb750+LyTYeIbBGR88eNSFWn7APsAC6I0v08oBFYBeQC3wWe9v0uAtYBZYDg9shzfL99wJm+uRxYFeN/PwBsw600RcD/AD/2/RYBCvw7kA+cCPQBx8cY1z24Uv3HgT/4bke7WTcyzC+BfwMKgVnAC8Bf+X63Aj8JGzb0/9m+/UlgF7AcyAaCwJ8BR/npPxu3Aazyw58D1I8zzxVY6ufhh3y3rwD3+OZ5QBOuFJgFvMO3V4fF86GIZbXBN78deAN4Pqzfy775GKDLjy8I/J1fBjlh68J6YD6QH75++PVgF3BZjGlaAvQC3wTOBYoi+v8t8BxQi1uf/g241/dbBnQCZ/l+3wAG8etlaPmGjWtk/vr5sw74IpDj49gOXBS2bHv9vAwA/wg8F2v9j7Lsj3Q5rwAO4LaVMt+8grHr5lO4o6A83E6zATjf97sd+D1Q4ZfLq5Oc9p/EiC1yOitwRxzvxa3ja3x7ZaxtYBK5ZBfwkchliSu09ABfAk4HcifIVePNp3GXcyK3Qf//PUCVnxf7gb1AMS5f9QCVwLHAbmBu2Dw/arxpTFbJ/hrgLlV9SVX7gM8AbxNXBzuAm7DjAFHVzaq6z/9uAFgmIiWq2qKqL40z/m+o6nZV7fTjvzpiL/slVe1R1ZeBl3FJfzz/BiwQkfDDZkSkBnco/beq2qWqB3FJaTKHcPeo6kZVHVTVAVX9taq+oc5TwGO4lTdeCnwB+KKI5Eb0uxZ4RFUfUdVhVf0tsBa34kXzLLBU3Em2s3CHzfNEpAiXoJ7yw10F/FpVf6uqA8A/41bOt4eN6zuqultVe8K6nQk8BFyvqg9HnRjV7bjkNw93KN7oS6RFfpC/Aj6nqvV+fboVuNIv7yuBh1X1ad/vC8Rf/XMKbid4m6r2+zj+nbHL9g9+Xg7hStoTrUfh03Wky7kX+BVu3l+Nm4+9oZ4iMh84A/i0qvaq6nrgh7ikC/Bu4KvqSr27ge9Mctrj9WfAVlX9sV/H7wVeAy4PG2bMNjCJce/F7UzGUNXfA3+OK0j8GmgSkW9IlJPjccwnmPxyPqxtUFV7ffNZQB3wCvAH3A7rNNx8bAKGcIWXZSISVNUdqvrGeAElK9nPBXaGWnxCbgLmqervgO8B3wcOiMidIlLiB/0LXFLaKSJPicjb4hm/b87G1SWG7A9r7sYdAcTkE8WX/UfCei3ElWT3+UOyVtyOYdZ444uwO7xFRC4Rkef8IWUrbpqrJjE+VPURXKnnhoheC4G/DMXqx38GMCfGeHpwK9/ZuBXwKeAZ3MoXnuwjl+mwn655sabTuxF4RlWfmGB6nlPVd6tqNS4hngWEqt8WAg+ETc9m3MZQ4+PaHTaeLty6Fo+FwNyIefVZxl+P8mIdukdKxHLGVdtcR5QqHNy0N+vYqoudjC6TMfOGsdtMPNMer8jtMTIOiL5uxGMe7pzaIVT1f1X1ctzOYDXwPiDaxQcTzSc4jOV8BNvgU7jCTWh7exK3rY1sb6q6DXdEeytwUETuE18VHkuykv1e3AQDICKFuEOTPQCq+h1VPRl3WHcM8Cnf/UVVXY1LpL8k9gmXMePH1Z0O4g5zj8TduBOE7wrrthtXDVSlqmX+U6Kqy33/LqAgbPjZUcY7cvWELwX8AlcyrlHVMuARxu5g4vV5XEIM///duCqtsrBPoareHhlLmKdwVTYnAS/69ouAU4Gn/TCRy1RwVQN7ok1nmBtxR0zfjHeiVPVFXNVcqK52N3BJxDTlqeoeXNXf/LC4CnDrWsh4y2c37lxB+HiLVTXWUdAhocbqkcDl/HtckqjBlQDD7QUqZOwVKAsYXSZj5o3vF3Kk0x4Zx8KIbuFxwDjzKhYROQWXkCOnewxfen4c+B2j60xkfOPNpyNxONtgZLJ/iohk76frv1T1DNy8VeCfxgtkOpJ9UETywj7ZwH8B7xeRlX6l/wdcPfAOETlFRN4q7vKtLtxh6ZCI5IjINSJS6g/z2nGlt2juBf6fiCz2h/r/APxMVQePZEL872/FXQUQ6rYPd/j9LyJSIu5k4VEyejJ3PXCWuGusS3FVSuPJwR2eNQCDvtrowsOM90lgA3B9WOefAJeLyEUiEvDL5BwRqfX9D3DoCbKncCXHTaraj6/XxyWDBj/M/cCficj5ftl9ArcTfGaCMDuAi3Hz6PZoA4jIGeJO6M/y7ccBV+Dq6QHuAL4qIgt9/2oRWe37/TdwmR9HDnAbY9f79cClIlIhIrNxpaWQF4B2fyIs38+vFT7JxCPavAxJyHJWVcVVh1zhm8P77cbN/3/0y/kE4IPAT/0g9wOfEZFyv/w/FvbzI532cI8Ax4jIe8RdHnoV7lxK1Gq7ifjt7DLgPtx5gw1RhlktIlf7aRNxl06fzeg6MyKO+XTYDnMbfAZXJ38q8IKqbsQl9LfiC1cicqyInOfzZy+uLn/cS3qnI9k/4gMJfW71e9kv4Eo2+3AnqUJ1gSW4usEW3KFUE670A64ObYeItONKhNfG+M+7cPVqTwNv4mbGx2IMO1n3+pjDXYfbeDf5uP8bf0jm6+N+hqt7W8cEK7g/lPw4bkNsAd6Dq4s9XJ8nrE7Tr9ircYfkDbhSxqcYXRe+javvbhGRUB3uM7j691ApfhNunj4dNt4tuOXxXdzJ98txl8r1TxSgqrbiTlJdIiJfjjJIKy65bxCRTuA3wAO4Kx5CMT8EPCYiHbgN+q1+3BuBv8YVMPbh5mn4jUM/xp2z2YHbaf8sLK4hPx0rcetRI64ut3SiafL+Efi8P1Qfc+VYIpezr+uOdZnjGtzJu724efb3fp0Ed/JyJ27aHsPNi9A4j3Taw+NrAi7DFQCacCfvL1PVxkmO6ld++e7GlZa/Abw/xrAtwIeBrbiC4U+Ar6tqrAQ+3nw6UpPaBn1V40vAxrDt51lgp7pzguAKCrfjlst+XG3HZ8cLQiIKA8akPRHZgbtK4v+SHYsx08VuqjLGmAxgyd4YYzKAVeMYY0wGsJK9McZkgEk9kCqRqqqqdNGiRcn6e2OMSUnr1q1r9DcXTkrSkv2iRYtYu3Ztsv7eGGNSkohE3o0cF6vGMcaYDGDJ3hhjMoAle2OMyQBJq7M3xpjJGBgYoL6+nt7e3okHTgN5eXnU1tYSDAYTMj5L9saYlFBfX09xcTGLFi3CPVQ1fakqTU1N1NfXs3jx4oSMc8JqHBG5S9yr/V6N0f8aca8OfEXca7bifnmDMcbEq7e3l8rKyrRP9AAiQmVlZUKPYuKps78H9wjaWN4EzlbVE3Av9rgzAXEZY8whMiHRhyR6WidM9qr6NDHeBOP7P6OqoTfah94DOmW27O/gnx/dQnPXhE/ONcYY4yX6apwPAv8bq6eI3CAia0VkbUNDQ6zBxvVmYyffe2Ib+9sy4ySNMSb5mpqaWLlyJStXrmT27NnMmzdvpL2/P76C5/vf/362bNkyxZHGlrATtCJyLi7ZnxFrGFW9E1/NU1dXd1hPYCvJc2em23sn805iY4w5fJWVlaxfvx6AW2+9laKiIj75yTHvo0FVUVWysqKXoe++++4pj3M8CSnZ+9d4/RBY7d9KM2WKfbLv6D2iNwwaY8wR27ZtGytWrODGG29k1apV7Nu3jxtuuIG6ujqWL1/ObbfdNjLsGWecwfr16xkcHKSsrIxbbrmFE088kbe97W0cPHhwnH9JjCMu2YvIAtzLn9+rqq8feUjjK8l3Ibf3WMnemEz1pV9tZNPe9oSOc9ncEv7+8uWT/t2mTZu4++67ueOOOwC4/fbbqaioYHBwkHPPPZcrr7ySZcuWjflNW1sbZ599Nrfffjs333wzd911F7fccktCpiOWeC69vBf3/sNjRaReRD4oIjeKyI1+kC8ClcAPRGS9iEzp081GS/aW7I0xyXfUUUdxyimj72G/9957WbVqFatWrWLz5s1s2rTpkN/k5+dzySWXAHDyySezY8eOKY9zwpK9qq6ZoP+HgA8lLKIJFOf5kr1V4xiTsQ6nBD5VCgsLR5q3bt3Kt7/9bV544QXKysq49tpro14rn5OTM9IcCAQYHJz6fJZyz8YJBrLIDwasZG+MmXHa29spLi6mpKSEffv28eijjyY7pBEp+biEkvxs2nusZG+MmVlWrVrFsmXLWLFiBUuWLOH0009PdkgjkvYO2rq6Oj3cl5e84xtPsbSmiB9cc3KCozLGzFSbN2/m+OOPT3YY0yraNIvIOlWtm+y4Uq4aB1y9vZXsjTEmfimZ7Evyg3ZTlTHGTEJKJvvivKDdVGWMMZOQksm+JC/bbqoyxphJSMlkHyrZJ+vksjHGpJqUTPYl+dn0Dw3TNzic7FCMMSYlpGSyL7YnXxpjptk555xzyE1S3/rWt/joRz8a8zdFRUVTHVbcUjLZl4QemWCXXxpjpsmaNWu47777xnS77777WLNm3CfKzBgpmuztYWjGmOl15ZVX8vDDD9PX1wfAjh072Lt3LytXruT8889n1apVvOUtb+HBBx9McqTRpezjEsAehmZMxvrfW2D/hsSOc/Zb4JLbY/aurKzk1FNP5Te/+Q2rV6/mvvvu46qrriI/P58HHniAkpISGhsbOe2007jiiitm3PtyrWRvjDFxCq/KCVXhqCqf/exnOeGEE7jgggvYs2cPBw4cSHKkh0rJkv3ICVqrszcmM41TAp9K73znO7n55pt56aWX6OnpYdWqVdxzzz00NDSwbt06gsEgixYtivpY42RLzZL9SDWOleyNMdOnqKiIc845hw984AMjJ2bb2tqYNWsWwWCQJ554gp07dyY5yuhSL9nvep78B97HnKwWq8Yxxky7NWvW8PLLL3P11VcDcM0117B27Vrq6ur46U9/ynHHHZfkCKNLvWqc7iZk869YnHu6VeMYY6bdu971rjF371dVVfHss89GHbazs3O6wppQ6pXs88sBmJPTYyV7Y4yJU+ol+4IKAGqC3bTZw9CMMSYuqZfsfcl+VrYle2MyTSY9/DDR05qyyb4yy5K9MZkkLy+PpqamjEj4qkpTUxN5eXkJG2fqnaANBCGnmPKsLtrsBK0xGaO2tpb6+noaGhqSHcq0yMvLo7a2NmHjS71kD5BfThkdtPX0o6oz7rZkY0ziBYNBFi9enOwwUlbqVeMAFJRTrB0MDCk9A0PJjsYYY2a81Ez2+eUUDrUD0Npt9fbGGDORFE32FeQNumRvJ2mNMWZiEyZ7EblLRA6KyKsx+ouIfEdEtonIKyKyKvFhRsgvJ3egDbCSvTHGxCOekv09wMXj9L8EWOo/NwD/euRhTaCgguz+NoRhK9kbY0wcJkz2qvo00DzOIKuBH6nzHFAmInMSFWBU+eWIDlNMN209/VP6V8YYkw4SUWc/D9gd1l7vux1CRG4QkbUisvaIrpXNd49MKJMuK9kbY0wcEpHso13kHvUWN1W9U1XrVLWuurr68P9x5C7aTquzN8aYOCQi2dcD88Paa4G9CRhvbD7Zz8vttZK9McbEIRHJ/iHgOn9VzmlAm6ruS8B4Y/NPvpwd7KHVkr0xxkxowscliMi9wDlAlYjUA38PBAFU9Q7gEeBSYBvQDbx/qoId4Uv2NcFutliyN8aYCU2Y7FV1zQT9FfjrhEUUj7wyAKoC3VZnb4wxcUjNO2gD2ZBbSmVWp9XZG2NMHFIz2QMUlFNKJ63ddp29McZMJHWTfX4FpdpOR98gw8Pp/zIDY4w5Eqmb7AsqKRxqRxU6eu0lJsYYM57UTfaFVRQMtgDQao9MMMaYcaVusi+oJK+/FbAnXxpjzERSOtkHhnrIo89urDLGmAmkbrIvrAKgknZauqwaxxhjxpO6yb6gEoBy6aDZkr0xxowrhZO9K9lXSQctdq29McaMK3WTva/GmZ/bbcneGGMmkLrJ3j/5ck5OFy1ddoLWGGPGk7rJPq8MsrKZnd1ldfbGGDOBCZ96OWOJQEElVdJp1TjGGDOB1C3ZAxRUUUG7leyNMWYCKZ7sKyjVNlq7B3CP1TfGGBNNaif7wiqKh9vpHxqmq38o2dEYY8yMldrJvqCK/AH3MDS7i9YYY2JL8WRfSe5AGwGGrN7eGGPGkdrJ3t9YVUYnzXZFjjHGxJTayd4/H6dCOuz1hMYYM47UTvaFoefjtNFsd9EaY0xMqZ3si2oAqJY2O0FrjDHjSO1kX1gNwIIcq7M3xpjxpHayzy+HrCDzgh1WsjfGmHGkdrIXgaJZzA7YIxOMMWY8cSV7EblYRLaIyDYRuSVK/wUi8oSI/ElEXhGRSxMfagxFs5glbTRZsjfGmJgmTPYiEgC+D1wCLAPWiMiyiME+D9yvqicBVwM/SHSgMRXOolxbaezsm7a/NMaYVBNPyf5UYJuqblfVfuA+YHXEMAqU+OZSYG/iQpxA0SxKhppp7R5gYGh42v7WGGNSSTzJfh6wO6y93ncLdytwrYjUA48AH0tIdPEoqqFgoAVhmKZOq8oxxpho4kn2EqVb5POE1wD3qGotcCnwYxE5ZNwicoOIrBWRtQ0NDZOPNpqiWWTpEOV0WlWOMcbEEE+yrwfmh7XXcmg1zQeB+wFU9VkgD6iKHJGq3qmqdapaV11dfXgRRyqaBUC1tNJgyd4YY6KKJ9m/CCwVkcUikoM7AftQxDC7gPMBROR4XLJPUNF9AoUu2VdJG40dluyNMSaaCZO9qg4CNwGPAptxV91sFJHbROQKP9gngA+LyMvAvcD7dLpeHRV6ZAJtNFqdvTHGRBXXC8dV9RHcidfwbl8Ma94EnJ7Y0OLkq3HmZLdbnb0xxsSQ2nfQAuQWQ3YeC3I6aLJkb4wxUaV+svePTJgTaLdqHGOMiSH1kz1A4SxmZbVZNY4xxsSQHsm+eDaVw82W7I0xJob0SPYlcykdbKS5q5+h4em5CMgYY1JJeiT74jnkDXWSp732qGNjjIkiPZJ9iXtUz2yxqhxjjIkmTZL9HABqpMWSvTHGRJEeyb54LgCzaeZAuyV7Y4yJlB7J3pfs50gzB9p7kxyMMcbMPOmR7HMKIa+UBcFW9rdZsjfGmEjpkewBiucyP7uN/VayN8aYQ6RPsi+Zw5wsq8Yxxpho0ijZz6VquMmqcYwxJor0SfbFcykebKa1s4tBe/G4McaMkT7JvmQOglKpbfZ6QmOMiZBGyX70LlqryjHGmLHSKNm7G6vmSpOdpDXGmAjpk+zLFgAwTxqsZG+MMRHSJ9nnlaJ5ZSzIamS/PTLBGGPGSJ9kD0jZApYE7Vp7Y4yJlFbJnrIFzLdqHGOMOUSaJfuFzBo+yL7W7mRHYowxM0qaJfsF5GovPW0HGbbXExpjzIi0S/YAs4cPcLDDTtIaY0xIeiX78oUA1Eoj9S1WlWOMMSHplexL5wNQKw3Ut/QkORhjjJk54kr2InKxiGwRkW0ickuMYd4tIptEZKOI/Fdiw4xTXgmaX+6TvZXsjTEmJHuiAUQkAHwfeAdQD7woIg+p6qawYZYCnwFOV9UWEZk1VQFPRMoWsKS3iV9Zyd4YY0bEU7I/FdimqttVtR+4D1gdMcyHge+raguAqh5MbJiTULaQhVkN7LaSvTHGjIgn2c8Ddoe11/tu4Y4BjhGRP4rIcyJycbQRicgNIrJWRNY2NDQcXsQTqTyK2UP72dfcOTXjN8aYFBRPspco3SIvYs8GlgLnAGuAH4pI2SE/Ur1TVetUta66unqyscan8miyGUTadjFk19obYwwQX7KvB+aHtdcCe6MM86CqDqjqm8AWXPKffpXub+frXg522GMTjDEG4kv2LwJLRWSxiOQAVwMPRQzzS+BcABGpwlXrbE9koHGrPBqAJbLfLr80xhhvwmSvqoPATcCjwGbgflXdKCK3icgVfrBHgSYR2QQ8AXxKVZumKuhxFVQwlFvGYtnHziY7SWuMMRDHpZcAqvoI8EhEty+GNStws/8klwhSdTRLevbzbGNXsqMxxpgZIb3uoPWyKo/m6MB+tjfaFTnGGANpmuypPJoabWTPweTUJBljzEyTpsn+KAC0abs96tgYY0jbZO+uyJk/vIe9bXZFjjHGpGeyrzoGlSyOydrN9gY7SWuMMemZ7IN5DJUfxfGyizftihxjjEnTZA8E5qzg+KzdbG+wK3KMMSZtk73ULGe+HGTvwSl64JoxxqSQtE321KwAQA5uTnIgxhiTfGmc7JcDUNW9jbbugSQHY4wxyZW+yb50PoPBIo6TXWze357saIwxJqnSN9mLMFy9jOOydrF5nyV7Y0xmS99kDwTnncjyrJ28trcl2aEYY0xSpXWyl9pTKKSXrvpNEw9sjDFpLK2TPbV1AJQ1r2dwaDjJwRhjTPKkd7KvWEJfThlv0a1stztpjTEZLL2TvQgDs1dxUtZWNu21k7TGmMyV3skeyF98GkfLXjbvqE92KMYYkzRpn+wDC04hS5SeN19IdijGGJM0aZ/smXcyw2Qxq2UdvQNDyY7GGGOSIv2TfV4p7ZUn8DZ5lQ172pIdjTHGJEX6J3sgZ+l5nChvsOGNXckOxRhjkiIjkn3BceeTLcP0vP50skMxxpikyIhkT+0p9EkelQefQdVeQG6MyTyZkeyzc2muqqNu6GW2HOhIdjTGGDPtMiPZA4XLLuLorL28sn5dskMxxphplzHJvmTVn7uGzQ8mNxBjjEmCuJK9iFwsIltEZJuI3DLOcFeKiIpIXeJCTJDSWuoLl7Os9Sm73t4Yk3EmTPYiEgC+D1wCLAPWiMiyKMMVAx8Hnk90kInSt/QyVsh2Xnn15WSHYowx0yqekv2pwDZV3a6q/cB9wOoow30Z+BrQm8D4Emre268GoO2FnyU5EmOMmV7xJPt5wO6w9nrfbYSInATMV9WHxxuRiNwgImtFZG1DQ8Okgz1SebOWsDXvLRy375cMD1lVjjEmc8ST7CVKt5GL1UUkC/gm8ImJRqSqd6pqnarWVVdXxx9lAnWsuI757Of1536dlP83xphkiCfZ1wPzw9prgb1h7cXACuBJEdkBnAY8NCNP0gLHnnsNLVrM0Iv/kexQjDFm2mTHMcyLwFIRWQzsAa4G3hPqqaptQFWoXUSeBD6pqmsTG2piFBYW8seySziv9b8ZbNxOdtWSZIdkjEkH/V3Q1QjdTe7T1QjdjaPf3c2+XzOsei+c/jfTGt6EyV5VB0XkJuBRIADcpaobReQ2YK2qPjTVQSZazpk3MfSrB2j49Vepvd5K+MaYCKrQ2xY9aXf5ZD7SzQ8z2BN9XFlBKKiEwiooqICa5VBaO73TA0iynhVTV1ena9cmp/A/ODTML756DX8x/CjZH18HFYuTEocxZpoMD/mSdWNE8o4shfsk3t0Ew4PRxxUsgIIqKKz031UumY8k9LBuhVWQWwIS7dTn4RGRdao66WryeKpx0k52IIu2VTcx+OJv6XvkCxRe+5Nkh2SMmYzBvgmSduPY9p5Wwq4rGSuvdDRBly+C2pN98q4KS96+vaAScgqmc0oTJiOTPcDlZ67iB8+9k5u3/RxefwyOuTDZIRmTmVShvzMiQUepJhlpb4b+GA80lKyxibpmWfSkHV4aDwSnd3qTJGOT/ZzSfPYuv4Ftrz3D4odvJvDRP7o9vDHmyAwPQ29r2InJKNUk4SctuxphqC/6uAK5YxNzxZIYSds355VBVsY88mtSMjbZA3zkgmX83YYP8/P2L8MvPwpX/SShdWvGpIWhgSgl7oikHX7SsrsZNMZNiznFPlFXQvFcmH3C2LrukWb/nVNk22SCZHSyP6q6iHlvOYevbd7OLa/9CJ68Hc79TLLDMmbqqEJfx2gdd6JrJuwAABPzSURBVGQS726KSNxN7qqUWPLLR0vVlUfBgrfGPmlZUAnBvOmbVjNGRid7gE9ceAzv2HgJ55Yd4K1P3Q65xfD2m5IdljGOqrt+u6/dJd3e0Hcb9LWN7TbQ7YYd6HHNA92uebAXhgZhqN+NZ6g/+n8FckaTckEFzD1pNGEfcqVJlUv0gYxPISkj45fUwspCbjzrKN7zu2t54dgAlY99Droa4Py/t7o/k1ihuuxDStShT/Noabq7eTSpx6oSCckKuvNNOYXuE8x3lwcWzXbN2XnuJGQg6C4DDK8qKfDXfluVSdrL+GQP8JFzjuaB9Xu48uAHePSkWeT88Vuw/xVY/X0omZvs8MxMFCpxh9dTR60aCeve0ww6HH18wYLREnXoRGReqfvklvjm0HfZ2G7ZeZakzYQy8qaqaNbtbOYv73iW1SfO5ZtH/wke/ZwrMZ39KTj1BsjOTXaIZioN9keUssf5hOq0Y11BIoGIqo+KsOqQqrFJPfRJ0Wu3zfSzm6qO0MkLK/j4+Uv51v9t5cT55/K+G/8Aj3wKHvs8PPNdWHU9nPw+KJ034bhMkg0PuZtoeprHVo90HoDOg67U3dPiqlR6Wlz/vvbY4wvddFNQCSW1MPvE0StKon3ySq2kbWYcS/ZhPnbeUl7d085tD29iQWUd5733f2D7k/DsD+Dpr8Pv/wWOOg+OuQiWXgjlC5MdcuYYHnKJu/PAaNIO/+7w3Se8W7Js9HrsgkqoPPrQ0nb4lST55Rlz041Jb1aNE6Grb5Cr7nyW1w908sPr6jjrGP/c/eY3Yd3dsOlBaNnhulUsgXl1MO9kmHMiVC11CcJKdfEJ3Tk5krD3H5rER0rjDdFPVOYUQdEsdzKyaNZoos4PVZOUh1WpzLJL/0zKO9xqHEv2UbR09fOeHz7PGw2dfOfqk7h4xezRnqrQtA1efxR2PQv1a6Fz/2j/vFKoXOpKjOULoXQ+lM1336W16V/3P+aZJQ2ufrurIXp750F3eWCkrGyXmItmQbFP4kU1/jNr9LtwFuQWTf80GpNEluwTrLmrnw/c8yIv17fy6YuP44Yzl5CVFaXErgrte+DAJrcTaNrqv9+A9r0cUp1QVAPFc0aTVZH/FFaPJrLCald9kKwjhOFhV+LubXP12r1h13P3tftv363L30XZ1eCqWWLVfWcFXak7/FrtkQReMzah55fbZa/GxGDJfgr09A/xiZ+v55EN+zn7mGr++S9PpLp4EiXzwX63I2jbDa27oa0e2naN1i93NbhPtEepSgDyy1zdcn552Md3C+a5BBrIcTe2SJb7v6E+V7oe7Ito7nc314w3zECPS/L9nRNPW3aeO4oprPZVJNURyTzUXm0nLY1JIEv2U0RV+enzu/jyw5sozgvy9StP4NzjZiXuD0I32oTXTXcecFeI9LSMfkJXjvS0jH/7eohkuYdIZfvPmOYcl6yzc3z3UHueu4M4p8h955dFudbbN2fnJG4eGGPiZsl+im3Z38HH7n2J1w90ctHyGr5w2TJqy5N0bfTwsCupDw+4h1QNDbibdcITu93GbkxasmQ/DfoGh/jh79/ku7/bCsBfnXUUHzpzMcV5dmmeMWZ6HG6yt7Ngk5CbHeCvzz2axz9xDucfV8O3H9/KmV97gn998g26+2O8wswYY2YAK9kfgVfqW/nGb1/nyS0NlBcEueatC7nubQuZVWLXchtjpoZV4yTRup3N3PHUdv5v8wGys4TLT5zLB89YzPK59uYrY0xiWbKfAXY0dnH3H9/k5+vq6e4f4uSF5Vx9ynwuO2Eu+TmBZIdnjEkDluxnkLaeAe5/cTf3vriL7Q1dFOdms/qkuVx9ygKWzy1B7HpzY8xhsmQ/A6kqL+5o4b4XdvHrDfvoGxzm6FlFrD5xLqtXzmNBpT3W1hgzOZbsZ7i27gEe3rCXB9fv5YU3mwE4aUEZ71w5jz87YQ5VRWn+zBxjTEJYsk8he1p7eGj9Xh5cv4fX9neQJXDKogouXD6bC5fVML/CSvzGmOimNNmLyMXAt4EA8ENVvT2i/83Ah4BBoAH4gKruHG+cmZzsw23Z38HDr+zlsY0H2HKgA4Blc0q4cHkNFy2fzXGzi62O3xgzYsqSvYgEgNeBdwD1wIvAGlXdFDbMucDzqtotIh8BzlHVq8YbryX7Q+1o7OKxTft5bOMB1u1qQRXmV+Rz/nE1nH1MNW9dUkFBjj0GwZhMNpXJ/m3Arap6kW//DICq/mOM4U8Cvqeqp483Xkv242vo6OPxzQd4dON+nnmjib7BYXICWZyyuJyzllZz5tJqjp9jpX5jMs1UJvsrgYtV9UO+/b3AW1X1phjDfw/Yr6pfidLvBuAGgAULFpy8c+e4NT3G6x0Y4oU3m3n69QZ+v7VxpLqnqiiHUxdXcOqiCk5dXMlxs4ujP3PfGJM2pvKF49GyR9Q9hIhcC9QBZ0frr6p3AneCK9nHGWPGywsGOOuY6pFXJO5v6+XprQ08+0YTz29v4pEN7k1ZJXnZnLKoglMWV7BqQTlvmVdqN3MZY4D4kn09MD+svRbYGzmQiFwAfA44W1X7EhOeiWZ2aR7vrpvPu+vcYqlv6eaFN5tHPo+/dhCAQJZw/JxiTppfzkkLylg5v4xFlYVW+jcmA8VTjZONO0F7PrAHd4L2Paq6MWyYk4D/xlX3bI3nj63Ofuo0dvaxflcrf9rdwp92tfLy7la6+t3Luotzs1k2t4QV80pZMa+EFXNLWVJdRMB2AMakhCmrxlHVQRG5CXgUd+nlXaq6UURuA9aq6kPA14Ei4Of+hOEuVb1issGYxKgqyuWCZTVcsKwGgKFhZevBDtbvauXVvW28uqednzy3k77BYQDygwGOn1PMsrklHFNTPPKpKLS3URmTLuymqgw1ODTMGw1dvLqnze8A2nhtfwcdvaPP5a8qyhlJ/Etriji2ppilNcWU5tvLWoxJlqk8QWvSUHYgi2NnF3Ps7GL+4uRawD3LZ397L68f6OT1/R28fqCD1w92cv/a3XT7aiCAmpJcjqouYlFVIYsqC1hUWcjiqkIWVBaQm20nhI2ZiSzZmxEiwpzSfOaU5nO2v/IHYHhY2dPaw9aDHWzZ38nWAx1sb+zikQ37aO0eCPs9zC3NZ3FVIYuqRncCteUFzCvPpyjXVjdjksW2PjOhrCxhfkUB8ysKOO+4mjH9Wrv72dHUzY7GLt5s7GJHUxc7Grt4aP1e2nvHvqqxrCBIbXk+88rymVdW4JrL86ktz6e2rICS/Gy7ScyYKWLJ3hyRsoIcVhbksHJ+2Zjuqkpr9wA7mrqob+mhvqWHPa3d1Lf0sL2hi6dfb6RnYGjMbwpzAtSU5lFTnEdNSS41JXlhH9c+qyTXqoqMOQyW7M2UEBHKC3MoL8zhpAXlh/RXVVq6B6hv6WbPyM6gh4aOPva397J2ZwsH2/voHxo+5LflBUGf+POYHdoJFOdSWZRLRWEOlYU5VBTmUFaQY5eUGuNZsjdJISJU+KR8Qm1Z1GFCRwf723s50N7LwfY+DrT3+vY+Dnb0smV/Ow0dfQxHuahMBMoLckb+p6Igh4qi0Z2B2zH4HURRDuUFOeRkZ03xlBuTHJbszYwVfnRw/JySmMMNDg3T1NVPU2c/zV39NHf309zZR3NXP01d/SPf2xo6ad7RT0t3P7GuOM4PBijND458SsKa3Seb0oJg1GGsesnMZJbsTcrLDmSN1O3HY2hYae32O4awnUFLVz9tPQNjPvUt3Wza65q7+ofGHW8wIBTkZFOUm01BToDC3GwKcwMUhrrl+m452f577DCuebR7fjBgj7YwCWPJ3mScQJZQWeTq+CdjYGiYjt7BQ3YIbT0DtPcM0Nk3SHffIJ19Q3T3D7r2/iGaOrvp6h+kq2+Irr7BkTuX41GYE6AgN2wHkpNNfk6AgpzAyHdBjtsxFIx0zx7tH/T9c8L7B8gJZNmVTxnGkr0xcQoGskbq+o/EwNAw3f1uh9DVN7oT6OoPfUfvHtqBtHb3s7d1iO7+IXoG3Hh6B+LfgYDb4RUEA2E7juzRnUFwdKeQH8yO2LH4YYPhO5yxw+Rl2xHJTGTJ3phpFgxkUZqfldDHTgwPq0/8Q/T0D9E9MDja7HcsoebQDqK7f4jegVB/N2xX3yANHX1jx9U/GPUE+Hjyg2OPPg7dQbidRF5w7BFHQdgOJurOJBggO2An0Q+HJXtj0kBWlozU+SeaqtI3OOx3IkP0+B3FITuTMTuO0Z1N+I7jQEfvmN/19A9Fvbx2PDmBrEOrsoLjV2/ljxyxHHqkUjByVJOd1ldjWbI3xoxLRMgLBsgLBjj0jokjNzg07Hcihx6FuCORyKOUsB3OwOjRR2vPAPvaesbuTAbGP6keKRgQ8oP+BHnYOZLCMTsFdx6lIOi+C/3OozDUPzf7kB1Jbnbyz5FYsjfGJFV2IIuSQBYleYl/murwsNI7GOUoJNTsj1S6+tyOoatvcGSYrrCqraaufnY1d7t2339gKP66rSxhdMeRm817Tl3Ah89akvDpHY8le2NM2srKEl+6Tnyq6x+p2vI7i/4huvoHR86HdPdF7DT6B123gSGqiyd3JVgiWLI3xpjDkJOdRU52FqWkxvsd0vdshDHGmBGW7I0xJgNYsjfGmAxgyd4YYzKAJXtjjMkAluyNMSYDWLI3xpgMYMneGGMygGisV/ZM9R+LNAA7D+OnVUBjgsNJBItr8mZqbBbX5MzUuGDmxnYkcS1U1erJ/ihpyf5wichaVa1LdhyRLK7Jm6mxWVyTM1PjgpkbWzLismocY4zJAJbsjTEmA6Risr8z2QHEYHFN3kyNzeKanJkaF8zc2KY9rpSrszfGGDN5qViyN8YYM0mW7I0xJgOkTLIXkYtFZIuIbBORW5Icy3wReUJENovIRhH5G9/9VhHZIyLr/efSJMS2Q0Q2+P9f67tViMhvRWSr/56KV4mOF9OxYfNkvYi0i8jfJmt+ichdInJQRF4N6xZ1HonzHb/evSIiq6Y5rq+LyGv+vx8QkTLffZGI9ITNuzumOa6Yy05EPuPn1xYRuWia4/pZWEw7RGS97z6d8ytWfkjuOqaqM/4DBIA3gCVADvAysCyJ8cwBVvnmYuB1YBlwK/DJJM+rHUBVRLevAbf45luAf0rystwPLEzW/ALOAlYBr040j4BLgf8FBDgNeH6a47oQyPbN/xQW16Lw4ZIwv6IuO78dvAzkAov9dhuYrrgi+v8L8MUkzK9Y+SGp61iqlOxPBbap6nZV7QfuA1YnKxhV3aeqL/nmDmAzMC9Z8cRhNfCfvvk/gXcmMZbzgTdU9XDunk4IVX0aaI7oHGserQZ+pM5zQJmIzJmuuFT1MVUd9K3PAbVT8d+TjWscq4H7VLVPVd8EtuG232mNS0QEeDdw71T893jGyQ9JXcdSJdnPA3aHtdczQ5KriCwCTgKe951u8odid013dYmnwGMisk5EbvDdalR1H7gVEZiVhLhCrmbsBpjs+RUSax7NpHXvA7gSYMhiEfmTiDwlImcmIZ5oy26mzK8zgQOqujWs27TPr4j8kNR1LFWSvUTplvRrRkWkCPgF8Leq2g78K3AUsBLYhzuMnG6nq+oq4BLgr0XkrCTEEJWI5ABXAD/3nWbC/JrIjFj3RORzwCDwU99pH7BAVU8Cbgb+S0RKpjGkWMtuRswvYA1jCxXTPr+i5IeYg0bplvB5lirJvh6YH9ZeC+xNUiwAiEgQtyB/qqr/A6CqB1R1SFWHgX9nig5fx6Oqe/33QeABH8OB0GGh/z443XF5lwAvqeoBH2PS51eYWPMo6eueiFwPXAZco76S11eTNPnmdbi68WOmK6Zxlt1MmF/ZwJ8DPwt1m+75FS0/kOR1LFWS/YvAUhFZ7EuHVwMPJSsYXx/4H8BmVf1GWPfwerZ3Aa9G/naK4yoUkeJQM+7k3qu4eXW9H+x64MHpjCvMmNJWsudXhFjz6CHgOn/FxGlAW+hQfDqIyMXAp4ErVLU7rHu1iAR88xJgKbB9GuOKteweAq4WkVwRWezjemG64vIuAF5T1fpQh+mcX7HyA8lex6bj7HQiPrgz1q/j9sifS3IsZ+AOs14B1vvPpcCPgQ2++0PAnGmOawnuSoiXgY2h+QRUAo8DW/13RRLmWQHQBJSGdUvK/MLtcPYBA7hS1QdjzSPcIfb3/Xq3Aaib5ri24epzQ+vZHX7Yv/DL+GXgJeDyaY4r5rIDPufn1xbgkumMy3e/B7gxYtjpnF+x8kNS1zF7XIIxxmSAVKnGMcYYcwQs2RtjTAawZG+MMRnAkr0xxmQAS/bGGJMBLNmbtCUiQzL2aZsJe1qqf4piMu8LMGZSspMdgDFTqEdVVyY7CGNmAivZm4zjn3P+TyLygv8c7bsvFJHH/cO9HheRBb57jbhnyb/sP2/3owqIyL/7Z5Y/JiL5SZsoYyZgyd6ks/yIapyrwvq1q+qpwPeAb/lu38M9avYE3APHvuO7fwd4SlVPxD0/faPvvhT4vqouB1pxd2kaMyPZHbQmbYlIp6oWRem+AzhPVbf7B1btV9VKEWnE3fY/4LvvU9UqEWkAalW1L2wci4DfqupS3/5pIKiqX5n6KTNm8qxkbzKVxmiONUw0fWHNQ9g5MDODWbI3meqqsO9nffMzuCeqAlwD/ME3Pw58BEBEAtP83HhjEsJKIiad5Yt/4bT3G1UNXX6ZKyLP4wo8a3y3jwN3icingAbg/b773wB3isgHcSX4j+CetmhMyrA6e5NxfJ19nao2JjsWY6aLVeMYY0wGsJK9McZkACvZG2NMBrBkb4wxGcCSvTHGZABL9sYYkwEs2RtjTAb4/zWL1AFzGBPTAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plotting loss\n",
    "epochs = 200\n",
    "loss_df = pd.DataFrame(\n",
    "    {\n",
    "        \"Epoch\": range(1, epochs + 1),\n",
    "        \"Train\": model_3.history[\"loss\"],\n",
    "        \"Val\": model_3.history[\"val_loss\"],\n",
    "    }\n",
    ")\n",
    "loss_df.set_index(\"Epoch\", inplace=True)\n",
    "\n",
    "plt.figure()\n",
    "loss_df.plot(title=f\"Loss on Neural Network Sequential Model for {ticker} on News\")\n",
    "plt.savefig(f'../Images/{ticker}_loss_nn.png')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 702,
   "metadata": {},
   "outputs": [],
   "source": [
    "import h5py\n",
    "\n",
    "nn.save('../Models/disney_nn_loss_3_layers.h5')  # creates a HDF5 file 'my_model.h5'\n",
    "del nn  # deletes the existing model"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "mlenv",
   "language": "python",
   "name": "mlenv"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}