s&p500.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Deep Learning"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## A. Data Preparation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import cufflinks as cf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "sp = pd.read_csv('data_csv.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sp.isnull().values.any()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Date                      0\n",
       "SP500                     0\n",
       "Dividend                  1\n",
       "Earnings                  4\n",
       "Consumer Price Index      0\n",
       "Long Interest Rate        0\n",
       "Real Price                0\n",
       "Real Dividend             1\n",
       "Real Earnings             4\n",
       "PE10                    120\n",
       "dtype: int64"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sp.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Date</th>\n",
       "      <th>SP500</th>\n",
       "      <th>Dividend</th>\n",
       "      <th>Earnings</th>\n",
       "      <th>Consumer Price Index</th>\n",
       "      <th>Long Interest Rate</th>\n",
       "      <th>Real Price</th>\n",
       "      <th>Real Dividend</th>\n",
       "      <th>Real Earnings</th>\n",
       "      <th>PE10</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1871-01-01</td>\n",
       "      <td>4.44</td>\n",
       "      <td>0.26</td>\n",
       "      <td>0.4</td>\n",
       "      <td>12.46</td>\n",
       "      <td>5.32</td>\n",
       "      <td>89.00</td>\n",
       "      <td>5.21</td>\n",
       "      <td>8.02</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1871-02-01</td>\n",
       "      <td>4.50</td>\n",
       "      <td>0.26</td>\n",
       "      <td>0.4</td>\n",
       "      <td>12.84</td>\n",
       "      <td>5.32</td>\n",
       "      <td>87.53</td>\n",
       "      <td>5.06</td>\n",
       "      <td>7.78</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1871-03-01</td>\n",
       "      <td>4.61</td>\n",
       "      <td>0.26</td>\n",
       "      <td>0.4</td>\n",
       "      <td>13.03</td>\n",
       "      <td>5.33</td>\n",
       "      <td>88.36</td>\n",
       "      <td>4.98</td>\n",
       "      <td>7.67</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1871-04-01</td>\n",
       "      <td>4.74</td>\n",
       "      <td>0.26</td>\n",
       "      <td>0.4</td>\n",
       "      <td>12.56</td>\n",
       "      <td>5.33</td>\n",
       "      <td>94.29</td>\n",
       "      <td>5.17</td>\n",
       "      <td>7.96</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1871-05-01</td>\n",
       "      <td>4.86</td>\n",
       "      <td>0.26</td>\n",
       "      <td>0.4</td>\n",
       "      <td>12.27</td>\n",
       "      <td>5.33</td>\n",
       "      <td>98.93</td>\n",
       "      <td>5.29</td>\n",
       "      <td>8.14</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         Date  SP500  Dividend  Earnings  Consumer Price Index  \\\n",
       "0  1871-01-01   4.44      0.26       0.4                 12.46   \n",
       "1  1871-02-01   4.50      0.26       0.4                 12.84   \n",
       "2  1871-03-01   4.61      0.26       0.4                 13.03   \n",
       "3  1871-04-01   4.74      0.26       0.4                 12.56   \n",
       "4  1871-05-01   4.86      0.26       0.4                 12.27   \n",
       "\n",
       "   Long Interest Rate  Real Price  Real Dividend  Real Earnings  PE10  \n",
       "0                5.32       89.00           5.21           8.02   NaN  \n",
       "1                5.32       87.53           5.06           7.78   NaN  \n",
       "2                5.33       88.36           4.98           7.67   NaN  \n",
       "3                5.33       94.29           5.17           7.96   NaN  \n",
       "4                5.33       98.93           5.29           8.14   NaN  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sp[sp.isnull().values.any(axis=1)].head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Date</th>\n",
       "      <th>SP500</th>\n",
       "      <th>Dividend</th>\n",
       "      <th>Earnings</th>\n",
       "      <th>Consumer Price Index</th>\n",
       "      <th>Long Interest Rate</th>\n",
       "      <th>Real Price</th>\n",
       "      <th>Real Dividend</th>\n",
       "      <th>Real Earnings</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1763</th>\n",
       "      <td>2017-12-01</td>\n",
       "      <td>2664.34</td>\n",
       "      <td>48.93</td>\n",
       "      <td>109.88</td>\n",
       "      <td>246.52</td>\n",
       "      <td>2.4</td>\n",
       "      <td>2700.13</td>\n",
       "      <td>49.59</td>\n",
       "      <td>111.36</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1764</th>\n",
       "      <td>2018-01-01</td>\n",
       "      <td>2789.8</td>\n",
       "      <td>49.29</td>\n",
       "      <td>49.29</td>\n",
       "      <td>247.87</td>\n",
       "      <td>2.58</td>\n",
       "      <td>2811.96</td>\n",
       "      <td>49.68</td>\n",
       "      <td>49.68</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1765</th>\n",
       "      <td>2018-02-01</td>\n",
       "      <td>2705.16</td>\n",
       "      <td>49.64</td>\n",
       "      <td>49.64</td>\n",
       "      <td>248.99</td>\n",
       "      <td>2.86</td>\n",
       "      <td>2714.34</td>\n",
       "      <td>49.81</td>\n",
       "      <td>49.81</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1766</th>\n",
       "      <td>2018-03-01</td>\n",
       "      <td>2702.77</td>\n",
       "      <td>50</td>\n",
       "      <td>50</td>\n",
       "      <td>249.55</td>\n",
       "      <td>2.84</td>\n",
       "      <td>2705.82</td>\n",
       "      <td>50.06</td>\n",
       "      <td>50.06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1767</th>\n",
       "      <td>2018-04-01</td>\n",
       "      <td>2642.19</td>\n",
       "      <td>2642.19</td>\n",
       "      <td>2642.19</td>\n",
       "      <td>249.84</td>\n",
       "      <td>2.8</td>\n",
       "      <td>2642.19</td>\n",
       "      <td>2642.19</td>\n",
       "      <td>2642.19</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            Date    SP500 Dividend Earnings Consumer Price Index  \\\n",
       "1763  2017-12-01  2664.34    48.93   109.88               246.52   \n",
       "1764  2018-01-01   2789.8    49.29    49.29               247.87   \n",
       "1765  2018-02-01  2705.16    49.64    49.64               248.99   \n",
       "1766  2018-03-01  2702.77       50       50               249.55   \n",
       "1767  2018-04-01  2642.19  2642.19  2642.19               249.84   \n",
       "\n",
       "     Long Interest Rate Real Price Real Dividend Real Earnings  \n",
       "1763                2.4    2700.13         49.59        111.36  \n",
       "1764               2.58    2811.96         49.68         49.68  \n",
       "1765               2.86    2714.34         49.81         49.81  \n",
       "1766               2.84    2705.82         50.06         50.06  \n",
       "1767                2.8    2642.19       2642.19       2642.19  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## Droping the column\n",
    "sp = sp.drop('PE10', axis=1)\n",
    "## fillna with forward fill\n",
    "sp=sp.fillna(method='ffill', axis=1)\n",
    "sp.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Date                    0\n",
       "SP500                   0\n",
       "Dividend                0\n",
       "Earnings                0\n",
       "Consumer Price Index    0\n",
       "Long Interest Rate      0\n",
       "Real Price              0\n",
       "Real Dividend           0\n",
       "Real Earnings           0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sp.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "12.84 > 12.46 1\n",
      "13.03 > 12.84 1\n",
      "12.56 < 13.03 0\n",
      "12.27 < 12.56 0\n",
      "12.08 < 12.27 0\n",
      "12.08 < 12.08 0\n",
      "11.89 < 12.08 0\n",
      "12.18 > 11.89 1\n",
      "12.37 > 12.18 1\n",
      "12.37 < 12.37 0\n",
      "12.65 > 12.37 1\n",
      "12.65 < 12.65 0\n",
      "12.65 < 12.65 0\n",
      "12.84 > 12.65 1\n",
      "13.13 > 12.84 1\n",
      "13.13 < 13.13 0\n",
      "13.03 < 13.13 0\n",
      "12.84 < 13.03 0\n",
      "12.94 > 12.84 1\n",
      "13.03 > 12.94 1\n",
      "12.75 < 13.03 0\n",
      "13.13 > 12.75 1\n",
      "12.94 < 13.13 0\n",
      "12.94 < 12.94 0\n",
      "13.23 > 12.94 1\n",
      "13.23 < 13.23 0\n",
      "13.23 < 13.23 0\n",
      "12.94 < 13.23 0\n",
      "12.56 < 12.94 0\n",
      "12.56 < 12.56 0\n",
      "12.56 < 12.56 0\n",
      "12.56 < 12.56 0\n",
      "12.27 < 12.56 0\n",
      "11.89 < 12.27 0\n",
      "12.18 > 11.89 1\n",
      "12.37 > 12.18 1\n",
      "12.37 < 12.37 0\n",
      "12.37 < 12.37 0\n",
      "12.18 < 12.37 0\n",
      "12.08 < 12.18 0\n",
      "11.8 < 12.08 0\n",
      "11.89 > 11.8 1\n",
      "11.8 < 11.89 0\n",
      "11.8 < 11.8 0\n",
      "11.61 < 11.8 0\n",
      "11.51 < 11.61 0\n",
      "11.51 < 11.51 0\n",
      "11.51 < 11.51 0\n",
      "11.51 < 11.51 0\n",
      "11.51 < 11.51 0\n",
      "11.61 > 11.51 1\n",
      "11.32 < 11.61 0\n",
      "11.13 < 11.32 0\n",
      "11.13 < 11.13 0\n",
      "11.23 > 11.13 1\n",
      "11.13 < 11.23 0\n",
      "11.13 < 11.13 0\n",
      "11.04 < 11.13 0\n",
      "10.94 < 11.04 0\n",
      "10.85 < 10.94 0\n",
      "10.85 < 10.85 0\n",
      "10.85 < 10.85 0\n",
      "10.75 < 10.85 0\n",
      "10.37 < 10.75 0\n",
      "10.09 < 10.37 0\n",
      "10.09 < 10.09 0\n",
      "10.18 > 10.09 1\n",
      "10.28 > 10.18 1\n",
      "10.47 > 10.28 1\n",
      "10.56 > 10.47 1\n",
      "10.75 > 10.56 1\n",
      "10.94 > 10.75 1\n",
      "10.66 < 10.94 0\n",
      "10.18 < 10.66 0\n",
      "10.47 > 10.18 1\n",
      "10.66 > 10.47 1\n",
      "10.09 < 10.66 0\n",
      "10.18 > 10.09 1\n",
      "9.8 < 10.18 0\n",
      "9.7 < 9.8 0\n",
      "9.7 < 9.7 0\n",
      "9.51 < 9.7 0\n",
      "9.51 < 9.51 0\n",
      "9.23 < 9.51 0\n",
      "9.13 < 9.23 0\n",
      "8.94 < 9.13 0\n",
      "8.85 < 8.94 0\n",
      "8.56 < 8.85 0\n",
      "8.37 < 8.56 0\n",
      "8.47 > 8.37 1\n",
      "8.56 > 8.47 1\n",
      "8.56 < 8.56 0\n",
      "8.47 < 8.56 0\n",
      "8.37 < 8.47 0\n",
      "8.18 < 8.37 0\n",
      "8.28 > 8.18 1\n",
      "8.37 > 8.28 1\n",
      "8.28 < 8.37 0\n",
      "8.18 < 8.28 0\n",
      "8.18 < 8.18 0\n",
      "8.09 < 8.18 0\n",
      "8.18 > 8.09 1\n",
      "8.18 < 8.18 0\n",
      "8.47 > 8.18 1\n",
      "8.94 > 8.47 1\n",
      "9.42 > 8.94 1\n",
      "9.7 > 9.42 1\n",
      "9.99 > 9.7 1\n",
      "9.99 < 9.99 0\n",
      "10.09 > 9.99 1\n",
      "9.7 < 10.09 0\n",
      "9.42 < 9.7 0\n",
      "9.23 < 9.42 0\n",
      "9.23 < 9.23 0\n",
      "9.23 < 9.23 0\n",
      "9.32 > 9.23 1\n",
      "9.32 < 9.32 0\n",
      "9.42 > 9.32 1\n",
      "9.51 > 9.42 1\n",
      "9.42 < 9.51 0\n",
      "9.51 > 9.42 1\n",
      "9.51 < 9.51 0\n",
      "9.61 > 9.51 1\n",
      "9.51 < 9.61 0\n",
      "9.51 < 9.51 0\n",
      "9.61 > 9.51 1\n",
      "9.8 > 9.61 1\n",
      "10.18 > 9.8 1\n",
      "10.28 > 10.18 1\n",
      "10.18 < 10.28 0\n",
      "10.18 < 10.18 0\n",
      "10.18 < 10.18 0\n",
      "10.28 > 10.18 1\n",
      "10.28 < 10.28 0\n",
      "10.37 > 10.28 1\n",
      "10.47 > 10.37 1\n",
      "10.56 > 10.47 1\n",
      "10.47 < 10.56 0\n",
      "10.56 > 10.47 1\n",
      "10.28 < 10.56 0\n",
      "10.18 < 10.28 0\n",
      "10.09 < 10.18 0\n",
      "9.99 < 10.09 0\n",
      "9.99 < 9.99 0\n",
      "10.09 > 9.99 1\n",
      "9.99 < 10.09 0\n",
      "9.9 < 9.99 0\n",
      "9.8 < 9.9 0\n",
      "9.51 < 9.8 0\n",
      "9.32 < 9.51 0\n",
      "9.32 < 9.32 0\n",
      "9.23 < 9.32 0\n",
      "9.23 < 9.23 0\n",
      "9.13 < 9.23 0\n",
      "9.23 > 9.13 1\n",
      "9.23 < 9.23 0\n",
      "9.23 < 9.23 0\n",
      "9.23 < 9.23 0\n",
      "9.04 < 9.23 0\n",
      "8.85 < 9.04 0\n",
      "8.85 < 8.85 0\n",
      "8.75 < 8.85 0\n",
      "8.75 < 8.75 0\n",
      "8.66 < 8.75 0\n",
      "8.56 < 8.66 0\n",
      "8.37 < 8.56 0\n",
      "8.28 < 8.37 0\n",
      "8.28 < 8.28 0\n",
      "8.37 > 8.28 1\n",
      "8.18 < 8.37 0\n",
      "8.28 > 8.18 1\n",
      "8.09 < 8.28 0\n",
      "7.9 < 8.09 0\n",
      "7.99 > 7.9 1\n",
      "7.99 < 7.99 0\n",
      "7.9 < 7.99 0\n",
      "7.9 < 7.9 0\n",
      "7.99 > 7.9 1\n",
      "8.18 > 7.99 1\n",
      "7.99 < 8.18 0\n",
      "7.99 < 7.99 0\n",
      "7.9 < 7.99 0\n",
      "7.8 < 7.9 0\n",
      "7.61 < 7.8 0\n",
      "7.52 < 7.61 0\n",
      "7.61 > 7.52 1\n",
      "7.71 > 7.61 1\n",
      "7.71 < 7.71 0\n",
      "7.71 < 7.71 0\n",
      "7.71 < 7.71 0\n",
      "7.8 > 7.71 1\n",
      "7.99 > 7.8 1\n",
      "8.09 > 7.99 1\n",
      "8.09 < 8.09 0\n",
      "8.09 < 8.09 0\n",
      "8.09 < 8.09 0\n",
      "7.99 < 8.09 0\n",
      "7.9 < 7.99 0\n",
      "7.99 > 7.9 1\n",
      "7.9 < 7.99 0\n",
      "7.99 > 7.9 1\n",
      "8.09 > 7.99 1\n",
      "8.28 > 8.09 1\n",
      "8.37 > 8.28 1\n",
      "8.28 < 8.37 0\n",
      "8.28 < 8.28 0\n",
      "8.18 < 8.28 0\n",
      "8.09 < 8.18 0\n",
      "7.99 < 8.09 0\n",
      "8.09 > 7.99 1\n",
      "8.09 < 8.09 0\n",
      "8.09 < 8.09 0\n",
      "8.18 > 8.09 1\n",
      "8.28 > 8.18 1\n",
      "8.28 < 8.28 0\n",
      "7.99 < 8.28 0\n",
      "7.9 < 7.99 0\n",
      "7.8 < 7.9 0\n",
      "7.8 < 7.8 0\n",
      "7.61 < 7.8 0\n",
      "7.61 < 7.61 0\n",
      "7.61 < 7.61 0\n",
      "7.61 < 7.61 0\n",
      "7.71 > 7.61 1\n",
      "7.71 < 7.71 0\n",
      "7.71 < 7.71 0\n",
      "7.8 > 7.71 1\n",
      "7.61 < 7.8 0\n",
      "7.61 < 7.61 0\n",
      "7.61 < 7.61 0\n",
      "7.61 < 7.61 0\n",
      "7.71 > 7.61 1\n",
      "7.71 < 7.71 0\n",
      "7.71 < 7.71 0\n",
      "7.99 > 7.71 1\n",
      "8.09 > 7.99 1\n",
      "8.09 < 8.09 0\n",
      "7.9 < 8.09 0\n",
      "7.9 < 7.9 0\n",
      "7.8 < 7.9 0\n",
      "7.9 > 7.8 1\n",
      "7.99 > 7.9 1\n",
      "8.09 > 7.99 1\n",
      "7.99 < 8.09 0\n",
      "7.8 < 7.99 0\n",
      "7.71 < 7.8 0\n",
      "7.71 < 7.71 0\n",
      "7.61 < 7.71 0\n",
      "7.61 < 7.61 0\n",
      "7.52 < 7.61 0\n",
      "7.52 < 7.52 0\n",
      "7.33 < 7.52 0\n",
      "7.33 < 7.33 0\n",
      "7.14 < 7.33 0\n",
      "7.04 < 7.14 0\n",
      "7.04 < 7.04 0\n",
      "7.04 < 7.04 0\n",
      "7.23 > 7.04 1\n",
      "7.33 > 7.23 1\n",
      "7.33 < 7.33 0\n",
      "7.33 < 7.33 0\n",
      "7.52 > 7.33 1\n",
      "7.61 > 7.52 1\n",
      "7.9 > 7.61 1\n",
      "7.99 > 7.9 1\n",
      "7.8 < 7.99 0\n",
      "7.71 < 7.8 0\n",
      "7.61 < 7.71 0\n",
      "7.42 < 7.61 0\n",
      "7.23 < 7.42 0\n",
      "6.95 < 7.23 0\n",
      "7.23 > 6.95 1\n",
      "7.33 > 7.23 1\n",
      "7.14 < 7.33 0\n",
      "7.04 < 7.14 0\n",
      "6.85 < 7.04 0\n",
      "6.76 < 6.85 0\n",
      "6.57 < 6.76 0\n",
      "6.57 < 6.57 0\n",
      "6.57 < 6.57 0\n",
      "6.57 < 6.57 0\n",
      "6.57 < 6.57 0\n",
      "6.76 > 6.57 1\n",
      "6.85 > 6.76 1\n",
      "6.66 < 6.85 0\n",
      "6.66 < 6.66 0\n",
      "6.57 < 6.66 0\n",
      "6.57 < 6.57 0\n",
      "6.57 < 6.57 0\n",
      "6.57 < 6.57 0\n",
      "6.85 > 6.57 1\n",
      "6.95 > 6.85 1\n",
      "7.04 > 6.95 1\n",
      "6.95 < 7.04 0\n",
      "6.85 < 6.95 0\n",
      "6.85 < 6.85 0\n",
      "6.85 < 6.85 0\n",
      "6.85 < 6.85 0\n",
      "6.76 < 6.85 0\n",
      "6.66 < 6.76 0\n",
      "6.57 < 6.66 0\n",
      "6.57 < 6.57 0\n",
      "6.47 < 6.57 0\n",
      "6.37 < 6.47 0\n",
      "6.28 < 6.37 0\n",
      "6.28 < 6.28 0\n",
      "6.28 < 6.28 0\n",
      "6.28 < 6.28 0\n",
      "6.47 > 6.28 1\n",
      "6.66 > 6.47 1\n",
      "6.66 < 6.66 0\n",
      "6.47 < 6.66 0\n",
      "6.47 < 6.47 0\n",
      "6.47 < 6.47 0\n",
      "6.37 < 6.47 0\n",
      "6.28 < 6.37 0\n",
      "6.28 < 6.28 0\n",
      "6.28 < 6.28 0\n",
      "6.57 > 6.28 1\n",
      "6.76 > 6.57 1\n",
      "6.66 < 6.76 0\n",
      "6.66 < 6.66 0\n",
      "6.66 < 6.66 0\n",
      "6.66 < 6.66 0\n",
      "6.76 > 6.66 1\n",
      "6.76 < 6.76 0\n",
      "6.76 < 6.76 0\n",
      "7.23 > 6.76 1\n",
      "6.76 < 7.23 0\n",
      "6.66 < 6.76 0\n",
      "6.66 < 6.66 0\n",
      "6.66 < 6.66 0\n",
      "6.66 < 6.66 0\n",
      "6.66 < 6.66 0\n",
      "6.76 > 6.66 1\n",
      "6.76 < 6.76 0\n",
      "6.95 > 6.76 1\n",
      "6.95 < 6.95 0\n",
      "7.04 > 6.95 1\n",
      "7.04 < 7.04 0\n",
      "7.14 > 7.04 1\n",
      "7.23 > 7.14 1\n",
      "7.33 > 7.23 1\n",
      "7.61 > 7.33 1\n",
      "7.71 > 7.61 1\n",
      "7.8 > 7.71 1\n",
      "7.9 > 7.8 1\n",
      "7.9 < 7.9 0\n",
      "7.99 > 7.9 1\n",
      "7.99 < 7.99 0\n",
      "7.99 < 7.99 0\n",
      "7.8 < 7.99 0\n",
      "7.71 < 7.8 0\n",
      "7.8 > 7.71 1\n",
      "7.71 < 7.8 0\n",
      "7.8 > 7.71 1\n",
      "7.71 < 7.8 0\n",
      "7.71 < 7.71 0\n",
      "7.61 < 7.71 0\n",
      "7.71 > 7.61 1\n",
      "7.61 < 7.71 0\n",
      "7.61 < 7.61 0\n",
      "7.52 < 7.61 0\n",
      "7.52 < 7.52 0\n",
      "7.52 < 7.52 0\n",
      "7.61 > 7.52 1\n",
      "7.71 > 7.61 1\n",
      "7.8 > 7.71 1\n",
      "7.8 < 7.8 0\n",
      "7.9 > 7.8 1\n",
      "7.99 > 7.9 1\n",
      "7.9 < 7.99 0\n",
      "7.9 < 7.9 0\n",
      "7.9 < 7.9 0\n",
      "7.99 > 7.9 1\n",
      "8.09 > 7.99 1\n",
      "8.18 > 8.09 1\n",
      "8.18 < 8.18 0\n",
      "8.09 < 8.18 0\n",
      "8.18 > 8.09 1\n",
      "8.75 > 8.18 1\n",
      "8.47 < 8.75 0\n",
      "8.56 > 8.47 1\n",
      "8.66 > 8.56 1\n",
      "8.66 < 8.66 0\n",
      "8.37 < 8.66 0\n",
      "8.37 < 8.37 0\n",
      "8.18 < 8.37 0\n",
      "8.18 < 8.18 0\n",
      "8.18 < 8.18 0\n",
      "8.18 < 8.18 0\n",
      "8.28 > 8.18 1\n",
      "8.18 < 8.28 0\n",
      "8.09 < 8.18 0\n",
      "8.09 < 8.09 0\n",
      "8.28 > 8.09 1\n",
      "8.47 > 8.28 1\n",
      "8.37 < 8.47 0\n",
      "8.28 < 8.37 0\n",
      "8.09 < 8.28 0\n",
      "8.09 < 8.09 0\n",
      "8.09 < 8.09 0\n",
      "8.18 > 8.09 1\n",
      "8.28 > 8.18 1\n",
      "8.28 < 8.28 0\n",
      "8.47 > 8.28 1\n",
      "8.47 < 8.47 0\n",
      "8.47 < 8.47 0\n",
      "8.47 < 8.47 0\n",
      "8.37 < 8.47 0\n",
      "8.37 < 8.37 0\n",
      "8.28 < 8.37 0\n",
      "8.28 < 8.28 0\n",
      "8.28 < 8.28 0\n",
      "8.37 > 8.28 1\n",
      "8.28 < 8.37 0\n",
      "8.28 < 8.28 0\n",
      "8.37 > 8.28 1\n",
      "8.47 > 8.37 1\n",
      "8.47 < 8.47 0\n",
      "8.47 < 8.47 0\n",
      "8.47 < 8.47 0\n",
      "8.47 < 8.47 0\n",
      "8.56 > 8.47 1\n",
      "8.56 < 8.56 0\n",
      "8.28 < 8.56 0\n",
      "8.47 > 8.28 1\n",
      "8.56 > 8.47 1\n",
      "8.75 > 8.56 1\n",
      "8.85 > 8.75 1\n",
      "8.94 > 8.85 1\n",
      "8.85 < 8.94 0\n",
      "9.04 > 8.85 1\n",
      "8.94 < 9.04 0\n",
      "8.94 < 8.94 0\n",
      "9.13 > 8.94 1\n",
      "9.23 > 9.13 1\n",
      "9.23 < 9.23 0\n",
      "9.23 < 9.23 0\n",
      "9.23 < 9.23 0\n",
      "9.32 > 9.23 1\n",
      "8.94 < 9.32 0\n",
      "8.75 < 8.94 0\n",
      "8.66 < 8.75 0\n",
      "8.56 < 8.66 0\n",
      "8.56 < 8.56 0\n",
      "8.66 > 8.56 1\n",
      "8.66 < 8.66 0\n",
      "8.66 < 8.66 0\n",
      "8.75 > 8.66 1\n",
      "8.75 < 8.75 0\n",
      "8.75 < 8.75 0\n",
      "8.85 > 8.75 1\n",
      "8.94 > 8.85 1\n",
      "9.04 > 8.94 1\n",
      "8.94 < 9.04 0\n",
      "9.04 > 8.94 1\n",
      "9.04 < 9.04 0\n",
      "9.23 > 9.04 1\n",
      "9.32 > 9.23 1\n",
      "9.42 > 9.32 1\n",
      "9.42 < 9.42 0\n",
      "9.51 > 9.42 1\n",
      "9.61 > 9.51 1\n",
      "9.8 > 9.61 1\n",
      "9.9 > 9.8 1\n",
      "9.99 > 9.9 1\n",
      "9.9 < 9.99 0\n",
      "9.9 < 9.9 0\n",
      "10.09 > 9.9 1\n",
      "10.18 > 10.09 1\n",
      "9.99 < 10.18 0\n",
      "9.9 < 9.99 0\n",
      "9.9 < 9.9 0\n",
      "9.8 < 9.9 0\n",
      "9.7 < 9.8 0\n",
      "9.42 < 9.7 0\n",
      "9.23 < 9.42 0\n",
      "9.23 < 9.23 0\n",
      "9.23 < 9.23 0\n",
      "8.94 < 9.23 0\n",
      "9.04 > 8.94 1\n",
      "8.75 < 9.04 0\n",
      "8.75 < 8.75 0\n",
      "8.75 < 8.75 0\n",
      "8.85 > 8.75 1\n",
      "9.13 > 8.85 1\n",
      "9.23 > 9.13 1\n",
      "9.23 < 9.23 0\n",
      "9.13 < 9.23 0\n",
      "9.04 < 9.13 0\n",
      "9.13 > 9.04 1\n",
      "9.23 > 9.13 1\n",
      "9.42 > 9.23 1\n",
      "9.7 > 9.42 1\n",
      "9.7 < 9.7 0\n",
      "9.61 < 9.7 0\n",
      "9.61 < 9.61 0\n",
      "9.7 > 9.61 1\n",
      "9.8 > 9.7 1\n",
      "9.8 < 9.8 0\n",
      "9.8 < 9.8 0\n",
      "9.7 < 9.8 0\n",
      "9.8 > 9.7 1\n",
      "9.8 < 9.8 0\n",
      "9.8 < 9.8 0\n",
      "9.8 < 9.8 0\n",
      "9.7 < 9.8 0\n",
      "9.8 > 9.7 1\n",
      "9.9 > 9.8 1\n",
      "9.9 < 9.9 0\n",
      "10.0 > 9.9 1\n",
      "10.0 < 10.0 0\n",
      "10.1 > 10.0 1\n",
      "10.0 < 10.1 0\n",
      "10.0 < 10.0 0\n",
      "9.9 < 10.0 0\n",
      "9.9 < 9.9 0\n",
      "9.8 < 9.9 0\n",
      "9.9 > 9.8 1\n",
      "9.9 < 9.9 0\n",
      "10.0 > 9.9 1\n",
      "10.2 > 10.0 1\n",
      "10.2 < 10.2 0\n",
      "10.1 < 10.2 0\n",
      "10.2 > 10.1 1\n",
      "10.1 < 10.2 0\n",
      "10.1 < 10.1 0\n",
      "10.0 < 10.1 0\n",
      "9.9 < 10.0 0\n",
      "10.0 > 9.9 1\n",
      "10.1 > 10.0 1\n",
      "10.1 < 10.1 0\n",
      "10.1 < 10.1 0\n",
      "10.1 < 10.1 0\n",
      "10.1 < 10.1 0\n",
      "10.2 > 10.1 1\n",
      "10.3 > 10.2 1\n",
      "10.3 < 10.3 0\n",
      "10.4 > 10.3 1\n",
      "10.4 < 10.4 0\n",
      "10.5 > 10.4 1\n",
      "10.6 > 10.5 1\n",
      "10.7 > 10.6 1\n",
      "10.8 > 10.7 1\n",
      "10.8 < 10.8 0\n",
      "10.9 > 10.8 1\n",
      "11.1 > 10.9 1\n",
      "11.3 > 11.1 1\n",
      "11.5 > 11.3 1\n",
      "11.6 > 11.5 1\n",
      "11.7 > 11.6 1\n",
      "12.0 > 11.7 1\n",
      "12.0 < 12.0 0\n",
      "12.6 > 12.0 1\n",
      "12.8 > 12.6 1\n",
      "13.0 > 12.8 1\n",
      "12.8 < 13.0 0\n",
      "13.0 > 12.8 1\n",
      "13.3 > 13.0 1\n",
      "13.5 > 13.3 1\n",
      "13.5 < 13.5 0\n",
      "13.7 > 13.5 1\n",
      "14.0 > 13.7 1\n",
      "14.1 > 14.0 1\n",
      "14.0 < 14.1 0\n",
      "14.2 > 14.0 1\n",
      "14.5 > 14.2 1\n",
      "14.7 > 14.5 1\n",
      "15.1 > 14.7 1\n",
      "15.4 > 15.1 1\n",
      "15.7 > 15.4 1\n",
      "16.0 > 15.7 1\n",
      "16.3 > 16.0 1\n",
      "16.5 > 16.3 1\n",
      "16.5 < 16.5 0\n",
      "16.2 < 16.5 0\n",
      "16.4 > 16.2 1\n",
      "16.7 > 16.4 1\n",
      "16.9 > 16.7 1\n",
      "16.9 < 16.9 0\n",
      "17.4 > 16.9 1\n",
      "17.7 > 17.4 1\n",
      "17.8 > 17.7 1\n",
      "18.1 > 17.8 1\n",
      "18.5 > 18.1 1\n",
      "18.9 > 18.5 1\n",
      "19.3 > 18.9 1\n",
      "19.5 > 19.3 1\n",
      "19.7 > 19.5 1\n",
      "20.3 > 19.7 1\n",
      "20.6 > 20.3 1\n",
      "20.9 > 20.6 1\n",
      "20.8 < 20.9 0\n",
      "20.3 < 20.8 0\n",
      "20.0 < 20.3 0\n",
      "19.9 < 20.0 0\n",
      "19.8 < 19.9 0\n",
      "19.4 < 19.8 0\n",
      "19.0 < 19.4 0\n",
      "18.4 < 19.0 0\n",
      "18.3 < 18.4 0\n",
      "18.1 < 18.3 0\n",
      "17.7 < 18.1 0\n",
      "17.6 < 17.7 0\n",
      "17.7 > 17.6 1\n",
      "17.7 < 17.7 0\n",
      "17.5 < 17.7 0\n",
      "17.5 < 17.5 0\n",
      "17.4 < 17.5 0\n",
      "17.3 < 17.4 0\n",
      "16.9 < 17.3 0\n",
      "16.9 < 16.9 0\n",
      "16.7 < 16.9 0\n",
      "16.7 < 16.7 0\n",
      "16.7 < 16.7 0\n",
      "16.7 < 16.7 0\n",
      "16.8 > 16.7 1\n",
      "16.6 < 16.8 0\n",
      "16.6 < 16.6 0\n",
      "16.7 > 16.6 1\n",
      "16.8 > 16.7 1\n",
      "16.9 > 16.8 1\n",
      "16.8 < 16.9 0\n",
      "16.8 < 16.8 0\n",
      "16.8 < 16.8 0\n",
      "16.9 > 16.8 1\n",
      "16.9 < 16.9 0\n",
      "17.0 > 16.9 1\n",
      "17.2 > 17.0 1\n",
      "17.1 < 17.2 0\n",
      "17.2 > 17.1 1\n",
      "17.3 > 17.2 1\n",
      "17.3 < 17.3 0\n",
      "17.3 < 17.3 0\n",
      "17.3 < 17.3 0\n",
      "17.2 < 17.3 0\n",
      "17.1 < 17.2 0\n",
      "17.0 < 17.1 0\n",
      "17.0 < 17.0 0\n",
      "17.0 < 17.0 0\n",
      "17.1 > 17.0 1\n",
      "17.0 < 17.1 0\n",
      "17.1 > 17.0 1\n",
      "17.2 > 17.1 1\n",
      "17.2 < 17.2 0\n",
      "17.3 > 17.2 1\n",
      "17.3 < 17.3 0\n",
      "17.2 < 17.3 0\n",
      "17.3 > 17.2 1\n",
      "17.2 < 17.3 0\n",
      "17.3 > 17.2 1\n",
      "17.5 > 17.3 1\n",
      "17.7 > 17.5 1\n",
      "17.7 < 17.7 0\n",
      "17.7 < 17.7 0\n",
      "17.7 < 17.7 0\n",
      "18.0 > 17.7 1\n",
      "17.9 < 18.0 0\n",
      "17.9 < 17.9 0\n",
      "17.9 < 17.9 0\n",
      "17.8 < 17.9 0\n",
      "17.9 > 17.8 1\n",
      "17.8 < 17.9 0\n",
      "17.7 < 17.8 0\n",
      "17.5 < 17.7 0\n",
      "17.4 < 17.5 0\n",
      "17.5 > 17.4 1\n",
      "17.6 > 17.5 1\n",
      "17.7 > 17.6 1\n",
      "17.7 < 17.7 0\n",
      "17.5 < 17.7 0\n",
      "17.4 < 17.5 0\n",
      "17.3 < 17.4 0\n",
      "17.3 < 17.3 0\n",
      "17.4 > 17.3 1\n",
      "17.6 > 17.4 1\n",
      "17.3 < 17.6 0\n",
      "17.2 < 17.3 0\n",
      "17.3 > 17.2 1\n",
      "17.4 > 17.3 1\n",
      "17.3 < 17.4 0\n",
      "17.3 < 17.3 0\n",
      "17.3 < 17.3 0\n",
      "17.1 < 17.3 0\n",
      "17.1 < 17.1 0\n",
      "17.1 < 17.1 0\n",
      "17.2 > 17.1 1\n",
      "17.1 < 17.2 0\n",
      "17.1 < 17.1 0\n",
      "17.1 < 17.1 0\n",
      "17.3 > 17.1 1\n",
      "17.2 < 17.3 0\n",
      "17.2 < 17.2 0\n",
      "17.1 < 17.2 0\n",
      "17.1 < 17.1 0\n",
      "17.1 < 17.1 0\n",
      "17.0 < 17.1 0\n",
      "16.9 < 17.0 0\n",
      "17.0 > 16.9 1\n",
      "17.1 > 17.0 1\n",
      "17.3 > 17.1 1\n",
      "17.3 < 17.3 0\n",
      "17.3 < 17.3 0\n",
      "17.3 < 17.3 0\n",
      "17.3 < 17.3 0\n",
      "17.2 < 17.3 0\n",
      "17.1 < 17.2 0\n",
      "17.0 < 17.1 0\n",
      "16.9 < 17.0 0\n",
      "17.0 > 16.9 1\n",
      "16.9 < 17.0 0\n",
      "16.8 < 16.9 0\n",
      "16.6 < 16.8 0\n",
      "16.5 < 16.6 0\n",
      "16.6 > 16.5 1\n",
      "16.5 < 16.6 0\n",
      "16.4 < 16.5 0\n",
      "16.1 < 16.4 0\n",
      "15.9 < 16.1 0\n",
      "15.7 < 15.9 0\n",
      "15.6 < 15.7 0\n",
      "15.5 < 15.6 0\n",
      "15.3 < 15.5 0\n",
      "15.1 < 15.3 0\n",
      "15.1 < 15.1 0\n",
      "15.1 < 15.1 0\n",
      "15.0 < 15.1 0\n",
      "14.9 < 15.0 0\n",
      "14.7 < 14.9 0\n",
      "14.6 < 14.7 0\n",
      "14.3 < 14.6 0\n",
      "14.1 < 14.3 0\n",
      "14.0 < 14.1 0\n",
      "13.9 < 14.0 0\n",
      "13.7 < 13.9 0\n",
      "13.6 < 13.7 0\n",
      "13.6 < 13.6 0\n",
      "13.5 < 13.6 0\n",
      "13.4 < 13.5 0\n",
      "13.3 < 13.4 0\n",
      "13.2 < 13.3 0\n",
      "13.1 < 13.2 0\n",
      "12.9 < 13.1 0\n",
      "12.7 < 12.9 0\n",
      "12.6 < 12.7 0\n",
      "12.6 < 12.6 0\n",
      "12.6 < 12.6 0\n",
      "12.7 > 12.6 1\n",
      "13.1 > 12.7 1\n",
      "13.2 > 13.1 1\n",
      "13.2 < 13.2 0\n",
      "13.2 < 13.2 0\n",
      "13.2 < 13.2 0\n",
      "13.2 < 13.2 0\n",
      "13.2 < 13.2 0\n",
      "13.3 > 13.2 1\n",
      "13.3 < 13.3 0\n",
      "13.3 < 13.3 0\n",
      "13.3 < 13.3 0\n",
      "13.4 > 13.3 1\n",
      "13.4 < 13.4 0\n",
      "13.4 < 13.4 0\n",
      "13.6 > 13.4 1\n",
      "13.5 < 13.6 0\n",
      "13.5 < 13.5 0\n",
      "13.4 < 13.5 0\n",
      "13.6 > 13.4 1\n",
      "13.7 > 13.6 1\n",
      "13.7 < 13.7 0\n",
      "13.8 > 13.7 1\n",
      "13.8 < 13.8 0\n",
      "13.7 < 13.8 0\n",
      "13.7 < 13.7 0\n",
      "13.7 < 13.7 0\n",
      "13.7 < 13.7 0\n",
      "13.7 < 13.7 0\n",
      "13.8 > 13.7 1\n",
      "13.8 < 13.8 0\n",
      "13.8 < 13.8 0\n",
      "13.8 < 13.8 0\n",
      "13.7 < 13.8 0\n",
      "13.7 < 13.7 0\n",
      "13.7 < 13.7 0\n",
      "13.8 > 13.7 1\n",
      "13.9 > 13.8 1\n",
      "14.0 > 13.9 1\n",
      "14.0 < 14.0 0\n",
      "14.0 < 14.0 0\n",
      "14.0 < 14.0 0\n",
      "14.0 < 14.0 0\n",
      "14.1 > 14.0 1\n",
      "14.1 < 14.1 0\n",
      "14.2 > 14.1 1\n",
      "14.3 > 14.2 1\n",
      "14.4 > 14.3 1\n",
      "14.4 < 14.4 0\n",
      "14.5 > 14.4 1\n",
      "14.5 < 14.5 0\n",
      "14.6 > 14.5 1\n",
      "14.6 < 14.6 0\n",
      "14.5 < 14.6 0\n",
      "14.4 < 14.5 0\n",
      "14.2 < 14.4 0\n",
      "14.1 < 14.2 0\n",
      "14.1 < 14.1 0\n",
      "14.2 > 14.1 1\n",
      "14.1 < 14.2 0\n",
      "14.1 < 14.1 0\n",
      "14.1 < 14.1 0\n",
      "14.1 < 14.1 0\n",
      "14.1 < 14.1 0\n",
      "14.0 < 14.1 0\n",
      "14.0 < 14.0 0\n",
      "14.0 < 14.0 0\n",
      "14.0 < 14.0 0\n",
      "13.9 < 14.0 0\n",
      "13.9 < 13.9 0\n",
      "13.8 < 13.9 0\n",
      "13.8 < 13.8 0\n",
      "13.8 < 13.8 0\n",
      "13.8 < 13.8 0\n",
      "13.8 < 13.8 0\n",
      "14.1 > 13.8 1\n",
      "14.0 < 14.1 0\n",
      "14.0 < 14.0 0\n",
      "14.0 < 14.0 0\n",
      "13.9 < 14.0 0\n",
      "14.0 > 13.9 1\n",
      "14.0 < 14.0 0\n",
      "14.0 < 14.0 0\n",
      "14.0 < 14.0 0\n",
      "14.1 > 14.0 1\n",
      "14.0 < 14.1 0\n",
      "14.0 < 14.0 0\n",
      "14.0 < 14.0 0\n",
      "14.0 < 14.0 0\n",
      "14.0 < 14.0 0\n",
      "14.1 > 14.0 1\n",
      "14.1 < 14.1 0\n",
      "14.1 < 14.1 0\n",
      "14.2 > 14.1 1\n",
      "14.3 > 14.2 1\n",
      "14.4 > 14.3 1\n",
      "14.7 > 14.4 1\n",
      "14.7 < 14.7 0\n",
      "14.9 > 14.7 1\n",
      "15.1 > 14.9 1\n",
      "15.3 > 15.1 1\n",
      "15.4 > 15.3 1\n",
      "15.5 > 15.4 1\n",
      "15.7 > 15.5 1\n",
      "15.8 > 15.7 1\n",
      "16.0 > 15.8 1\n",
      "16.1 > 16.0 1\n",
      "16.3 > 16.1 1\n",
      "16.3 < 16.3 0\n",
      "16.4 > 16.3 1\n",
      "16.5 > 16.4 1\n",
      "16.5 < 16.5 0\n",
      "16.7 > 16.5 1\n",
      "16.8 > 16.7 1\n",
      "16.9 > 16.8 1\n",
      "16.9 < 16.9 0\n",
      "16.9 < 16.9 0\n",
      "17.2 > 16.9 1\n",
      "17.4 > 17.2 1\n",
      "17.5 > 17.4 1\n",
      "17.5 < 17.5 0\n",
      "17.4 < 17.5 0\n",
      "17.3 < 17.4 0\n",
      "17.4 > 17.3 1\n",
      "17.4 < 17.4 0\n",
      "17.4 < 17.4 0\n",
      "17.4 < 17.4 0\n",
      "17.4 < 17.4 0\n",
      "17.4 < 17.4 0\n",
      "17.4 < 17.4 0\n",
      "17.5 > 17.4 1\n",
      "17.5 < 17.5 0\n",
      "17.6 > 17.5 1\n",
      "17.7 > 17.6 1\n",
      "17.7 < 17.7 0\n",
      "17.7 < 17.7 0\n",
      "17.7 < 17.7 0\n",
      "17.7 < 17.7 0\n",
      "17.8 > 17.7 1\n",
      "17.8 < 17.8 0\n",
      "17.8 < 17.8 0\n",
      "17.8 < 17.8 0\n",
      "17.8 < 17.8 0\n",
      "17.9 > 17.8 1\n",
      "18.1 > 17.9 1\n",
      "18.1 < 18.1 0\n",
      "18.1 < 18.1 0\n",
      "18.1 < 18.1 0\n",
      "18.1 < 18.1 0\n",
      "18.1 < 18.1 0\n",
      "18.2 > 18.1 1\n",
      "18.2 < 18.2 0\n",
      "18.1 < 18.2 0\n",
      "18.3 > 18.1 1\n",
      "18.4 > 18.3 1\n",
      "18.5 > 18.4 1\n",
      "18.7 > 18.5 1\n",
      "19.8 > 18.7 1\n",
      "20.2 > 19.8 1\n",
      "20.4 > 20.2 1\n",
      "20.8 > 20.4 1\n",
      "21.3 > 20.8 1\n",
      "21.5 > 21.3 1\n",
      "21.5 < 21.5 0\n",
      "21.5 < 21.5 0\n",
      "21.9 > 21.5 1\n",
      "21.9 < 21.9 0\n",
      "21.9 < 21.9 0\n",
      "22.0 > 21.9 1\n",
      "22.2 > 22.0 1\n",
      "22.5 > 22.2 1\n",
      "23.0 > 22.5 1\n",
      "23.0 < 23.0 0\n",
      "23.1 > 23.0 1\n",
      "23.4 > 23.1 1\n",
      "23.7 > 23.4 1\n",
      "23.5 < 23.7 0\n",
      "23.4 < 23.5 0\n",
      "23.8 > 23.4 1\n",
      "23.9 > 23.8 1\n",
      "24.1 > 23.9 1\n",
      "24.4 > 24.1 1\n",
      "24.5 > 24.4 1\n",
      "24.5 < 24.5 0\n",
      "24.4 < 24.5 0\n",
      "24.2 < 24.4 0\n",
      "24.1 < 24.2 0\n",
      "24.0 < 24.1 0\n",
      "23.8 < 24.0 0\n",
      "23.8 < 23.8 0\n",
      "23.9 > 23.8 1\n",
      "23.8 < 23.9 0\n",
      "23.9 > 23.8 1\n",
      "23.7 < 23.9 0\n",
      "23.8 > 23.7 1\n",
      "23.9 > 23.8 1\n",
      "23.7 < 23.9 0\n",
      "23.8 > 23.7 1\n",
      "23.6 < 23.8 0\n",
      "23.5 < 23.6 0\n",
      "23.5 < 23.5 0\n",
      "23.6 > 23.5 1\n",
      "23.6 < 23.6 0\n",
      "23.7 > 23.6 1\n",
      "23.8 > 23.7 1\n",
      "24.1 > 23.8 1\n",
      "24.3 > 24.1 1\n",
      "24.4 > 24.3 1\n",
      "24.6 > 24.4 1\n",
      "24.7 > 24.6 1\n",
      "25.0 > 24.7 1\n",
      "25.4 > 25.0 1\n",
      "25.7 > 25.4 1\n",
      "25.8 > 25.7 1\n",
      "25.8 < 25.8 0\n",
      "25.9 > 25.8 1\n",
      "25.9 < 25.9 0\n",
      "25.9 < 25.9 0\n",
      "25.9 < 25.9 0\n",
      "26.1 > 25.9 1\n",
      "26.2 > 26.1 1\n",
      "26.4 > 26.2 1\n",
      "26.5 > 26.4 1\n",
      "26.5 < 26.5 0\n",
      "26.3 < 26.5 0\n",
      "26.3 < 26.3 0\n",
      "26.4 > 26.3 1\n",
      "26.4 < 26.4 0\n",
      "26.5 > 26.4 1\n",
      "26.7 > 26.5 1\n",
      "26.7 < 26.7 0\n",
      "26.7 < 26.7 0\n",
      "26.7 < 26.7 0\n",
      "26.7 < 26.7 0\n",
      "26.7 < 26.7 0\n",
      "26.6 < 26.7 0\n",
      "26.5 < 26.6 0\n",
      "26.6 > 26.5 1\n",
      "26.6 < 26.6 0\n",
      "26.7 > 26.6 1\n",
      "26.8 > 26.7 1\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "26.8 < 26.8 0\n",
      "26.9 > 26.8 1\n",
      "26.9 < 26.9 0\n",
      "27.0 > 26.9 1\n",
      "26.9 < 27.0 0\n",
      "26.9 < 26.9 0\n",
      "26.9 < 26.9 0\n",
      "26.9 < 26.9 0\n",
      "26.9 < 26.9 0\n",
      "26.8 < 26.9 0\n",
      "26.9 > 26.8 1\n",
      "26.9 < 26.9 0\n",
      "26.9 < 26.9 0\n",
      "26.9 < 26.9 0\n",
      "26.8 < 26.9 0\n",
      "26.8 < 26.8 0\n",
      "26.8 < 26.8 0\n",
      "26.7 < 26.8 0\n",
      "26.7 < 26.7 0\n",
      "26.7 < 26.7 0\n",
      "26.7 < 26.7 0\n",
      "26.7 < 26.7 0\n",
      "26.7 < 26.7 0\n",
      "26.7 < 26.7 0\n",
      "26.8 > 26.7 1\n",
      "26.8 < 26.8 0\n",
      "26.9 > 26.8 1\n",
      "26.9 < 26.9 0\n",
      "26.9 < 26.9 0\n",
      "26.8 < 26.9 0\n",
      "26.8 < 26.8 0\n",
      "26.8 < 26.8 0\n",
      "26.8 < 26.8 0\n",
      "26.9 > 26.8 1\n",
      "27.0 > 26.9 1\n",
      "27.2 > 27.0 1\n",
      "27.4 > 27.2 1\n",
      "27.3 < 27.4 0\n",
      "27.4 > 27.3 1\n",
      "27.5 > 27.4 1\n",
      "27.5 < 27.5 0\n",
      "27.6 > 27.5 1\n",
      "27.6 < 27.6 0\n",
      "27.7 > 27.6 1\n",
      "27.8 > 27.7 1\n",
      "27.9 > 27.8 1\n",
      "28.0 > 27.9 1\n",
      "28.1 > 28.0 1\n",
      "28.3 > 28.1 1\n",
      "28.3 < 28.3 0\n",
      "28.3 < 28.3 0\n",
      "28.3 < 28.3 0\n",
      "28.4 > 28.3 1\n",
      "28.4 < 28.4 0\n",
      "28.6 > 28.4 1\n",
      "28.6 < 28.6 0\n",
      "28.8 > 28.6 1\n",
      "28.9 > 28.8 1\n",
      "28.9 < 28.9 0\n",
      "28.9 < 28.9 0\n",
      "29.0 > 28.9 1\n",
      "28.9 < 29.0 0\n",
      "28.9 < 28.9 0\n",
      "28.9 < 28.9 0\n",
      "29.0 > 28.9 1\n",
      "28.9 < 29.0 0\n",
      "29.0 > 28.9 1\n",
      "28.9 < 29.0 0\n",
      "28.9 < 28.9 0\n",
      "29.0 > 28.9 1\n",
      "29.0 < 29.0 0\n",
      "29.1 > 29.0 1\n",
      "29.2 > 29.1 1\n",
      "29.2 < 29.2 0\n",
      "29.3 > 29.2 1\n",
      "29.4 > 29.3 1\n",
      "29.4 < 29.4 0\n",
      "29.4 < 29.4 0\n",
      "29.3 < 29.4 0\n",
      "29.4 > 29.3 1\n",
      "29.4 < 29.4 0\n",
      "29.5 > 29.4 1\n",
      "29.5 < 29.5 0\n",
      "29.6 > 29.5 1\n",
      "29.6 < 29.6 0\n",
      "29.6 < 29.6 0\n",
      "29.6 < 29.6 0\n",
      "29.8 > 29.6 1\n",
      "29.8 < 29.8 0\n",
      "29.8 < 29.8 0\n",
      "29.8 < 29.8 0\n",
      "29.8 < 29.8 0\n",
      "29.8 < 29.8 0\n",
      "29.8 < 29.8 0\n",
      "29.8 < 29.8 0\n",
      "29.8 < 29.8 0\n",
      "30.0 > 29.8 1\n",
      "29.9 < 30.0 0\n",
      "30.0 > 29.9 1\n",
      "30.0 < 30.0 0\n",
      "30.0 < 30.0 0\n",
      "30.0 < 30.0 0\n",
      "30.0 < 30.0 0\n",
      "30.1 > 30.0 1\n",
      "30.1 < 30.1 0\n",
      "30.2 > 30.1 1\n",
      "30.2 < 30.2 0\n",
      "30.2 < 30.2 0\n",
      "30.3 > 30.2 1\n",
      "30.3 < 30.3 0\n",
      "30.4 > 30.3 1\n",
      "30.4 < 30.4 0\n",
      "30.4 < 30.4 0\n",
      "30.4 < 30.4 0\n",
      "30.4 < 30.4 0\n",
      "30.4 < 30.4 0\n",
      "30.5 > 30.4 1\n",
      "30.5 < 30.5 0\n",
      "30.5 < 30.5 0\n",
      "30.6 > 30.5 1\n",
      "30.7 > 30.6 1\n",
      "30.7 < 30.7 0\n",
      "30.7 < 30.7 0\n",
      "30.8 > 30.7 1\n",
      "30.8 < 30.8 0\n",
      "30.9 > 30.8 1\n",
      "30.9 < 30.9 0\n",
      "30.9 < 30.9 0\n",
      "30.9 < 30.9 0\n",
      "30.9 < 30.9 0\n",
      "30.9 < 30.9 0\n",
      "31.0 > 30.9 1\n",
      "31.1 > 31.0 1\n",
      "31.0 < 31.1 0\n",
      "31.1 > 31.0 1\n",
      "31.1 < 31.1 0\n",
      "31.2 > 31.1 1\n",
      "31.2 < 31.2 0\n",
      "31.2 < 31.2 0\n",
      "31.2 < 31.2 0\n",
      "31.3 > 31.2 1\n",
      "31.4 > 31.3 1\n",
      "31.4 < 31.4 0\n",
      "31.6 > 31.4 1\n",
      "31.6 < 31.6 0\n",
      "31.6 < 31.6 0\n",
      "31.6 < 31.6 0\n",
      "31.7 > 31.6 1\n",
      "31.7 < 31.7 0\n",
      "31.8 > 31.7 1\n",
      "31.8 < 31.8 0\n",
      "32.0 > 31.8 1\n",
      "32.1 > 32.0 1\n",
      "32.3 > 32.1 1\n",
      "32.3 < 32.3 0\n",
      "32.4 > 32.3 1\n",
      "32.5 > 32.4 1\n",
      "32.7 > 32.5 1\n",
      "32.7 < 32.7 0\n",
      "32.9 > 32.7 1\n",
      "32.9 < 32.9 0\n",
      "32.9 < 32.9 0\n",
      "32.9 < 32.9 0\n",
      "32.9 < 32.9 0\n",
      "33.0 > 32.9 1\n",
      "33.1 > 33.0 1\n",
      "33.2 > 33.1 1\n",
      "33.3 > 33.2 1\n",
      "33.4 > 33.3 1\n",
      "33.5 > 33.4 1\n",
      "33.6 > 33.5 1\n",
      "33.7 > 33.6 1\n",
      "33.8 > 33.7 1\n",
      "33.9 > 33.8 1\n",
      "34.1 > 33.9 1\n",
      "34.2 > 34.1 1\n",
      "34.3 > 34.2 1\n",
      "34.4 > 34.3 1\n",
      "34.5 > 34.4 1\n",
      "34.7 > 34.5 1\n",
      "34.9 > 34.7 1\n",
      "35.0 > 34.9 1\n",
      "35.1 > 35.0 1\n",
      "35.3 > 35.1 1\n",
      "35.4 > 35.3 1\n",
      "35.5 > 35.4 1\n",
      "35.6 > 35.5 1\n",
      "35.8 > 35.6 1\n",
      "36.1 > 35.8 1\n",
      "36.3 > 36.1 1\n",
      "36.4 > 36.3 1\n",
      "36.6 > 36.4 1\n",
      "36.8 > 36.6 1\n",
      "37.0 > 36.8 1\n",
      "37.1 > 37.0 1\n",
      "37.3 > 37.1 1\n",
      "37.5 > 37.3 1\n",
      "37.7 > 37.5 1\n",
      "37.8 > 37.7 1\n",
      "38.0 > 37.8 1\n",
      "38.2 > 38.0 1\n",
      "38.5 > 38.2 1\n",
      "38.6 > 38.5 1\n",
      "38.8 > 38.6 1\n",
      "39.0 > 38.8 1\n",
      "39.0 < 39.0 0\n",
      "39.2 > 39.0 1\n",
      "39.4 > 39.2 1\n",
      "39.6 > 39.4 1\n",
      "39.8 > 39.6 1\n",
      "39.8 < 39.8 0\n",
      "39.9 > 39.8 1\n",
      "40.0 > 39.9 1\n",
      "40.1 > 40.0 1\n",
      "40.3 > 40.1 1\n",
      "40.6 > 40.3 1\n",
      "40.7 > 40.6 1\n",
      "40.8 > 40.7 1\n",
      "40.8 < 40.8 0\n",
      "40.9 > 40.8 1\n",
      "40.9 < 40.9 0\n",
      "41.1 > 40.9 1\n",
      "41.1 < 41.1 0\n",
      "41.3 > 41.1 1\n",
      "41.4 > 41.3 1\n",
      "41.5 > 41.4 1\n",
      "41.6 > 41.5 1\n",
      "41.7 > 41.6 1\n",
      "41.9 > 41.7 1\n",
      "42.0 > 41.9 1\n",
      "42.1 > 42.0 1\n",
      "42.3 > 42.1 1\n",
      "42.4 > 42.3 1\n",
      "42.5 > 42.4 1\n",
      "42.6 > 42.5 1\n",
      "42.9 > 42.6 1\n",
      "43.3 > 42.9 1\n",
      "43.6 > 43.3 1\n",
      "43.9 > 43.6 1\n",
      "44.2 > 43.9 1\n",
      "44.3 > 44.2 1\n",
      "45.1 > 44.3 1\n",
      "45.2 > 45.1 1\n",
      "45.6 > 45.2 1\n",
      "45.9 > 45.6 1\n",
      "46.2 > 45.9 1\n",
      "46.6 > 46.2 1\n",
      "47.2 > 46.6 1\n",
      "47.8 > 47.2 1\n",
      "48.0 > 47.8 1\n",
      "48.6 > 48.0 1\n",
      "49.0 > 48.6 1\n",
      "49.4 > 49.0 1\n",
      "50.0 > 49.4 1\n",
      "50.6 > 50.0 1\n",
      "51.1 > 50.6 1\n",
      "51.5 > 51.1 1\n",
      "51.9 > 51.5 1\n",
      "52.1 > 51.9 1\n",
      "52.5 > 52.1 1\n",
      "52.7 > 52.5 1\n",
      "52.9 > 52.7 1\n",
      "53.2 > 52.9 1\n",
      "53.6 > 53.2 1\n",
      "54.2 > 53.6 1\n",
      "54.3 > 54.2 1\n",
      "54.6 > 54.3 1\n",
      "54.9 > 54.6 1\n",
      "55.3 > 54.9 1\n",
      "55.5 > 55.3 1\n",
      "55.6 > 55.5 1\n",
      "55.8 > 55.6 1\n",
      "55.9 > 55.8 1\n",
      "56.1 > 55.9 1\n",
      "56.5 > 56.1 1\n",
      "56.8 > 56.5 1\n",
      "57.1 > 56.8 1\n",
      "57.4 > 57.1 1\n",
      "57.6 > 57.4 1\n",
      "57.9 > 57.6 1\n",
      "58.0 > 57.9 1\n",
      "58.2 > 58.0 1\n",
      "58.5 > 58.2 1\n",
      "59.1 > 58.5 1\n",
      "59.5 > 59.1 1\n",
      "60.0 > 59.5 1\n",
      "60.3 > 60.0 1\n",
      "60.7 > 60.3 1\n",
      "61.0 > 60.7 1\n",
      "61.2 > 61.0 1\n",
      "61.4 > 61.2 1\n",
      "61.6 > 61.4 1\n",
      "61.9 > 61.6 1\n",
      "62.1 > 61.9 1\n",
      "62.5 > 62.1 1\n",
      "62.9 > 62.5 1\n",
      "63.4 > 62.9 1\n",
      "63.9 > 63.4 1\n",
      "64.5 > 63.9 1\n",
      "65.2 > 64.5 1\n",
      "65.7 > 65.2 1\n",
      "66.0 > 65.7 1\n",
      "66.5 > 66.0 1\n",
      "67.1 > 66.5 1\n",
      "67.4 > 67.1 1\n",
      "67.7 > 67.4 1\n",
      "68.3 > 67.7 1\n",
      "69.1 > 68.3 1\n",
      "69.8 > 69.1 1\n",
      "70.6 > 69.8 1\n",
      "71.5 > 70.6 1\n",
      "72.3 > 71.5 1\n",
      "73.1 > 72.3 1\n",
      "73.8 > 73.1 1\n",
      "74.6 > 73.8 1\n",
      "75.2 > 74.6 1\n",
      "75.9 > 75.2 1\n",
      "76.7 > 75.9 1\n",
      "77.8 > 76.7 1\n",
      "78.9 > 77.8 1\n",
      "80.1 > 78.9 1\n",
      "81.0 > 80.1 1\n",
      "81.8 > 81.0 1\n",
      "82.7 > 81.8 1\n",
      "82.7 < 82.7 0\n",
      "83.3 > 82.7 1\n",
      "84.0 > 83.3 1\n",
      "84.8 > 84.0 1\n",
      "85.5 > 84.8 1\n",
      "86.3 > 85.5 1\n",
      "87.0 > 86.3 1\n",
      "87.9 > 87.0 1\n",
      "88.5 > 87.9 1\n",
      "89.1 > 88.5 1\n",
      "89.8 > 89.1 1\n",
      "90.6 > 89.8 1\n",
      "91.6 > 90.6 1\n",
      "92.3 > 91.6 1\n",
      "93.2 > 92.3 1\n",
      "93.4 > 93.2 1\n",
      "93.7 > 93.4 1\n",
      "94.0 > 93.7 1\n",
      "94.3 > 94.0 1\n",
      "94.6 > 94.3 1\n",
      "94.5 < 94.6 0\n",
      "94.9 > 94.5 1\n",
      "95.8 > 94.9 1\n",
      "97.0 > 95.8 1\n",
      "97.5 > 97.0 1\n",
      "97.7 > 97.5 1\n",
      "97.9 > 97.7 1\n",
      "98.2 > 97.9 1\n",
      "98.0 < 98.2 0\n",
      "97.6 < 98.0 0\n",
      "97.8 > 97.6 1\n",
      "97.9 > 97.8 1\n",
      "97.9 < 97.9 0\n",
      "98.6 > 97.9 1\n",
      "99.2 > 98.6 1\n",
      "99.5 > 99.2 1\n",
      "99.9 > 99.5 1\n",
      "100.2 > 99.9 1\n",
      "100.7 > 100.2 1\n",
      "101.0 > 100.7 1\n",
      "101.2 > 101.0 1\n",
      "101.3 > 101.2 1\n",
      "101.9 > 101.3 1\n",
      "102.4 > 101.9 1\n",
      "102.6 > 102.4 1\n",
      "103.1 > 102.6 1\n",
      "103.4 > 103.1 1\n",
      "103.7 > 103.4 1\n",
      "104.1 > 103.7 1\n",
      "104.5 > 104.1 1\n",
      "105.0 > 104.5 1\n",
      "105.3 > 105.0 1\n",
      "105.3 < 105.3 0\n",
      "105.3 < 105.3 0\n",
      "105.5 > 105.3 1\n",
      "106.0 > 105.5 1\n",
      "106.4 > 106.0 1\n",
      "106.9 > 106.4 1\n",
      "107.3 > 106.9 1\n",
      "107.6 > 107.3 1\n",
      "107.8 > 107.6 1\n",
      "108.0 > 107.8 1\n",
      "108.3 > 108.0 1\n",
      "108.7 > 108.3 1\n",
      "109.0 > 108.7 1\n",
      "109.3 > 109.0 1\n",
      "109.6 > 109.3 1\n",
      "109.3 < 109.6 0\n",
      "108.8 < 109.3 0\n",
      "108.6 < 108.8 0\n",
      "108.9 > 108.6 1\n",
      "109.5 > 108.9 1\n",
      "109.5 < 109.5 0\n",
      "109.7 > 109.5 1\n",
      "110.2 > 109.7 1\n",
      "110.3 > 110.2 1\n",
      "110.4 > 110.3 1\n",
      "110.5 > 110.4 1\n",
      "111.2 > 110.5 1\n",
      "111.6 > 111.2 1\n",
      "112.1 > 111.6 1\n",
      "112.7 > 112.1 1\n",
      "113.1 > 112.7 1\n",
      "113.5 > 113.1 1\n",
      "113.8 > 113.5 1\n",
      "114.4 > 113.8 1\n",
      "115.0 > 114.4 1\n",
      "115.3 > 115.0 1\n",
      "115.4 > 115.3 1\n",
      "115.4 < 115.4 0\n",
      "115.7 > 115.4 1\n",
      "116.0 > 115.7 1\n",
      "116.5 > 116.0 1\n",
      "117.1 > 116.5 1\n",
      "117.5 > 117.1 1\n",
      "118.0 > 117.5 1\n",
      "118.5 > 118.0 1\n",
      "119.0 > 118.5 1\n",
      "119.8 > 119.0 1\n",
      "120.2 > 119.8 1\n",
      "120.3 > 120.2 1\n",
      "120.5 > 120.3 1\n",
      "121.1 > 120.5 1\n",
      "121.6 > 121.1 1\n",
      "122.3 > 121.6 1\n",
      "123.1 > 122.3 1\n",
      "123.8 > 123.1 1\n",
      "124.1 > 123.8 1\n",
      "124.4 > 124.1 1\n",
      "124.6 > 124.4 1\n",
      "125.0 > 124.6 1\n",
      "125.6 > 125.0 1\n",
      "125.9 > 125.6 1\n",
      "126.1 > 125.9 1\n",
      "127.4 > 126.1 1\n",
      "128.0 > 127.4 1\n",
      "128.7 > 128.0 1\n",
      "128.9 > 128.7 1\n",
      "129.2 > 128.9 1\n",
      "129.9 > 129.2 1\n",
      "130.4 > 129.9 1\n",
      "131.6 > 130.4 1\n",
      "132.7 > 131.6 1\n",
      "133.5 > 132.7 1\n",
      "133.8 > 133.5 1\n",
      "133.8 < 133.8 0\n",
      "134.6 > 133.8 1\n",
      "134.8 > 134.6 1\n",
      "135.0 > 134.8 1\n",
      "135.2 > 135.0 1\n",
      "135.6 > 135.2 1\n",
      "136.0 > 135.6 1\n",
      "136.2 > 136.0 1\n",
      "136.6 > 136.2 1\n",
      "137.2 > 136.6 1\n",
      "137.4 > 137.2 1\n",
      "137.8 > 137.4 1\n",
      "137.9 > 137.8 1\n",
      "138.1 > 137.9 1\n",
      "138.6 > 138.1 1\n",
      "139.3 > 138.6 1\n",
      "139.5 > 139.3 1\n",
      "139.7 > 139.5 1\n",
      "140.2 > 139.7 1\n",
      "140.5 > 140.2 1\n",
      "140.9 > 140.5 1\n",
      "141.3 > 140.9 1\n",
      "141.8 > 141.3 1\n",
      "142.0 > 141.8 1\n",
      "141.9 < 142.0 0\n",
      "142.6 > 141.9 1\n",
      "143.1 > 142.6 1\n",
      "143.6 > 143.1 1\n",
      "144.0 > 143.6 1\n",
      "144.2 > 144.0 1\n",
      "144.4 > 144.2 1\n",
      "144.4 < 144.4 0\n",
      "144.8 > 144.4 1\n",
      "145.1 > 144.8 1\n",
      "145.7 > 145.1 1\n",
      "145.8 > 145.7 1\n",
      "145.8 < 145.8 0\n",
      "146.2 > 145.8 1\n",
      "146.7 > 146.2 1\n",
      "147.2 > 146.7 1\n",
      "147.4 > 147.2 1\n",
      "147.5 > 147.4 1\n",
      "148.0 > 147.5 1\n",
      "148.4 > 148.0 1\n",
      "149.0 > 148.4 1\n",
      "149.4 > 149.0 1\n",
      "149.5 > 149.4 1\n",
      "149.7 > 149.5 1\n",
      "149.7 < 149.7 0\n",
      "150.3 > 149.7 1\n",
      "150.9 > 150.3 1\n",
      "151.4 > 150.9 1\n",
      "151.9 > 151.4 1\n",
      "152.2 > 151.9 1\n",
      "152.5 > 152.2 1\n",
      "152.5 < 152.5 0\n",
      "152.9 > 152.5 1\n",
      "153.2 > 152.9 1\n",
      "153.7 > 153.2 1\n",
      "153.6 < 153.7 0\n",
      "153.5 < 153.6 0\n",
      "154.4 > 153.5 1\n",
      "154.9 > 154.4 1\n",
      "155.7 > 154.9 1\n",
      "156.3 > 155.7 1\n",
      "156.6 > 156.3 1\n",
      "156.7 > 156.6 1\n",
      "157.0 > 156.7 1\n",
      "157.3 > 157.0 1\n",
      "157.8 > 157.3 1\n",
      "158.3 > 157.8 1\n",
      "158.6 > 158.3 1\n",
      "158.6 < 158.6 0\n",
      "159.1 > 158.6 1\n",
      "159.6 > 159.1 1\n",
      "160.0 > 159.6 1\n",
      "160.2 > 160.0 1\n",
      "160.1 < 160.2 0\n",
      "160.3 > 160.1 1\n",
      "160.5 > 160.3 1\n",
      "160.8 > 160.5 1\n",
      "161.2 > 160.8 1\n",
      "161.6 > 161.2 1\n",
      "161.5 < 161.6 0\n",
      "161.3 < 161.5 0\n",
      "161.6 > 161.3 1\n",
      "161.9 > 161.6 1\n",
      "162.2 > 161.9 1\n",
      "162.5 > 162.2 1\n",
      "162.8 > 162.5 1\n",
      "163.0 > 162.8 1\n",
      "163.2 > 163.0 1\n",
      "163.4 > 163.2 1\n",
      "163.6 > 163.4 1\n",
      "164.0 > 163.6 1\n",
      "164.0 < 164.0 0\n",
      "163.9 < 164.0 0\n",
      "164.3 > 163.9 1\n",
      "164.5 > 164.3 1\n",
      "165.0 > 164.5 1\n",
      "166.2 > 165.0 1\n",
      "166.2 < 166.2 0\n",
      "166.2 < 166.2 0\n",
      "166.7 > 166.2 1\n",
      "167.1 > 166.7 1\n",
      "167.9 > 167.1 1\n",
      "168.2 > 167.9 1\n",
      "168.3 > 168.2 1\n",
      "168.3 < 168.3 0\n",
      "168.8 > 168.3 1\n",
      "169.8 > 168.8 1\n",
      "171.2 > 169.8 1\n",
      "171.3 > 171.2 1\n",
      "171.5 > 171.3 1\n",
      "172.4 > 171.5 1\n",
      "172.8 > 172.4 1\n",
      "172.8 < 172.8 0\n",
      "173.7 > 172.8 1\n",
      "174.0 > 173.7 1\n",
      "174.1 > 174.0 1\n",
      "174.0 < 174.1 0\n",
      "175.1 > 174.0 1\n",
      "175.8 > 175.1 1\n",
      "176.2 > 175.8 1\n",
      "176.9 > 176.2 1\n",
      "177.7 > 176.9 1\n",
      "178.0 > 177.7 1\n",
      "177.5 < 178.0 0\n",
      "177.5 < 177.5 0\n",
      "178.3 > 177.5 1\n",
      "177.7 < 178.3 0\n",
      "177.4 < 177.7 0\n",
      "176.7 < 177.4 0\n",
      "177.1 > 176.7 1\n",
      "177.8 > 177.1 1\n",
      "178.8 > 177.8 1\n",
      "179.8 > 178.8 1\n",
      "179.8 < 179.8 0\n",
      "179.9 > 179.8 1\n",
      "180.1 > 179.9 1\n",
      "180.7 > 180.1 1\n",
      "181.0 > 180.7 1\n",
      "181.3 > 181.0 1\n",
      "181.3 < 181.3 0\n",
      "180.9 < 181.3 0\n",
      "181.7 > 180.9 1\n",
      "183.1 > 181.7 1\n",
      "184.2 > 183.1 1\n",
      "183.8 < 184.2 0\n",
      "183.5 < 183.8 0\n",
      "183.7 > 183.5 1\n",
      "183.9 > 183.7 1\n",
      "184.6 > 183.9 1\n",
      "185.2 > 184.6 1\n",
      "185.0 < 185.2 0\n",
      "184.5 < 185.0 0\n",
      "184.3 < 184.5 0\n",
      "185.2 > 184.3 1\n",
      "186.2 > 185.2 1\n",
      "187.4 > 186.2 1\n",
      "188.0 > 187.4 1\n",
      "189.1 > 188.0 1\n",
      "189.7 > 189.1 1\n",
      "189.4 < 189.7 0\n",
      "189.5 > 189.4 1\n",
      "189.9 > 189.5 1\n",
      "190.9 > 189.9 1\n",
      "191.0 > 190.9 1\n",
      "190.3 < 191.0 0\n",
      "190.7 > 190.3 1\n",
      "191.8 > 190.7 1\n",
      "193.3 > 191.8 1\n",
      "194.6 > 193.3 1\n",
      "194.4 < 194.6 0\n",
      "194.5 > 194.4 1\n",
      "195.4 > 194.5 1\n",
      "196.4 > 195.4 1\n",
      "198.8 > 196.4 1\n",
      "199.2 > 198.8 1\n",
      "197.6 < 199.2 0\n",
      "196.8 < 197.6 0\n",
      "198.3 > 196.8 1\n",
      "198.7 > 198.3 1\n",
      "199.8 > 198.7 1\n",
      "201.5 > 199.8 1\n",
      "202.5 > 201.5 1\n",
      "202.9 > 202.5 1\n",
      "203.5 > 202.9 1\n",
      "203.9 > 203.5 1\n",
      "202.9 < 203.9 0\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "201.8 < 202.9 0\n",
      "201.5 < 201.8 0\n",
      "201.8 > 201.5 1\n",
      "202.42 > 201.8 1\n",
      "203.5 > 202.42 1\n",
      "205.35 > 203.5 1\n",
      "206.69 > 205.35 1\n",
      "207.95 > 206.69 1\n",
      "208.35 > 207.95 1\n",
      "208.3 < 208.35 0\n",
      "207.92 < 208.3 0\n",
      "208.49 > 207.92 1\n",
      "208.94 > 208.49 1\n",
      "210.18 > 208.94 1\n",
      "210.04 < 210.18 0\n",
      "211.08 > 210.04 1\n",
      "211.69 > 211.08 1\n",
      "213.53 > 211.69 1\n",
      "214.82 > 213.53 1\n",
      "216.63 > 214.82 1\n",
      "218.81 > 216.63 1\n",
      "219.96 > 218.81 1\n",
      "219.09 < 219.96 0\n",
      "218.78 < 219.09 0\n",
      "216.57 < 218.78 0\n",
      "212.43 < 216.57 0\n",
      "210.23 < 212.43 0\n",
      "211.14 > 210.23 1\n",
      "212.19 > 211.14 1\n",
      "212.71 > 212.19 1\n",
      "213.24 > 212.71 1\n",
      "213.86 > 213.24 1\n",
      "215.69 > 213.86 1\n",
      "215.35 < 215.69 0\n",
      "215.83 > 215.35 1\n",
      "215.97 > 215.83 1\n",
      "216.18 > 215.97 1\n",
      "216.33 > 216.18 1\n",
      "215.95 < 216.33 0\n",
      "216.69 > 215.95 1\n",
      "216.74 > 216.69 1\n",
      "217.63 > 216.74 1\n",
      "218.01 > 217.63 1\n",
      "218.18 > 218.01 1\n",
      "217.97 < 218.18 0\n",
      "218.01 > 217.97 1\n",
      "218.31 > 218.01 1\n",
      "218.44 > 218.31 1\n",
      "218.71 > 218.44 1\n",
      "218.8 > 218.71 1\n",
      "219.18 > 218.8 1\n",
      "220.22 > 219.18 1\n",
      "221.31 > 220.22 1\n",
      "223.47 > 221.31 1\n",
      "224.91 > 223.47 1\n",
      "225.96 > 224.91 1\n",
      "225.72 < 225.96 0\n",
      "225.92 > 225.72 1\n",
      "226.54 > 225.92 1\n",
      "226.89 > 226.54 1\n",
      "226.42 < 226.89 0\n",
      "226.23 < 226.42 0\n",
      "225.67 < 226.23 0\n",
      "226.66 > 225.67 1\n",
      "227.66 > 226.66 1\n",
      "229.39 > 227.66 1\n",
      "230.09 > 229.39 1\n",
      "229.81 < 230.09 0\n",
      "229.48 < 229.81 0\n",
      "229.1 < 229.48 0\n",
      "230.38 > 229.1 1\n",
      "231.41 > 230.38 1\n",
      "231.32 < 231.41 0\n",
      "230.22 < 231.32 0\n",
      "229.6 < 230.22 0\n",
      "230.28 > 229.6 1\n",
      "232.17 > 230.28 1\n",
      "232.77 > 232.17 1\n",
      "232.53 < 232.77 0\n",
      "232.94 > 232.53 1\n",
      "233.5 > 232.94 1\n",
      "233.6 > 233.5 1\n",
      "233.88 > 233.6 1\n",
      "234.15 > 233.88 1\n",
      "233.55 < 234.15 0\n",
      "233.07 < 233.55 0\n",
      "233.05 < 233.07 0\n",
      "233.92 > 233.05 1\n",
      "234.78 > 233.92 1\n",
      "236.29 > 234.78 1\n",
      "237.07 > 236.29 1\n",
      "237.9 > 237.07 1\n",
      "238.34 > 237.9 1\n",
      "238.25 < 238.34 0\n",
      "237.85 < 238.25 0\n",
      "238.03 > 237.85 1\n",
      "237.43 < 238.03 0\n",
      "236.15 < 237.43 0\n",
      "234.81 < 236.15 0\n",
      "233.71 < 234.81 0\n",
      "234.72 > 233.71 1\n",
      "236.12 > 234.72 1\n",
      "236.6 > 236.12 1\n",
      "237.81 > 236.6 1\n",
      "238.64 > 237.81 1\n",
      "238.65 > 238.64 1\n",
      "238.32 < 238.65 0\n",
      "237.94 < 238.32 0\n",
      "237.84 < 237.94 0\n",
      "237.34 < 237.84 0\n",
      "236.53 < 237.34 0\n",
      "236.92 > 236.53 1\n",
      "237.11 > 236.92 1\n",
      "238.13 > 237.11 1\n",
      "239.26 > 238.13 1\n",
      "240.23 > 239.26 1\n",
      "241.02 > 240.23 1\n",
      "240.63 < 241.02 0\n",
      "240.85 > 240.63 1\n",
      "241.43 > 240.85 1\n",
      "241.73 > 241.43 1\n",
      "241.35 < 241.73 0\n",
      "241.43 > 241.35 1\n",
      "242.84 > 241.43 1\n",
      "243.6 > 242.84 1\n",
      "243.8 > 243.6 1\n",
      "244.52 > 243.8 1\n",
      "244.73 > 244.52 1\n",
      "244.96 > 244.73 1\n",
      "244.79 < 244.96 0\n",
      "245.52 > 244.79 1\n",
      "246.82 > 245.52 1\n",
      "246.66 < 246.82 0\n",
      "246.67 > 246.66 1\n",
      "246.52 < 246.67 0\n",
      "247.87 > 246.52 1\n",
      "248.99 > 247.87 1\n",
      "249.55 > 248.99 1\n",
      "249.84 > 249.55 1\n"
     ]
    },
    {
     "ename": "KeyError",
     "evalue": "1768",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mKeyError\u001b[0m                                  Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-11-04ec3e7aa352>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m      4\u001b[0m \u001b[0mtarget\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      5\u001b[0m \u001b[1;32mfor\u001b[0m \u001b[0mi\u001b[0m \u001b[1;32min\u001b[0m \u001b[0msp\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'Consumer Price Index'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 6\u001b[1;33m     \u001b[1;32mif\u001b[0m \u001b[0msp\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'Consumer Price Index'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mcnt\u001b[0m\u001b[1;33m]\u001b[0m \u001b[1;33m>\u001b[0m \u001b[0mi\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m      7\u001b[0m         \u001b[0mprint\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0msp\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'Consumer Price Index'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mcnt\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;34m'>'\u001b[0m \u001b[1;33m,\u001b[0m \u001b[0mi\u001b[0m \u001b[1;33m,\u001b[0m \u001b[1;34m'1'\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      8\u001b[0m         \u001b[0mtarget\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\pandas\\core\\series.py\u001b[0m in \u001b[0;36m__getitem__\u001b[1;34m(self, key)\u001b[0m\n\u001b[0;32m    765\u001b[0m         \u001b[0mkey\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mcom\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_apply_if_callable\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mkey\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    766\u001b[0m         \u001b[1;32mtry\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 767\u001b[1;33m             \u001b[0mresult\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mindex\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mget_value\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mkey\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    768\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    769\u001b[0m             \u001b[1;32mif\u001b[0m \u001b[1;32mnot\u001b[0m \u001b[0mis_scalar\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mresult\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\pandas\\core\\indexes\\base.py\u001b[0m in \u001b[0;36mget_value\u001b[1;34m(self, series, key)\u001b[0m\n\u001b[0;32m   3116\u001b[0m         \u001b[1;32mtry\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   3117\u001b[0m             return self._engine.get_value(s, k,\n\u001b[1;32m-> 3118\u001b[1;33m                                           tz=getattr(series.dtype, 'tz', None))\n\u001b[0m\u001b[0;32m   3119\u001b[0m         \u001b[1;32mexcept\u001b[0m \u001b[0mKeyError\u001b[0m \u001b[1;32mas\u001b[0m \u001b[0me1\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   3120\u001b[0m             \u001b[1;32mif\u001b[0m \u001b[0mlen\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;33m>\u001b[0m \u001b[1;36m0\u001b[0m \u001b[1;32mand\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0minferred_type\u001b[0m \u001b[1;32min\u001b[0m \u001b[1;33m[\u001b[0m\u001b[1;34m'integer'\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;34m'boolean'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32mpandas\\_libs\\index.pyx\u001b[0m in \u001b[0;36mpandas._libs.index.IndexEngine.get_value\u001b[1;34m()\u001b[0m\n",
      "\u001b[1;32mpandas\\_libs\\index.pyx\u001b[0m in \u001b[0;36mpandas._libs.index.IndexEngine.get_value\u001b[1;34m()\u001b[0m\n",
      "\u001b[1;32mpandas\\_libs\\index.pyx\u001b[0m in \u001b[0;36mpandas._libs.index.IndexEngine.get_loc\u001b[1;34m()\u001b[0m\n",
      "\u001b[1;32mpandas\\_libs\\hashtable_class_helper.pxi\u001b[0m in \u001b[0;36mpandas._libs.hashtable.Int64HashTable.get_item\u001b[1;34m()\u001b[0m\n",
      "\u001b[1;32mpandas\\_libs\\hashtable_class_helper.pxi\u001b[0m in \u001b[0;36mpandas._libs.hashtable.Int64HashTable.get_item\u001b[1;34m()\u001b[0m\n",
      "\u001b[1;31mKeyError\u001b[0m: 1768"
     ]
    }
   ],
   "source": [
    "## Use the below code to create a new variable so as to detect increament or decrement of CPI \n",
    "\n",
    "cnt=1\n",
    "target=[]\n",
    "for i in sp['Consumer Price Index']:\n",
    "    if sp['Consumer Price Index'][cnt] > i:\n",
    "        print(sp['Consumer Price Index'][cnt], '>' , i , '1')\n",
    "        target.append(1)\n",
    "    else:\n",
    "        print(sp['Consumer Price Index'][cnt], '<' , i , '0')\n",
    "        target.append(0)\n",
    "    cnt+=1\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "target=pd.DataFrame(data=target, columns=['CPI_Inc'],dtype=int)\n",
    "sp=sp[:1767]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Date</th>\n",
       "      <th>SP500</th>\n",
       "      <th>Dividend</th>\n",
       "      <th>Earnings</th>\n",
       "      <th>Consumer Price Index</th>\n",
       "      <th>Long Interest Rate</th>\n",
       "      <th>Real Price</th>\n",
       "      <th>Real Dividend</th>\n",
       "      <th>Real Earnings</th>\n",
       "      <th>CPI_Inc</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1762</th>\n",
       "      <td>2017-11-01</td>\n",
       "      <td>2593.61</td>\n",
       "      <td>48.68</td>\n",
       "      <td>108.95</td>\n",
       "      <td>246.67</td>\n",
       "      <td>2.35</td>\n",
       "      <td>2626.9</td>\n",
       "      <td>49.3</td>\n",
       "      <td>110.35</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1763</th>\n",
       "      <td>2017-12-01</td>\n",
       "      <td>2664.34</td>\n",
       "      <td>48.93</td>\n",
       "      <td>109.88</td>\n",
       "      <td>246.52</td>\n",
       "      <td>2.4</td>\n",
       "      <td>2700.13</td>\n",
       "      <td>49.59</td>\n",
       "      <td>111.36</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1764</th>\n",
       "      <td>2018-01-01</td>\n",
       "      <td>2789.8</td>\n",
       "      <td>49.29</td>\n",
       "      <td>49.29</td>\n",
       "      <td>247.87</td>\n",
       "      <td>2.58</td>\n",
       "      <td>2811.96</td>\n",
       "      <td>49.68</td>\n",
       "      <td>49.68</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1765</th>\n",
       "      <td>2018-02-01</td>\n",
       "      <td>2705.16</td>\n",
       "      <td>49.64</td>\n",
       "      <td>49.64</td>\n",
       "      <td>248.99</td>\n",
       "      <td>2.86</td>\n",
       "      <td>2714.34</td>\n",
       "      <td>49.81</td>\n",
       "      <td>49.81</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1766</th>\n",
       "      <td>2018-03-01</td>\n",
       "      <td>2702.77</td>\n",
       "      <td>50</td>\n",
       "      <td>50</td>\n",
       "      <td>249.55</td>\n",
       "      <td>2.84</td>\n",
       "      <td>2705.82</td>\n",
       "      <td>50.06</td>\n",
       "      <td>50.06</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            Date    SP500 Dividend Earnings Consumer Price Index  \\\n",
       "1762  2017-11-01  2593.61    48.68   108.95               246.67   \n",
       "1763  2017-12-01  2664.34    48.93   109.88               246.52   \n",
       "1764  2018-01-01   2789.8    49.29    49.29               247.87   \n",
       "1765  2018-02-01  2705.16    49.64    49.64               248.99   \n",
       "1766  2018-03-01  2702.77       50       50               249.55   \n",
       "\n",
       "     Long Interest Rate Real Price Real Dividend Real Earnings  CPI_Inc  \n",
       "1762               2.35     2626.9          49.3        110.35        0  \n",
       "1763                2.4    2700.13         49.59        111.36        1  \n",
       "1764               2.58    2811.96         49.68         49.68        1  \n",
       "1765               2.86    2714.34         49.81         49.81        1  \n",
       "1766               2.84    2705.82         50.06         50.06        1  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "frames=[sp,target]\n",
    "sp=pd.concat(frames, axis=1, join='outer')\n",
    "sp.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "sp3=sp.drop('Date',axis=1)\n",
    "\n",
    "test=sp3[-40:]\n",
    "train=sp3[:-40]\n",
    "#Train\n",
    "sp1=np.array(train)\n",
    "#Test\n",
    "sp2=np.array(test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "### To change the dimension of the data for deep learning model\n",
    "\n",
    "from numpy import array\n",
    "\n",
    "def split_sequences(sequences, n_steps):\n",
    "    X, y = list(), list()\n",
    "    for i in range(len(sequences)):\n",
    "        # find the end of this pattern\n",
    "        end_ix = i + n_steps\n",
    "        \n",
    "        # check if we are beyond the dataset\n",
    "        if end_ix > len(sequences):\n",
    "            break\n",
    "        # gather input and output parts of the pattern\n",
    "        seq_x, seq_y = sequences[i:end_ix, :-1], sequences[end_ix-1, -1]\n",
    "        \n",
    "        X.append(seq_x)\n",
    "        y.append(seq_y)\n",
    "    return array(X), array(y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1720, 8, 8) (1720,)\n",
      "(33, 8, 8) (33,)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "8"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# choose a number of time steps\n",
    "n_steps = 8\n",
    "\n",
    "# convert into input/output\n",
    "X1, y1 = split_sequences(sp1, n_steps)\n",
    "# convert into input/output\n",
    "X2, y2 = split_sequences(sp2, n_steps)\n",
    "\n",
    "print(X1.shape, y1.shape)\n",
    "print(X2.shape, y2.shape)\n",
    "\n",
    "\n",
    "n_features = X2.shape[2]\n",
    "n_features"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# B. CNN Model:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Part A - Deep Learning model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "___Hyperparameters:___<br>\n",
    "filters=128<br>\n",
    "kernel_size=3<br>\n",
    "Activation functions=  relu<br>\n",
    "loss=mean_squared_error<br>\n",
    "optimizer= adam<br>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "# univariate cnn example\n",
    "import keras\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout, Activation, Flatten\n",
    "from keras.layers import BatchNormalization\n",
    "from keras.layers.convolutional import Conv1D\n",
    "from keras.layers.convolutional import MaxPooling1D\n",
    "from keras import metrics\n",
    "from keras import optimizers\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "# define model\n",
    "model1 = Sequential()\n",
    "\n",
    "model1.add(Conv1D(filters=128, kernel_size=3, activation='relu', input_shape=(n_steps, n_features)))\n",
    "model1.add(MaxPooling1D(pool_size=2))\n",
    "model1.add(BatchNormalization())\n",
    "\n",
    "model1.add(Flatten())\n",
    "\n",
    "model1.add(Dropout(0.25))\n",
    "model1.add(Dense(20, activation='relu'))\n",
    "model1.add(Dropout(0.25))\n",
    "model1.add(Dense(1))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      " - 0s - loss: 0.8933 - mean_absolute_error: 0.7287 - accuracy: 0.4110\n",
      "Epoch 2/10\n",
      " - 0s - loss: 0.3926 - mean_absolute_error: 0.5035 - accuracy: 0.5651\n",
      "Epoch 3/10\n",
      " - 0s - loss: 0.2918 - mean_absolute_error: 0.4583 - accuracy: 0.6012\n",
      "Epoch 4/10\n",
      " - 0s - loss: 0.2818 - mean_absolute_error: 0.4488 - accuracy: 0.6029\n",
      "Epoch 5/10\n",
      " - 0s - loss: 0.2488 - mean_absolute_error: 0.4266 - accuracy: 0.6355\n",
      "Epoch 6/10\n",
      " - 0s - loss: 0.2464 - mean_absolute_error: 0.4329 - accuracy: 0.6349\n",
      "Epoch 7/10\n",
      " - 0s - loss: 0.2466 - mean_absolute_error: 0.4278 - accuracy: 0.6331\n",
      "Epoch 8/10\n",
      " - 0s - loss: 0.2363 - mean_absolute_error: 0.4306 - accuracy: 0.6355\n",
      "Epoch 9/10\n",
      " - 0s - loss: 0.2357 - mean_absolute_error: 0.4266 - accuracy: 0.6407\n",
      "Epoch 10/10\n",
      " - 0s - loss: 0.2327 - mean_absolute_error: 0.4268 - accuracy: 0.6302\n"
     ]
    }
   ],
   "source": [
    "model1.compile(optimizer='adam', loss='mean_squared_error',metrics=[metrics.mae, 'accuracy'])\n",
    "history=model1.fit(X1, y1, batch_size=32, epochs=10, verbose=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "ename": "SyntaxError",
     "evalue": "invalid syntax (<ipython-input-22-0e46dadb5838>, line 1)",
     "output_type": "error",
     "traceback": [
      "\u001b[1;36m  File \u001b[1;32m\"<ipython-input-22-0e46dadb5838>\"\u001b[1;36m, line \u001b[1;32m1\u001b[0m\n\u001b[1;33m    history[]\u001b[0m\n\u001b[1;37m            ^\u001b[0m\n\u001b[1;31mSyntaxError\u001b[0m\u001b[1;31m:\u001b[0m invalid syntax\n"
     ]
    }
   ],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEWCAYAAAB8LwAVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8li6FKAAAgAElEQVR4nO3deXxU5dn/8c9FAiRsAgmikGACIhI2hYAiaqso4kqfaquoj0uraB+tXbTPA/6sUrtZu2pLtVaxWq1LrbVYFWoVtypCsJQdgbBFtrBv2XP9/pgTGOIAE8kw2/f9es0rZ59rRjnXnOu+z7nN3REREWmsRbwDEBGRxKQEISIiESlBiIhIREoQIiISkRKEiIhEpAQhIiIRKUFI2jOzAjNzM8uMYtvrzOy9IxGXSLwpQUhSMbOVZlZtZrmNls8JTvIF8YlMJPUoQUgyWgGMbZgxswFAdvzCSQzRXAGJNIUShCSjPwLXhM1fCzwZvoGZHWVmT5pZuZmtMrO7zKxFsC7DzH5mZpvMrBS4MMK+j5nZOjP7xMx+YGYZ0QRmZn82s/Vmtt3M3jGzfmHrss3s50E8283sPTPLDtadbmbvm9k2M1tjZtcFy98ysxvCjrFfiSu4arrFzJYCS4NlDwTH2GFms83sjLDtM8zsTjNbbmY7g/X5ZjbJzH7e6LO8bGbfjOZzS2pSgpBkNAPoYGZ9gxP35cBTjbb5NXAU0BP4HKGEcn2w7kbgIuBkoBi4rNG+TwC1wPHBNqOAG4jOa0Bv4GjgI+DpsHU/A4YApwGdgf8F6s2sR7Dfr4EuwEnAnCjfD+ALwClAUTA/KzhGZ+BPwJ/NLCtY921CV18XAB2ArwB7gs88NiyJ5gIjgWeaEIekGnfXS6+keQErgXOAu4AfA6OB14FMwIECIAOoAorC9rsJeCuYfhO4OWzdqGDfTKBrsG922PqxwPRg+jrgvShj7Rgc9yhCP8YqgEERtpsA/PUAx3gLuCFsfr/3D45/9iHi2NrwvsASYMwBtlsEnBtM3wq8Gu//3nrF96WapSSrPwLvAIU0Ki8BuUArYFXYslVA92C6G7Cm0boGxwEtgXVm1rCsRaPtIwquZn4IfInQlUB9WDytgSxgeYRd8w+wPFr7xWZmtxO64ulGKIF0CGI41Hs9AVxNKOFeDTxwGDFJClCJSZKSu68i1Fh9AfBio9WbgBpCJ/sGPYBPgul1hE6U4esarCF0BZHr7h2DVwd378ehXQmMIXSFcxShqxkAC2KqBHpF2G/NAZYD7AbahM0fE2GbvY9kDtob/g/4MtDJ3TsC24MYDvVeTwFjzGwQ0Bd46QDbSZpQgpBk9lVC5ZXd4QvdvQ54HvihmbU3s+MI1d4b2imeB24zszwz6wSMD9t3HfAP4Odm1sHMWphZLzP7XBTxtCeUXDYTOqn/KOy49cBk4Bdm1i1oLB5uZq0JtVOcY2ZfNrNMM8sxs5OCXecAXzSzNmZ2fPCZDxVDLVAOZJrZ3YSuIBo8CnzfzHpbyEAzywliLCPUfvFH4C/uXhHFZ5YUpgQhScvdl7t7yQFWf53Qr+9S4D1CjbWTg3W/B6YB/yHUkNz4CuQaQiWqhYTq9y8Ax0YR0pOEylWfBPvOaLT+DmAeoZPwFuAnQAt3X03oSuj2YPkcYFCwzy+BamADoRLQ0xzcNEIN3h8HsVSyfwnqF4QS5D+AHcBj7N9F+AlgAKEkIWnO3DVgkIiEmNmZhK60CoKrHkljuoIQEQDMrCXwDeBRJQcBJQgRAcysL7CNUCntV3EORxKESkwiIhKRriBERCSilLlRLjc31wsKCuIdhohIUpk9e/Ymd+8SaV3KJIiCggJKSg7U41FERCIxs1UHWqcSk4iIRKQEISIiESlBiIhIRCnTBhFJTU0NZWVlVFZWxjuUIyYrK4u8vDxatmwZ71BEJMmldIIoKyujffv2FBQUEPbo5pTl7mzevJmysjIKCwvjHY6IJLmULjFVVlaSk5OTFskBwMzIyclJqysmEYmdlE4QQNokhwbp9nlFJHZSPkGIiMTCzBVbePGjMqprU/e5hindBhFvmzdvZuTIkQCsX7+ejIwMunQJ3bA4c+ZMWrVqdchjXH/99YwfP54+ffrENFYROTR354Plm/nVG0uZuWILAL95cxnfvbiIs/ocHefomp8SRAzl5OQwZ84cACZOnEi7du2444479tumYXDwFi0iX8w9/vjjMY9TRA7O3Xl36SYefGMpJau20rVDa+65uIj8Tm340auLuP7xWZzT92juurCIgty28Q632ajEFAfLli2jf//+3HzzzQwePJh169Yxbtw4iouL6devH/fee+/ebU8//XTmzJlDbW0tHTt2ZPz48QwaNIjhw4ezcePGOH4KkdTn7kxfvJH/+u37XDN5Jmu3VfD9L/Tn7e+cxfUjCjmnqCtTv3kmE84/kQ+Wb2bUL9/hp9MWs7uqNt6hN4u0uYL43ssLWLh2R7Mes6hbB+65OJqx7D9t4cKFPP744zz88MMA3HfffXTu3Jna2lrOOussLrvsMoqKivbbZ/v27Xzuc5/jvvvu49vf/jaTJ09m/PjxkQ4vIofB3fnnoo08+MZS5n2ynbxO2fz4iwO4dHAerTL3/13dKrMFN32uF/91cnfue20xk6Yv5y+zP2HCBSdyyaBuSd1xRFcQcdKrVy+GDh26d/6ZZ55h8ODBDB48mEWLFrFw4cJP7ZOdnc35558PwJAhQ1i5cuWRClckLdTXO1Pnr+OCB9/jxidL2FFZw/2XDWT6HZ9n7LAen0oO4Y7ukMUvLj+Jv3xtOLntW/GNZ+dw+e9mNPsP0yMpba4gPusv/Vhp23ZfnXLp0qU88MADzJw5k44dO3L11VdHvJchvFE7IyOD2trUuIwVibe6eue1+ev49RvLWLJhJ4W5bfn5lwYx5qRuZGY07Xf0kOM687dbTuf5kjX8dNoSLvr1u1x1ynHcPuoEOrY5dMeURJI2CSKR7dixg/bt29OhQwfWrVvHtGnTGD16dLzDEkl5dfXO3+eu5ddvLmPZxl306tKWB644iYsGdiOjxWcvDWW0MMYO68EF/Y/ll//8mD/OWMXLc9dyx6g+jB3W47COfSQpQSSAwYMHU1RURP/+/enZsycjRoyId0giKa22rp6/zVnLpOnLKN20mz5d2/ObK0/m/P7HNuvJ+6g2LZl4ST+uGJbPxCkLuOul+fzpw9V8b0w/hhZ0brb3iZWUGZO6uLjYGw8YtGjRIvr27RuniOInXT+3yKHU1NXz139/wqTpy1i1eQ99j+3AN0Yez6iiY2gR41/17s6r89bzw1cWsnZ7JWNO6saE8/tyzFFZMX3fQzGz2e5eHGmdriBEJOVV19bzl4/KmDR9GWVbKxjQ/Sh+f00x5/Q9+oj1MjIzLhx4LGed2IWH31rOw++U8vrCDdx69vF89fRCWmdmHJE4mkIJQkRSVlVtHc+XlPHQ9GWs3V7JoPyOfH9Mfz7fp0vcup+2aZXJt0f14bIh+fzglYXcP3UJz89aw90XF3H2iV3jEtOBpHyCcPek7ofcVKlSMhQ5HJU1dTw7czUPv13K+h2VDDmuE/ddOpAzeucmzPmgR04bHrmmmHc+Lmfiywv4yh9KOKtPF+6+uB+FCXI3dkq3QaxYsYL27dunzSO/G8aD2Llzp8aDkLRUUV3H0x+u4nfvlFK+s4phhZ355sjeDO+V2OeA6tp6nnh/JQ+8sZSq2jq+enpPvn728bRtHfvf8Adrg0jpBKER5UTSw+6qWp6asYrfv1vKpl3VnNYrh9tG9ubUnjnxDq1JNu6s5P6pS3hhdhldO7Rmwvl9GXNSbO/GjluCMLPRwANABvCou98XYZsvAxMBB/7j7lcGy+uAecFmq939koO9V6QEISKpbWdlDU9+sIpH3y1l654azuidy20jeydFF9KD+Wj1ViZOWcDcsu0UH9eJiZf0o3/3o2LyXnFJEGaWAXwMnAuUAbOAse6+MGyb3sDzwNnuvtXMjnb3jcG6Xe7eLtr3U4IQSR/bK2p44v2VPPbeCrZX1HBWny58fWRvBvfoFO/Qmk19vfPn2Wu4f+oStuypZuywHtwxqg+d2zbv3djx6uY6DFjm7qVBEM8CY4DwhwzdCExy960ADclBRCScu7Ni024+KN3MB8s38/bH5eysrOWcvl25beTxDMzrGO8Qm12LFsblQ3swuv+x/OqfH/PkB6t4Ze46bh91AlcO69HkR4B8FrFMEN2BNWHzZcApjbY5AcDM/kWoDDXR3acG67LMrASoBe5z95cav4GZjQPGAfTo0aN5oxeRuFqzZQ/vL9/EB8s3M6N0C+t3hNoSu3ZozblFXfnKiMKYlV0SyVHZLbnn4n6MHdaDiVMWcPffFvCnD1cz8ZJ+MW9jiWWCiNSq0rielQn0Bj4P5AHvmll/d98G9HD3tWbWE3jTzOa5+/L9Dub+CPAIhEpMzf0BROTIWbutgg+Wb957lfDJtgoActu14pSeOZzWK4fhPXMozG2b0D2SYuWEru15+oZTmDp/PT94ZRFXPDKDiwd1Y8L5J9KtY3ZM3jOWCaIMyA+bzwPWRthmhrvXACvMbAmhhDHL3dcCuHupmb0FnAwsR0RSwsYdlXuTwQelm1m1eQ8AHdu05NTCHMad2ZPhvXLofXS7tEwIkZgZ5w84ls/3OZqH317Ow28v558LN3DLWb245azjm/17imWCmAX0NrNC4BPgCuDKRtu8BIwF/mBmuYRKTqVm1gnY4+5VwfIRwP0xjFVEYmzTripmhCWE0vLdALTPyuSUwhyuGV7A8J45nHhM+5g/FynZZbfK4FvnnsBlQ/L44SuLmP/Jjpgk0ZglCHevNbNbgWmE2hcmu/sCM7sXKHH3KcG6UWa2EKgDvuPum83sNOB3ZlZPaFCj+8J7P4lI4tu6u5oPV+xLCB9v2AVA21YZDCvszBVD8xneM5eibh2S5vHXiSa/cxse/u8hVNXWxeT4KX2jnIgcOdsrapi5YsvehLB4/Q7cIatlC4YWdObUnjkM75XDgO5H0fII9MCR6OhpriLS7HZV1TJr5RZmLN/M+8s3s2Dtduo9NEbzkB6d+NY5JzC8Vw6D8joedKhOSVxKECJySDV19azbVknppl2hq4TSzcwt205dvdMywzg5vxO3nt2b4T1zOLlHR7JaJt6jq6XplCBEhLp6Z/2OStZs2UPZ1op9f7fu4ZOtFazbXkF9UI3OaGEMzDuKm4JeRsXHdSa7lRJCKlKCEEkD9fVO+a4qyrbuYc2Win1/t4X+rt1WQW39vvZIM+jaPov8ztkMK+xMfqds8jq1Ia9zNgPzOtLuCDxlVOJP/5VFUoC7s2V3NWu2VuyfBIL5sq0VVNfW77dPbrvW5HXKZlB+Ry4ceCz5ndqQ1ymb/M5t6NYxKyFHOJMjSwlCJEls31PDmq17DpgA9lTv39WxU5uW5HVqw4nHtOecvl33XgXkd86me8c2KgvJISlBiDRBfb1TXVdPVU09VbV1VNWG/lY2zNfU711WVRvarnLv8obt66mqCU1X1oQta7R/Zdg+lTWh+XDtW2eS17kNx+W05fTju+z99R9KANm0z9KYIHJ4lCBEDqC0fBffe3khC9bu2Hvyrq6rP/SOB2EGWZkZtG7ZgtaZLchqmUHrzBa0zgz+tmxBh+yWZLUMW5bZgtYtM+jSrjX5nYOrgE5t6JCdqUdQSEwpQYg0Ul1bzyPvLOfBN5fROrMFFw44luxWGbTOzNj/xB02ve9EHzqZh09nhS3LbGE6qUvSUIIQCfPR6q1M+Ms8lmzYyYUDjuWei4s4ukNWvMMSiQslCBFCQ1f+dNoS/jhjFcd0yOLRa4o5p6hrvMMSiSslCEl7ry/cwHdfms+GnZVcO7yAO87ro37+IihBSBrbuKOSe6Ys4LX56+nTtT2/vXpwSo1pLHK4lCAk7dTXO8/OWsOPX1tEVW093zmvD+PO7KknjIo0ogQhaWXZxl3c+eI8Zq7cwvCeOfzoiwMozG0b77BEEpIShKSF6tp6HnprOZOmLyO7VQb3XzqQLxXnqcupyEEoQUjKK1m5hQkvzmPpxl1cPKgbd19URJf2reMdlkjCU4KQlLWjsob7py7mqRmr6d4xm8evG8pZJx4d77BEkoYShKSkqfPXc8+U+ZTvrOIrIwq5fdQJtFXXVZEm0b8YSSnrt1dyz5T5TFuwgROPac8j/13MoPyO8Q5LJCkpQUhKqK93np65mvtfW0x1XT3/N/pEbjijUF1XRQ6DEoQkvaUbdjL+xXnMXrWVEcfn8MMvDKBAXVdFDpsShCStqto6Jk1fzkNvLaNt60x+9qVBXDq4u7quijQTJQhJSjNXbGH8i3MpLd/NF07qxl0XFZHbTl1XRZqTEoQkle0VNdz32mKembmavE7Z/OH6oXy+j7quisSCEoQkBXfntfnruWfKAjbvquLGMwr51rkn0KaV/hcWiRX965KEt3ZbBXf/bQH/XLSBft06MPnaoQzIOyreYYmkPCUISVh19c5TM1Zx/9TF1Llz5wUn8pURhWSq66rIEaEEIQlp2cadfOeFufx79TbO6J3LD78wgB45beIdlkhaUYKQhPPOx+Xc8vRHZGYYv/jyIP7rZHVdFYkHJQhJKH/6cDXf/dt8eh/djseuG0r3jtnxDkkkbSlBSEKor3d+MnUxv3unlM+d0IXfXHky7bNaxjsskbSmBCFxV1Fdx7eem8PUBeu5+tQeTLy4nxqiRRKAEoTE1cadldz4RAlzP9nOXRf25aunF6q9QSRBKEFI3Hy8YSfXPz6LLburefjqIZzX75h4hyQiYZQgJC7eXVrO/zz1EVmtMnjuplMZmKcxG0QSjRKEHHHPzFzNXS+pp5JIootpS6CZjTazJWa2zMzGH2CbL5vZQjNbYGZ/Clt+rZktDV7XxjJOOTLq650fv7aICS/OY8Txufz55uFKDiIJLGZXEGaWAUwCzgXKgFlmNsXdF4Zt0xuYAIxw961mdnSwvDNwD1AMODA72HdrrOKV2KqsCfVUem3+eq46pQffu0Q9lUQSXSxLTMOAZe5eCmBmzwJjgIVh29wITGo48bv7xmD5ecDr7r4l2Pd1YDTwTAzjlRgp31nFDU+WMLdsm3oqiSSRWCaI7sCasPky4JRG25wAYGb/AjKAie4+9QD7do9dqBIrSzfs5LrHZ7F5d5V6KokkmVgmiEg/ET3C+/cGPg/kAe+aWf8o98XMxgHjAHr06HE4sUoMvLd0E197ajZZrTJ4/qbh6qkkkmRiWQQuA/LD5vOAtRG2+Zu717j7CmAJoYQRzb64+yPuXuzuxV26dGnW4OXwPDtzNdc9PpNuHbN56ZYRSg4iSSiWCWIW0NvMCs2sFXAFMKXRNi8BZwGYWS6hklMpMA0YZWadzKwTMCpYJgmu4ZlK41+cx/BeObzwNfVUEklWMSsxuXutmd1K6MSeAUx29wVmdi9Q4u5T2JcIFgJ1wHfcfTOAmX2fUJIBuLehwVoSV2VNHbc//x9embeOK0/pwb3qqSSS1Mz9U6X9pFRcXOwlJSXxDiNtbdpVxQ1PlPCfsm3ceX5fbjhDPZVEkoGZzXb34kjrdCe1HLalG3Zy/R9msWlXFQ9dNYTR/dVTSSQVKEHIYfnXsk3c/NRsWmdm8Ny44QzKV2O0SKpQgpDP7LlZq/l/f51Pry7teOy6YvI6acxokVSiBCFNVl/v/PQfS3joreWc0TuXSVcNpoNGfxNJOUoQ0iSNeyp975J+tFRPJZGUpAQhUdu0q4obnyxhzppt3HnBidx4Rk/1VBJJYUoQEpVlG0M9lcp3VvHQVYMZ3f/YeIckIjGmBCGH9P6yTdyknkoiaUcJQg7q+VlruPOv8+jZpS2TrxuqnkoiaUQJQiKqr3d+9o8l/FY9lUTS1iG7n5jZrcED8yRNVNbU8fVn/81v31rO2GE9mHzdUCUHkTQUzRXEMYSGC/0ImAxM81R5gJN8ypote/jGs//mo9XqqSSS7g6ZINz9LjP7LqFHbl8P/MbMngcec/flsQ5QjozZq7by2HulTJ2/nlaZLXjoqsGcP0A9lUTSWVRtEO7uZrYeWA/UAp2AF8zsdXf/31gGKLFTW1fPtAUbePS9Uv69ehsdsjIZd2Yvrj3tOI49SmM4iKS7QyYIM7sNuBbYBDxKaMyGGjNrASwFlCCSzM7KGp6btYbH/7WST7ZVcFxOG+4d049LB+fRtrX6LYhISDRng1zgi+6+Knyhu9eb2UWxCUtiYc2WPfzh/ZU8N2sNu6pqGVbYmXsuLmJk365ktFA7g4jsL5oE8SqwdzQ3M2sPFLn7h+6+KGaRSbOZvWork99bwWvz19HCjAsHHstXTy/UONEiclDRJIiHgMFh87sjLJMEE6l94cYze3Lt8AK6aYxoEYlCNAnCwru1BqUlFaoTVKT2he9d0o/Lhqh9QUSaJpozRmnQUP1QMP8/QGnsQpLPYs2WPTzx/kqebWhfKOjM3RcXcY7aF0TkM4omQdwMPAjcBTjwBjAulkFJ9D5avZXH3g21L5gZF6l9QUSaSTQ3ym0ErjgCsUiUauvq+cfCDTz6bikfrd5Ge7UviEgMRHMfRBbwVaAfkNWw3N2/EsO4JIKG9oU/vL+Ssq0V9OjchokXF/Gl4ny1L4hIs4vmrPJHYDFwHnAvcBWg7q1HUNnWPfzhX/u3L9x1YRHnFql9QURiJ5oEcby7f8nMxrj7E2b2J2BarAOToH3hvRW8Ni/UvnDhgFD7ggbsEZEjIZoEURP83WZm/Qk9j6kgZhGluYjtC2f05NrT1L4gIkdWNAnikWA8iLuAKUA74LsxjSoN1dc7T3ywksfeW7Ff+8Jlxfm0U/uCiMTBQc88wQP5drj7VuAdoOcRiSoNvTp/Hd97eSFDjuuk9gURSQgHTRDBXdO3As8foXjS1t//s44u7Vvz/E3DlRhEJCEccshR4HUzu8PM8s2sc8Mr5pGlkd1VtUxfspEL+h+j5CAiCSOa4nbD/Q63hC1zVG5qNm8s3khVbT0XaAQ3EUkg0dxJXXgkAklnr84NlZeKC3RhJiKJI5o7qa+JtNzdn2z+cNJPQ3npiqH5Ki+JSEKJpsQ0NGw6CxgJfAQoQTQDlZdEJFFFU2L6evi8mR1F6PEb0gxUXhKRRBVNL6bG9gC9mzuQdNRQXjpfvZdEJAFF0wbxMqFeSxBKKEXovohm0VBeulDlJRFJQNG0QfwsbLoWWOXuZTGKJ62ovCQiiSyaEtNq4EN3f9vd/wVsNrOCaA5uZqPNbImZLTOz8RHWX2dm5WY2J3jdELauLmz5lCg/T9JQeUlEEl00VxB/Bk4Lm68Llg2NvHmImWUAk4BzgTJglplNcfeFjTZ9zt1vjXCICnc/KYr4ktKb6r0kIgkumiuITHevbpgJpltFsd8wYJm7lwb7PAuM+Wxhpp5XgvLSUJWXRCRBRZMgys3skoYZMxsDbIpiv+7AmrD5smBZY5ea2Vwze8HM8sOWZ5lZiZnNMLMvRHoDMxsXbFNSXl4eRUiJQeUlEUkG0SSIm4E7zWy1ma0G/g+4KYr9Ip35vNH8y0CBuw8E/gk8Ebauh7sXA1cCvzKzXp86mPsj7l7s7sVdunSJIqTEoPKSiCSDaG6UWw6cambtAHP3nVEeuwwIvyLIA9Y2OvbmsNnfAz8JW7c2+FtqZm8BJwPLo3zvhKbykogkg0NeQZjZj8yso7vvcvedZtbJzH4QxbFnAb3NrNDMWgFXEBqRLvzY4T+hLwEWBcs7mVnrYDoXGAE0btxOSioviUiyiKbEdL67b2uYCUaXu+BQO7l7LXArMI3Qif95d19gZveGtWncZmYLzOw/wG3AdcHyvkBJsHw6cF+E3k9JSeUlEUkW0XRzzTCz1u5eBWBm2UDraA7u7q8CrzZadnfY9ARgQoT93gcGRPMeyebVeevIbafykogkvmgSxFPAG2b2eDB/Pfs3JkuUdlfV8ubijVyuR3uLSBKIppH6fjObC5xDqGfSVOC4WAeWilReEpFkEu3TXNcD9cClhMaDWBSziFKYyksikkwOeAVhZicQ6nk0FtgMPEeom+tZRyi2lNLQe+lLQ1ReEpHkcLAS02LgXeBid18GYGbfOiJRpaA3F2+ksqaeCweqvCQiyeFgJaZLCZWWppvZ781sJJHvjpYoqLwkIsnmgAnC3f/q7pcDJwJvAd8CuprZQ2Y26gjFlxJ0c5yIJKNDNlK7+253f9rdLyL0uIw5wKfGdpADU3lJRJJRk8akdvct7v47dz87VgGlIpWXRCQZNSlBSNPtqVZ5SUSSkxJEjDWUl3RznIgkGyWIGHtlbqi8NKxQ5SURSS5KEDGk8pKIJDMliBhSeUlEkpkSRAypvCQiyUwJIkZUXhKRZKcEESMqL4lIslOCiJGGm+NUXhKRZKUEEQN7qkMjx6m8JCLJTAkiBlReEpFUoAQRAyoviUgqUIJoZg3lpdH9u6q8JCJJTQmime19tPeAbvEORUTksChBNDOVl0QkVShBNCOVl0QklShBNCOVl0QklShBNCOVl0QklShBNBOVl0Qk1ShBNJPpi8t1c5yIpBQliGbyyry15LZrxSmFOfEORUSkWShBNIN95SU9e0lEUocSRDNQeUlEUpESRDMI9V5SeUlEUosSxGHaU13LG4s3qLwkIilHCeIwqbwkIqlKCeIwqbwkIqlKCeIwqLwkIqkspgnCzEab2RIzW2Zm4yOsv87Mys1sTvC6IWzdtWa2NHhdG8s4PyuVl0QklWXG6sBmlgFMAs4FyoBZZjbF3Rc22vQ5d7+10b6dgXuAYsCB2cG+W2MV72eh8pKIpLJYXkEMA5a5e6m7VwPPAmOi3Pc84HV33xIkhdeB0TGK8zNpuDnuvH4qL4lIaoplgugOrAmbLwuWNXapmc01sxfMLL8p+5rZODMrMbOS8vLy5oo7KtMXl1NRU8eFA1VeEpHUFMsEEelntTeafxkocPeBwD+BJ5qwL+7+iLsXu3txly5dDivYplJ5SURSXSwTRBmQHzafB6wN38DdN7t7VTD7e2BItPvGU0V1ncpLIpLyYpkgZgG9zazQzFoBVwBTwjcws/D6zCXAomB6GjDKzDqZWSdgVLAsIc1fYAcAAAfFSURBVExfsjFUXlLvJRFJYTHrxeTutWZ2K6ETewYw2d0XmNm9QIm7TwFuM7NLgFpgC3BdsO8WM/s+oSQDcK+7b4lVrE31ytxQeUkjx4lIKotZggBw91eBVxstuztsegIw4QD7TgYmxzK+z6KhvPTFwd3JzNB9hiKSunSGayKVl0QkXShBNJHKSyKSLpQgmiC895LKSyKS6nSWawKVl0QknShBNMEr89aR01blJRFJD0oQUaqoruPNRRsZ3V/lJRFJDzrTRUnlJRFJN0oQUVJ5SUTSjRJEFBrKS+epvCQiaURnuyg0lJcuUnlJRNKIEkQUVF4SkXSkBHEIKi+JSLrSGe8QVF4SkXSlBHEIKi+JSLpSgjgIlZdEJJ3prHcQb+nmOBFJY0oQB/H3oLx0ispLIpKGlCAOQOUlEUl3OvMdgMpLIpLulCAO4O/z1tFZ5SURSWNKEBHo0d4iIkoQEam8JCKiBBHRKyoviYgoQTRWUV3HGyoviYgoQTSm8pKISIgSRCMqL4mIhChBhKmoruPNxRs5r5/KSyIiOguGeWvJRvZU13HRQJWXRESUIMKovCQiso8SREDlJRGR/elMGFB5SURkf0oQAZWXRET2pwQBVNaovCQi0pjOhuwrL+nmOBGRfZQggL/PDZWXTu2p8pKISIO0TxAqL4mIRJb2Z8QdFTWc07crY07qFu9QREQSSkwThJmNNrMlZrbMzMYfZLvLzMzNrDiYLzCzCjObE7wejlWMR3fI4sGxJ3Nqz5xYvYWISFLKjNWBzSwDmAScC5QBs8xsirsvbLRde+A24MNGh1ju7ifFKj4RETm4WF5BDAOWuXupu1cDzwJjImz3feB+oDKGsYiISBPFMkF0B9aEzZcFy/Yys5OBfHf/e4T9C83s32b2tpmdEekNzGycmZWYWUl5eXmzBS4iIrFNEBZhme9dadYC+CVwe4Tt1gE93P1k4NvAn8ysw6cO5v6Iuxe7e3GXLl2aKWwREYHYJogyID9sPg9YGzbfHugPvGVmK4FTgSlmVuzuVe6+GcDdZwPLgRNiGKuIiDQSywQxC+htZoVm1gq4ApjSsNLdt7t7rrsXuHsBMAO4xN1LzKxL0MiNmfUEegOlMYxVREQaiVkvJnevNbNbgWlABjDZ3ReY2b1AibtPOcjuZwL3mlktUAfc7O5bYhWriIh8mrn7obdKAsXFxV5SUhLvMEREkoqZzXb34ojrUiVBmFk5sOowDpELbGqmcJKdvov96fvYn76PfVLhuzjO3SP28kmZBHG4zKzkQFk03ei72J++j/3p+9gn1b+LtH8Wk4iIRKYEISIiESlB7PNIvANIIPou9qfvY3/6PvZJ6e9CbRAiIhKRriBERCQiJQgREYko7RNEtIMapQMzyzez6Wa2yMwWmNk34h1TvJlZRvBU4UhPHE4rZtbRzF4ws8XB/yPD4x1TPJnZt4J/J/PN7Bkzy4p3TM0trRNE2KBG5wNFwFgzK4pvVHFVC9zu7n0JPTzxljT/PgC+ASyKdxAJ4gFgqrufCAwijb8XM+tOaKCzYnfvT+hxQlfEN6rml9YJgugHNUoL7r7O3T8KpncSOgF0P/heqcvM8oALgUfjHUu8BY/bPxN4DMDdq919W3yjirtMINvMMoE27P+06pSQ7gnikIMapSszKwBO5tNDwaaTXwH/C9THO5AE0BMoBx4PSm6PmlnbeAcVL+7+CfAzYDWh8Wu2u/s/4htV80v3BHHQQY3SlZm1A/4CfNPdd8Q7nngws4uAjcF4JBL6tTwYeCgYyGs3kLZtdmbWiVC1oRDoBrQ1s6vjG1XzS/cEcahBjdKOmbUklByedvcX4x1PHI0ALgkGs3oWONvMnopvSHFVBpS5e8MV5QuEEka6OgdY4e7l7l4DvAicFueYml26J4iDDmqUbszMCNWYF7n7L+IdTzy5+wR3zwsGs7oCeNPdU+4XYrTcfT2wxsz6BItGAgvjGFK8rQZONbM2wb+bkaRgo33MBgxKBgca1CjOYcXTCOC/gXlmNidYdqe7vxrHmCRxfB14OvgxVQpcH+d44sbdPzSzF4CPCPX++zcp+NgNPWpDREQiSvcSk4iIHIAShIiIRKQEISIiESlBiIhIREoQIiISkRKESBOYWZ2ZzQl7NdvdxGZWYGbzm+t4Iocrre+DEPkMKtz9pHgHIXIk6ApCpBmY2Uoz+4mZzQxexwfLjzOzN8xsbvC3R7C8q5n91cz+E7waHtOQYWa/D8YZ+IeZZcftQ0naU4IQaZrsRiWmy8PW7XD3YcBvCD0JlmD6SXcfCDwNPBgsfxB4290HEXqmUcMd/L2BSe7eD9gGXBrjzyNyQLqTWqQJzGyXu7eLsHwlcLa7lwYPPFzv7jlmtgk41t1rguXr3D3XzMqBPHevCjtGAfC6u/cO5v8PaOnuP4j9JxP5NF1BiDQfP8D0gbaJpCpsug61E0ocKUGINJ/Lw/5+EEy/z76hKK8C3gum3wC+BnvHve5wpIIUiZZ+nYg0TXbYk24hNEZzQ1fX1mb2IaEfXmODZbcBk83sO4RGZGt4Auo3gEfM7KuErhS+RmhkMpGEoTYIkWYQtEEUu/umeMci0lxUYhIRkYh0BSEiIhHpCkJERCJSghARkYiUIEREJCIlCBERiUgJQkREIvr/vs1zggQtVIoAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "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": [
    "plt.plot(history.history['accuracy'])\n",
    "plt.plot(history.history['loss'])\n",
    "plt.title('Model accuracy, Loss')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Valid'], loc='upper left')\n",
    "plt.show()\n",
    "\n",
    "\n",
    "plt.title('Model loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Valid'], loc='upper left')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Validation:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy on Test Data :- \n",
      "\t loss: 0.2104390954429453   mean_absolute_error: 0.42969244718551636   accuracy: 0.6969696879386902\n"
     ]
    }
   ],
   "source": [
    "loss, mean_absolute_error, accuracy =model1.evaluate(X2, y2, batch_size=32, verbose=2)\n",
    "print('Accuracy on Test Data :- ')\n",
    "print('\\t loss:',loss , '  mean_absolute_error:',mean_absolute_error, '  accuracy:',accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__For the baseline model, Loss is Higher and Accuracy on validation data is 63%__"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Activation function"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "___Hyperparameters:___<br>\n",
    "filters=128<br>\n",
    "kernel_size=3<br>\n",
    "Activation functions=  elu, softmax<br>\n",
    "loss=mean_squared_error<br>\n",
    "optimizer= adam<br>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "# define model\n",
    "model2 = Sequential()\n",
    "\n",
    "model2.add(Conv1D(filters=128, kernel_size=3, activation='elu', input_shape=(n_steps, n_features)))\n",
    "model2.add(MaxPooling1D(pool_size=2))\n",
    "model2.add(BatchNormalization())\n",
    "\n",
    "model2.add(Flatten())\n",
    "\n",
    "model2.add(Dropout(0.25))\n",
    "model2.add(Dense(20, activation='softmax'))\n",
    "model2.add(Dropout(0.25))\n",
    "model2.add(Dense(1))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      " - 0s - loss: 0.3031 - mean_absolute_error: 0.4801 - accuracy: 0.5564\n",
      "Epoch 2/10\n",
      " - 0s - loss: 0.2518 - mean_absolute_error: 0.4515 - accuracy: 0.6116\n",
      "Epoch 3/10\n",
      " - 0s - loss: 0.2470 - mean_absolute_error: 0.4406 - accuracy: 0.6221\n",
      "Epoch 4/10\n",
      " - 0s - loss: 0.2340 - mean_absolute_error: 0.4341 - accuracy: 0.6331\n",
      "Epoch 5/10\n",
      " - 0s - loss: 0.2293 - mean_absolute_error: 0.4309 - accuracy: 0.6355\n",
      "Epoch 6/10\n",
      " - 0s - loss: 0.2236 - mean_absolute_error: 0.4220 - accuracy: 0.6477\n",
      "Epoch 7/10\n",
      " - 0s - loss: 0.2238 - mean_absolute_error: 0.4294 - accuracy: 0.6529\n",
      "Epoch 8/10\n",
      " - 0s - loss: 0.2196 - mean_absolute_error: 0.4231 - accuracy: 0.6552\n",
      "Epoch 9/10\n",
      " - 0s - loss: 0.2163 - mean_absolute_error: 0.4166 - accuracy: 0.6634\n",
      "Epoch 10/10\n",
      " - 0s - loss: 0.2103 - mean_absolute_error: 0.4116 - accuracy: 0.6907\n"
     ]
    }
   ],
   "source": [
    "adam=optimizers.Adam(learning_rate=0.4, beta_1=0.9, beta_2=0.999, amsgrad=False)\n",
    "model2.compile(optimizer='adam', loss='mean_squared_error',metrics=[metrics.mae, 'accuracy'])\n",
    "history=model2.fit(X1, y1, batch_size=32, epochs=10, verbose=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "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"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['accuracy'])\n",
    "plt.title('Model accuracy')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Valid'], loc='upper left')\n",
    "plt.show()\n",
    "\n",
    "plt.plot(history.history['loss'])\n",
    "plt.title('Model loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Valid'], loc='upper left')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Validation:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy on Test Data :- \n",
      "\t loss: 0.2764722759073431   mean_absolute_error: 0.523048996925354   accuracy: 0.3030303120613098\n"
     ]
    }
   ],
   "source": [
    "loss, mean_absolute_error, accuracy =model2.evaluate(X2, y2, batch_size=32, verbose=2)\n",
    "print('Accuracy on Test Data :- ')\n",
    "print('\\t loss:',loss , '  mean_absolute_error:',mean_absolute_error, '  accuracy:',accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__After changing the activation functions in the current network, loss get reduced but with decrease in accuracy on validation set which is 30%.__  "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Cost function / loss"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "___Hyperparameters:___<br>\n",
    "filters=128<br>\n",
    "kernel_size=3<br>\n",
    "Activation functions=  elu, softmax<br>\n",
    "loss= hinge, logcosh<br>\n",
    "optimizer= adam<br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__1. Cost function= hinge:__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      " - 0s - loss: 0.8280 - mean_absolute_error: 0.3903 - accuracy: 0.6785\n",
      "Epoch 2/10\n",
      " - 0s - loss: 0.7772 - mean_absolute_error: 0.4316 - accuracy: 0.6779\n",
      "Epoch 3/10\n",
      " - 0s - loss: 0.7588 - mean_absolute_error: 0.5018 - accuracy: 0.6605\n",
      "Epoch 4/10\n",
      " - 0s - loss: 0.7559 - mean_absolute_error: 0.5383 - accuracy: 0.5471\n",
      "Epoch 5/10\n",
      " - 0s - loss: 0.7499 - mean_absolute_error: 0.5934 - accuracy: 0.4378\n",
      "Epoch 6/10\n",
      " - 0s - loss: 0.7184 - mean_absolute_error: 0.6502 - accuracy: 0.3791\n",
      "Epoch 7/10\n",
      " - 0s - loss: 0.6962 - mean_absolute_error: 0.7026 - accuracy: 0.3599\n",
      "Epoch 8/10\n",
      " - 0s - loss: 0.6672 - mean_absolute_error: 0.7386 - accuracy: 0.3744\n",
      "Epoch 9/10\n",
      " - 0s - loss: 0.6864 - mean_absolute_error: 0.7716 - accuracy: 0.3860\n",
      "Epoch 10/10\n",
      " - 0s - loss: 0.6948 - mean_absolute_error: 0.7789 - accuracy: 0.3727\n"
     ]
    }
   ],
   "source": [
    "model2.compile(optimizer='adam', loss='hinge',metrics=[metrics.mae, 'accuracy'])\n",
    "history=model2.fit(X1, y1, batch_size=32, epochs=10, verbose=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "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"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['accuracy'])\n",
    "plt.title('Model accuracy')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Valid'], loc='upper left')\n",
    "plt.show()\n",
    "\n",
    "plt.plot(history.history['loss'])\n",
    "plt.title('Model loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Valid'], loc='upper left')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__Accuracy :__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy on Test Data :- \n",
      "\t loss: 0.6813373999162153   mean_absolute_error: 0.3783071041107178   accuracy: 0.6969696879386902\n"
     ]
    }
   ],
   "source": [
    "loss, mean_absolute_error, accuracy =model2.evaluate(X2, y2, batch_size=32, verbose=2)\n",
    "print('Accuracy on Test Data :- ')\n",
    "print('\\t loss:',loss , '  mean_absolute_error:',mean_absolute_error, '  accuracy:',accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__2. Cost function= logcosh:__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      " - 0s - loss: 0.2246 - mean_absolute_error: 0.5722 - accuracy: 0.5023\n",
      "Epoch 2/10\n",
      " - 0s - loss: 0.1104 - mean_absolute_error: 0.4184 - accuracy: 0.6401\n",
      "Epoch 3/10\n",
      " - 0s - loss: 0.1074 - mean_absolute_error: 0.4146 - accuracy: 0.6419\n",
      "Epoch 4/10\n",
      " - 0s - loss: 0.1061 - mean_absolute_error: 0.4171 - accuracy: 0.6535\n",
      "Epoch 5/10\n",
      " - 0s - loss: 0.1049 - mean_absolute_error: 0.4132 - accuracy: 0.6610\n",
      "Epoch 6/10\n",
      " - 0s - loss: 0.1005 - mean_absolute_error: 0.4059 - accuracy: 0.6762\n",
      "Epoch 7/10\n",
      " - 0s - loss: 0.1020 - mean_absolute_error: 0.4086 - accuracy: 0.6750\n",
      "Epoch 8/10\n",
      " - 0s - loss: 0.1022 - mean_absolute_error: 0.4060 - accuracy: 0.6680\n",
      "Epoch 9/10\n",
      " - 0s - loss: 0.1025 - mean_absolute_error: 0.4150 - accuracy: 0.6680\n",
      "Epoch 10/10\n",
      " - 0s - loss: 0.0997 - mean_absolute_error: 0.4058 - accuracy: 0.6849\n"
     ]
    }
   ],
   "source": [
    "model2.compile(optimizer='adam', loss='logcosh',metrics=[metrics.mae, 'accuracy'])\n",
    "history=model2.fit(X1, y1, batch_size=32, epochs=10, verbose=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "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"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['accuracy'])\n",
    "plt.title('Model accuracy')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Valid'], loc='upper left')\n",
    "plt.show()\n",
    "\n",
    "plt.plot(history.history['loss'])\n",
    "plt.title('Model loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Valid'], loc='upper left')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__Accuracy :__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy on Test Data :- \n",
      "\t loss: 0.10073961633624452   mean_absolute_error: 0.43285611271858215   accuracy: 0.6969696879386902\n"
     ]
    }
   ],
   "source": [
    "loss, mean_absolute_error, accuracy =model2.evaluate(X2, y2, batch_size=32, verbose=2)\n",
    "print('Accuracy on Test Data :- ')\n",
    "print('\\t loss:',loss , '  mean_absolute_error:',mean_absolute_error, '  accuracy:',accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__For the cost function, logcosh is better than hinge for better accuracy.__ "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Epochs"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "___Hyperparameters:___<br>\n",
    "filters=128<br>\n",
    "kernel_size=3<br>\n",
    "Activation functions=  elu, softmax<br>\n",
    "loss= logcosh<br>\n",
    "optimizer= adam<br>\n",
    "epochs=200 <br>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      " - 0s - loss: 0.0999 - mean_absolute_error: 0.4039 - accuracy: 0.6767\n",
      "Epoch 2/200\n",
      " - 0s - loss: 0.1023 - mean_absolute_error: 0.4106 - accuracy: 0.6634\n",
      "Epoch 3/200\n",
      " - 0s - loss: 0.0962 - mean_absolute_error: 0.3995 - accuracy: 0.6953\n",
      "Epoch 4/200\n",
      " - 0s - loss: 0.0974 - mean_absolute_error: 0.3998 - accuracy: 0.6773\n",
      "Epoch 5/200\n",
      " - 0s - loss: 0.0959 - mean_absolute_error: 0.3987 - accuracy: 0.6959\n",
      "Epoch 6/200\n",
      " - 0s - loss: 0.0977 - mean_absolute_error: 0.4032 - accuracy: 0.6924\n",
      "Epoch 7/200\n",
      " - 0s - loss: 0.0962 - mean_absolute_error: 0.3974 - accuracy: 0.6971\n",
      "Epoch 8/200\n",
      " - 0s - loss: 0.0941 - mean_absolute_error: 0.3946 - accuracy: 0.7029\n",
      "Epoch 9/200\n",
      " - 0s - loss: 0.0956 - mean_absolute_error: 0.3896 - accuracy: 0.6959\n",
      "Epoch 10/200\n",
      " - 0s - loss: 0.0956 - mean_absolute_error: 0.3949 - accuracy: 0.6942\n",
      "Epoch 11/200\n",
      " - 0s - loss: 0.0961 - mean_absolute_error: 0.3964 - accuracy: 0.7012\n",
      "Epoch 12/200\n",
      " - 0s - loss: 0.0953 - mean_absolute_error: 0.3893 - accuracy: 0.6965\n",
      "Epoch 13/200\n",
      " - 0s - loss: 0.0930 - mean_absolute_error: 0.3904 - accuracy: 0.7099\n",
      "Epoch 14/200\n",
      " - 0s - loss: 0.0952 - mean_absolute_error: 0.3931 - accuracy: 0.6971\n",
      "Epoch 15/200\n",
      " - 0s - loss: 0.0954 - mean_absolute_error: 0.3934 - accuracy: 0.6936\n",
      "Epoch 16/200\n",
      " - 0s - loss: 0.0927 - mean_absolute_error: 0.3870 - accuracy: 0.7105\n",
      "Epoch 17/200\n",
      " - 0s - loss: 0.0962 - mean_absolute_error: 0.3970 - accuracy: 0.7000\n",
      "Epoch 18/200\n",
      " - 0s - loss: 0.0951 - mean_absolute_error: 0.3962 - accuracy: 0.7017\n",
      "Epoch 19/200\n",
      " - 0s - loss: 0.0951 - mean_absolute_error: 0.3934 - accuracy: 0.7041\n",
      "Epoch 20/200\n",
      " - 0s - loss: 0.0921 - mean_absolute_error: 0.3829 - accuracy: 0.7134\n",
      "Epoch 21/200\n",
      " - 0s - loss: 0.0927 - mean_absolute_error: 0.3858 - accuracy: 0.7122\n",
      "Epoch 22/200\n",
      " - 0s - loss: 0.0945 - mean_absolute_error: 0.3858 - accuracy: 0.7000\n",
      "Epoch 23/200\n",
      " - 0s - loss: 0.0943 - mean_absolute_error: 0.3923 - accuracy: 0.6994\n",
      "Epoch 24/200\n",
      " - 0s - loss: 0.0939 - mean_absolute_error: 0.3927 - accuracy: 0.6988\n",
      "Epoch 25/200\n",
      " - 0s - loss: 0.0926 - mean_absolute_error: 0.3883 - accuracy: 0.7064\n",
      "Epoch 26/200\n",
      " - 0s - loss: 0.0945 - mean_absolute_error: 0.3919 - accuracy: 0.7070\n",
      "Epoch 27/200\n",
      " - 0s - loss: 0.0932 - mean_absolute_error: 0.3880 - accuracy: 0.7052\n",
      "Epoch 28/200\n",
      " - 0s - loss: 0.0930 - mean_absolute_error: 0.3871 - accuracy: 0.7093\n",
      "Epoch 29/200\n",
      " - 0s - loss: 0.0916 - mean_absolute_error: 0.3839 - accuracy: 0.7151\n",
      "Epoch 30/200\n",
      " - 0s - loss: 0.0928 - mean_absolute_error: 0.3829 - accuracy: 0.7081\n",
      "Epoch 31/200\n",
      " - 0s - loss: 0.0933 - mean_absolute_error: 0.3884 - accuracy: 0.7029\n",
      "Epoch 32/200\n",
      " - 0s - loss: 0.0912 - mean_absolute_error: 0.3795 - accuracy: 0.7058\n",
      "Epoch 33/200\n",
      " - 0s - loss: 0.0925 - mean_absolute_error: 0.3831 - accuracy: 0.7058\n",
      "Epoch 34/200\n",
      " - 0s - loss: 0.0921 - mean_absolute_error: 0.3817 - accuracy: 0.7052\n",
      "Epoch 35/200\n",
      " - 0s - loss: 0.0929 - mean_absolute_error: 0.3842 - accuracy: 0.7076\n",
      "Epoch 36/200\n",
      " - 0s - loss: 0.0930 - mean_absolute_error: 0.3866 - accuracy: 0.7029\n",
      "Epoch 37/200\n",
      " - 0s - loss: 0.0929 - mean_absolute_error: 0.3857 - accuracy: 0.7116\n",
      "Epoch 38/200\n",
      " - 0s - loss: 0.0925 - mean_absolute_error: 0.3848 - accuracy: 0.7076\n",
      "Epoch 39/200\n",
      " - 0s - loss: 0.0930 - mean_absolute_error: 0.3880 - accuracy: 0.7081\n",
      "Epoch 40/200\n",
      " - 0s - loss: 0.0916 - mean_absolute_error: 0.3812 - accuracy: 0.7110\n",
      "Epoch 41/200\n",
      " - 0s - loss: 0.0919 - mean_absolute_error: 0.3827 - accuracy: 0.7151\n",
      "Epoch 42/200\n",
      " - 0s - loss: 0.0939 - mean_absolute_error: 0.3940 - accuracy: 0.7116\n",
      "Epoch 43/200\n",
      " - 0s - loss: 0.0934 - mean_absolute_error: 0.3896 - accuracy: 0.6965\n",
      "Epoch 44/200\n",
      " - 0s - loss: 0.0928 - mean_absolute_error: 0.3857 - accuracy: 0.7017\n",
      "Epoch 45/200\n",
      " - 0s - loss: 0.0923 - mean_absolute_error: 0.3839 - accuracy: 0.7052\n",
      "Epoch 46/200\n",
      " - 0s - loss: 0.0915 - mean_absolute_error: 0.3786 - accuracy: 0.7116\n",
      "Epoch 47/200\n",
      " - 0s - loss: 0.0925 - mean_absolute_error: 0.3827 - accuracy: 0.7099\n",
      "Epoch 48/200\n",
      " - 0s - loss: 0.0924 - mean_absolute_error: 0.3854 - accuracy: 0.7110\n",
      "Epoch 49/200\n",
      " - 0s - loss: 0.0925 - mean_absolute_error: 0.3863 - accuracy: 0.7006\n",
      "Epoch 50/200\n",
      " - 0s - loss: 0.0912 - mean_absolute_error: 0.3840 - accuracy: 0.7116\n",
      "Epoch 51/200\n",
      " - 0s - loss: 0.0912 - mean_absolute_error: 0.3781 - accuracy: 0.7151\n",
      "Epoch 52/200\n",
      " - 0s - loss: 0.0908 - mean_absolute_error: 0.3796 - accuracy: 0.7151\n",
      "Epoch 53/200\n",
      " - 0s - loss: 0.0920 - mean_absolute_error: 0.3823 - accuracy: 0.7140\n",
      "Epoch 54/200\n",
      " - 0s - loss: 0.0912 - mean_absolute_error: 0.3817 - accuracy: 0.7145\n",
      "Epoch 55/200\n",
      " - 0s - loss: 0.0907 - mean_absolute_error: 0.3799 - accuracy: 0.7134\n",
      "Epoch 56/200\n",
      " - 0s - loss: 0.0924 - mean_absolute_error: 0.3799 - accuracy: 0.7041\n",
      "Epoch 57/200\n",
      " - 0s - loss: 0.0922 - mean_absolute_error: 0.3842 - accuracy: 0.7140\n",
      "Epoch 58/200\n",
      " - 0s - loss: 0.0901 - mean_absolute_error: 0.3835 - accuracy: 0.7163\n",
      "Epoch 59/200\n",
      " - 0s - loss: 0.0899 - mean_absolute_error: 0.3749 - accuracy: 0.7221\n",
      "Epoch 60/200\n",
      " - 0s - loss: 0.0925 - mean_absolute_error: 0.3836 - accuracy: 0.7105\n",
      "Epoch 61/200\n",
      " - 0s - loss: 0.0921 - mean_absolute_error: 0.3813 - accuracy: 0.7140\n",
      "Epoch 62/200\n",
      " - 0s - loss: 0.0910 - mean_absolute_error: 0.3793 - accuracy: 0.7145\n",
      "Epoch 63/200\n",
      " - 0s - loss: 0.0907 - mean_absolute_error: 0.3791 - accuracy: 0.7140\n",
      "Epoch 64/200\n",
      " - 0s - loss: 0.0910 - mean_absolute_error: 0.3810 - accuracy: 0.7180\n",
      "Epoch 65/200\n",
      " - 0s - loss: 0.0917 - mean_absolute_error: 0.3833 - accuracy: 0.7122\n",
      "Epoch 66/200\n",
      " - 0s - loss: 0.0928 - mean_absolute_error: 0.3863 - accuracy: 0.7081\n",
      "Epoch 67/200\n",
      " - 0s - loss: 0.0901 - mean_absolute_error: 0.3754 - accuracy: 0.7198\n",
      "Epoch 68/200\n",
      " - 0s - loss: 0.0907 - mean_absolute_error: 0.3781 - accuracy: 0.7221\n",
      "Epoch 69/200\n",
      " - 0s - loss: 0.0915 - mean_absolute_error: 0.3789 - accuracy: 0.7116\n",
      "Epoch 70/200\n",
      " - 0s - loss: 0.0914 - mean_absolute_error: 0.3838 - accuracy: 0.7151\n",
      "Epoch 71/200\n",
      " - 0s - loss: 0.0908 - mean_absolute_error: 0.3817 - accuracy: 0.7151\n",
      "Epoch 72/200\n",
      " - 0s - loss: 0.0914 - mean_absolute_error: 0.3804 - accuracy: 0.7099\n",
      "Epoch 73/200\n",
      " - 0s - loss: 0.0919 - mean_absolute_error: 0.3857 - accuracy: 0.7110\n",
      "Epoch 74/200\n",
      " - 0s - loss: 0.0907 - mean_absolute_error: 0.3788 - accuracy: 0.7244\n",
      "Epoch 75/200\n",
      " - 0s - loss: 0.0912 - mean_absolute_error: 0.3828 - accuracy: 0.7163\n",
      "Epoch 76/200\n",
      " - 0s - loss: 0.0906 - mean_absolute_error: 0.3791 - accuracy: 0.7174\n",
      "Epoch 77/200\n",
      " - 0s - loss: 0.0912 - mean_absolute_error: 0.3800 - accuracy: 0.7157\n",
      "Epoch 78/200\n",
      " - 0s - loss: 0.0893 - mean_absolute_error: 0.3739 - accuracy: 0.7209\n",
      "Epoch 79/200\n",
      " - 0s - loss: 0.0897 - mean_absolute_error: 0.3738 - accuracy: 0.7262\n",
      "Epoch 80/200\n",
      " - 0s - loss: 0.0886 - mean_absolute_error: 0.3717 - accuracy: 0.7256\n",
      "Epoch 81/200\n",
      " - 0s - loss: 0.0912 - mean_absolute_error: 0.3759 - accuracy: 0.7140\n",
      "Epoch 82/200\n",
      " - 0s - loss: 0.0926 - mean_absolute_error: 0.3874 - accuracy: 0.7076\n",
      "Epoch 83/200\n",
      " - 0s - loss: 0.0893 - mean_absolute_error: 0.3740 - accuracy: 0.7186\n",
      "Epoch 84/200\n",
      " - 0s - loss: 0.0908 - mean_absolute_error: 0.3800 - accuracy: 0.7238\n",
      "Epoch 85/200\n",
      " - 0s - loss: 0.0909 - mean_absolute_error: 0.3804 - accuracy: 0.7163\n",
      "Epoch 86/200\n",
      " - 0s - loss: 0.0894 - mean_absolute_error: 0.3737 - accuracy: 0.7157\n",
      "Epoch 87/200\n",
      " - 0s - loss: 0.0918 - mean_absolute_error: 0.3812 - accuracy: 0.7151\n",
      "Epoch 88/200\n",
      " - 0s - loss: 0.0913 - mean_absolute_error: 0.3790 - accuracy: 0.7134\n",
      "Epoch 89/200\n",
      " - 0s - loss: 0.0920 - mean_absolute_error: 0.3839 - accuracy: 0.7151\n",
      "Epoch 90/200\n",
      " - 0s - loss: 0.0908 - mean_absolute_error: 0.3760 - accuracy: 0.7163\n",
      "Epoch 91/200\n",
      " - 0s - loss: 0.0918 - mean_absolute_error: 0.3784 - accuracy: 0.7116\n",
      "Epoch 92/200\n",
      " - 0s - loss: 0.0904 - mean_absolute_error: 0.3806 - accuracy: 0.7215\n",
      "Epoch 93/200\n",
      " - 0s - loss: 0.0910 - mean_absolute_error: 0.3792 - accuracy: 0.7186\n",
      "Epoch 94/200\n",
      " - 0s - loss: 0.0911 - mean_absolute_error: 0.3765 - accuracy: 0.7169\n",
      "Epoch 95/200\n",
      " - 0s - loss: 0.0895 - mean_absolute_error: 0.3813 - accuracy: 0.7203\n",
      "Epoch 96/200\n",
      " - 0s - loss: 0.0901 - mean_absolute_error: 0.3784 - accuracy: 0.7233\n",
      "Epoch 97/200\n",
      " - 0s - loss: 0.0883 - mean_absolute_error: 0.3729 - accuracy: 0.7256\n",
      "Epoch 98/200\n",
      " - 0s - loss: 0.0904 - mean_absolute_error: 0.3777 - accuracy: 0.7140\n",
      "Epoch 99/200\n",
      " - 0s - loss: 0.0903 - mean_absolute_error: 0.3730 - accuracy: 0.7238\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 100/200\n",
      " - 0s - loss: 0.0899 - mean_absolute_error: 0.3750 - accuracy: 0.7163\n",
      "Epoch 101/200\n",
      " - 0s - loss: 0.0890 - mean_absolute_error: 0.3724 - accuracy: 0.7203\n",
      "Epoch 102/200\n",
      " - 0s - loss: 0.0910 - mean_absolute_error: 0.3784 - accuracy: 0.7192\n",
      "Epoch 103/200\n",
      " - 0s - loss: 0.0909 - mean_absolute_error: 0.3805 - accuracy: 0.7186\n",
      "Epoch 104/200\n",
      " - 0s - loss: 0.0907 - mean_absolute_error: 0.3801 - accuracy: 0.7169\n",
      "Epoch 105/200\n",
      " - 0s - loss: 0.0890 - mean_absolute_error: 0.3716 - accuracy: 0.7238\n",
      "Epoch 106/200\n",
      " - 0s - loss: 0.0886 - mean_absolute_error: 0.3696 - accuracy: 0.7203\n",
      "Epoch 107/200\n",
      " - 0s - loss: 0.0900 - mean_absolute_error: 0.3767 - accuracy: 0.7145\n",
      "Epoch 108/200\n",
      " - 0s - loss: 0.0915 - mean_absolute_error: 0.3812 - accuracy: 0.7070\n",
      "Epoch 109/200\n",
      " - 0s - loss: 0.0890 - mean_absolute_error: 0.3769 - accuracy: 0.7198\n",
      "Epoch 110/200\n",
      " - 0s - loss: 0.0886 - mean_absolute_error: 0.3701 - accuracy: 0.7227\n",
      "Epoch 111/200\n",
      " - 0s - loss: 0.0908 - mean_absolute_error: 0.3804 - accuracy: 0.7180\n",
      "Epoch 112/200\n",
      " - 0s - loss: 0.0899 - mean_absolute_error: 0.3737 - accuracy: 0.7163\n",
      "Epoch 113/200\n",
      " - 0s - loss: 0.0894 - mean_absolute_error: 0.3767 - accuracy: 0.7227\n",
      "Epoch 114/200\n",
      " - 0s - loss: 0.0900 - mean_absolute_error: 0.3786 - accuracy: 0.7186\n",
      "Epoch 115/200\n",
      " - 0s - loss: 0.0903 - mean_absolute_error: 0.3774 - accuracy: 0.7186\n",
      "Epoch 116/200\n",
      " - 0s - loss: 0.0907 - mean_absolute_error: 0.3804 - accuracy: 0.7163\n",
      "Epoch 117/200\n",
      " - 0s - loss: 0.0899 - mean_absolute_error: 0.3733 - accuracy: 0.7221\n",
      "Epoch 118/200\n",
      " - 0s - loss: 0.0887 - mean_absolute_error: 0.3712 - accuracy: 0.7122\n",
      "Epoch 119/200\n",
      " - 0s - loss: 0.0900 - mean_absolute_error: 0.3737 - accuracy: 0.7174\n",
      "Epoch 120/200\n",
      " - 0s - loss: 0.0891 - mean_absolute_error: 0.3733 - accuracy: 0.7174\n",
      "Epoch 121/200\n",
      " - 0s - loss: 0.0882 - mean_absolute_error: 0.3670 - accuracy: 0.7215\n",
      "Epoch 122/200\n",
      " - 0s - loss: 0.0883 - mean_absolute_error: 0.3737 - accuracy: 0.7198\n",
      "Epoch 123/200\n",
      " - 0s - loss: 0.0891 - mean_absolute_error: 0.3706 - accuracy: 0.7221\n",
      "Epoch 124/200\n",
      " - 0s - loss: 0.0893 - mean_absolute_error: 0.3697 - accuracy: 0.7215\n",
      "Epoch 125/200\n",
      " - 0s - loss: 0.0892 - mean_absolute_error: 0.3752 - accuracy: 0.7140\n",
      "Epoch 126/200\n",
      " - 0s - loss: 0.0897 - mean_absolute_error: 0.3764 - accuracy: 0.7174\n",
      "Epoch 127/200\n",
      " - 0s - loss: 0.0905 - mean_absolute_error: 0.3782 - accuracy: 0.7192\n",
      "Epoch 128/200\n",
      " - 0s - loss: 0.0880 - mean_absolute_error: 0.3704 - accuracy: 0.7238\n",
      "Epoch 129/200\n",
      " - 0s - loss: 0.0888 - mean_absolute_error: 0.3738 - accuracy: 0.7140\n",
      "Epoch 130/200\n",
      " - 0s - loss: 0.0896 - mean_absolute_error: 0.3748 - accuracy: 0.7122\n",
      "Epoch 131/200\n",
      " - 0s - loss: 0.0893 - mean_absolute_error: 0.3718 - accuracy: 0.7145\n",
      "Epoch 132/200\n",
      " - 0s - loss: 0.0887 - mean_absolute_error: 0.3712 - accuracy: 0.7192\n",
      "Epoch 133/200\n",
      " - 0s - loss: 0.0902 - mean_absolute_error: 0.3759 - accuracy: 0.7140\n",
      "Epoch 134/200\n",
      " - 0s - loss: 0.0896 - mean_absolute_error: 0.3749 - accuracy: 0.7174\n",
      "Epoch 135/200\n",
      " - 0s - loss: 0.0893 - mean_absolute_error: 0.3760 - accuracy: 0.7209\n",
      "Epoch 136/200\n",
      " - 0s - loss: 0.0888 - mean_absolute_error: 0.3719 - accuracy: 0.7151\n",
      "Epoch 137/200\n",
      " - 0s - loss: 0.0885 - mean_absolute_error: 0.3720 - accuracy: 0.7169\n",
      "Epoch 138/200\n",
      " - 0s - loss: 0.0908 - mean_absolute_error: 0.3756 - accuracy: 0.7087\n",
      "Epoch 139/200\n",
      " - 0s - loss: 0.0894 - mean_absolute_error: 0.3761 - accuracy: 0.7262\n",
      "Epoch 140/200\n",
      " - 0s - loss: 0.0884 - mean_absolute_error: 0.3697 - accuracy: 0.7151\n",
      "Epoch 141/200\n",
      " - 0s - loss: 0.0889 - mean_absolute_error: 0.3735 - accuracy: 0.7151\n",
      "Epoch 142/200\n",
      " - 0s - loss: 0.0886 - mean_absolute_error: 0.3714 - accuracy: 0.7238\n",
      "Epoch 143/200\n",
      " - 0s - loss: 0.0893 - mean_absolute_error: 0.3719 - accuracy: 0.7209\n",
      "Epoch 144/200\n",
      " - 0s - loss: 0.0883 - mean_absolute_error: 0.3723 - accuracy: 0.7145\n",
      "Epoch 145/200\n",
      " - 0s - loss: 0.0890 - mean_absolute_error: 0.3726 - accuracy: 0.7105\n",
      "Epoch 146/200\n",
      " - 0s - loss: 0.0881 - mean_absolute_error: 0.3680 - accuracy: 0.7262\n",
      "Epoch 147/200\n",
      " - 0s - loss: 0.0882 - mean_absolute_error: 0.3693 - accuracy: 0.7250\n",
      "Epoch 148/200\n",
      " - 0s - loss: 0.0900 - mean_absolute_error: 0.3725 - accuracy: 0.7198\n",
      "Epoch 149/200\n",
      " - 0s - loss: 0.0879 - mean_absolute_error: 0.3727 - accuracy: 0.7285\n",
      "Epoch 150/200\n",
      " - 0s - loss: 0.0883 - mean_absolute_error: 0.3675 - accuracy: 0.7256\n",
      "Epoch 151/200\n",
      " - 0s - loss: 0.0878 - mean_absolute_error: 0.3690 - accuracy: 0.7285\n",
      "Epoch 152/200\n",
      " - 0s - loss: 0.0885 - mean_absolute_error: 0.3710 - accuracy: 0.7238\n",
      "Epoch 153/200\n",
      " - 0s - loss: 0.0889 - mean_absolute_error: 0.3699 - accuracy: 0.7122\n",
      "Epoch 154/200\n",
      " - 0s - loss: 0.0883 - mean_absolute_error: 0.3698 - accuracy: 0.7221\n",
      "Epoch 155/200\n",
      " - 0s - loss: 0.0892 - mean_absolute_error: 0.3740 - accuracy: 0.7192\n",
      "Epoch 156/200\n",
      " - 0s - loss: 0.0884 - mean_absolute_error: 0.3688 - accuracy: 0.7244\n",
      "Epoch 157/200\n",
      " - 0s - loss: 0.0880 - mean_absolute_error: 0.3699 - accuracy: 0.7215\n",
      "Epoch 158/200\n",
      " - 0s - loss: 0.0884 - mean_absolute_error: 0.3680 - accuracy: 0.7267\n",
      "Epoch 159/200\n",
      " - 0s - loss: 0.0895 - mean_absolute_error: 0.3716 - accuracy: 0.7174\n",
      "Epoch 160/200\n",
      " - 0s - loss: 0.0877 - mean_absolute_error: 0.3760 - accuracy: 0.7262\n",
      "Epoch 161/200\n",
      " - 0s - loss: 0.0874 - mean_absolute_error: 0.3634 - accuracy: 0.7238\n",
      "Epoch 162/200\n",
      " - 0s - loss: 0.0887 - mean_absolute_error: 0.3692 - accuracy: 0.7140\n",
      "Epoch 163/200\n",
      " - 0s - loss: 0.0870 - mean_absolute_error: 0.3646 - accuracy: 0.7267\n",
      "Epoch 164/200\n",
      " - 0s - loss: 0.0888 - mean_absolute_error: 0.3755 - accuracy: 0.7203\n",
      "Epoch 165/200\n",
      " - 0s - loss: 0.0895 - mean_absolute_error: 0.3678 - accuracy: 0.7174\n",
      "Epoch 166/200\n",
      " - 0s - loss: 0.0878 - mean_absolute_error: 0.3682 - accuracy: 0.7297\n",
      "Epoch 167/200\n",
      " - 0s - loss: 0.0886 - mean_absolute_error: 0.3735 - accuracy: 0.7174\n",
      "Epoch 168/200\n",
      " - 0s - loss: 0.0878 - mean_absolute_error: 0.3711 - accuracy: 0.7186\n",
      "Epoch 169/200\n",
      " - 0s - loss: 0.0874 - mean_absolute_error: 0.3678 - accuracy: 0.7244\n",
      "Epoch 170/200\n",
      " - 0s - loss: 0.0894 - mean_absolute_error: 0.3724 - accuracy: 0.7134\n",
      "Epoch 171/200\n",
      " - 0s - loss: 0.0876 - mean_absolute_error: 0.3723 - accuracy: 0.7250\n",
      "Epoch 172/200\n",
      " - 0s - loss: 0.0887 - mean_absolute_error: 0.3692 - accuracy: 0.7221\n",
      "Epoch 173/200\n",
      " - 0s - loss: 0.0881 - mean_absolute_error: 0.3728 - accuracy: 0.7244\n",
      "Epoch 174/200\n",
      " - 0s - loss: 0.0887 - mean_absolute_error: 0.3688 - accuracy: 0.7169\n",
      "Epoch 175/200\n",
      " - 0s - loss: 0.0888 - mean_absolute_error: 0.3760 - accuracy: 0.7256\n",
      "Epoch 176/200\n",
      " - 0s - loss: 0.0877 - mean_absolute_error: 0.3651 - accuracy: 0.7273\n",
      "Epoch 177/200\n",
      " - 0s - loss: 0.0892 - mean_absolute_error: 0.3752 - accuracy: 0.7128\n",
      "Epoch 178/200\n",
      " - 0s - loss: 0.0868 - mean_absolute_error: 0.3653 - accuracy: 0.7256\n",
      "Epoch 179/200\n",
      " - 0s - loss: 0.0881 - mean_absolute_error: 0.3665 - accuracy: 0.7215\n",
      "Epoch 180/200\n",
      " - 0s - loss: 0.0909 - mean_absolute_error: 0.3740 - accuracy: 0.7163\n",
      "Epoch 181/200\n",
      " - 0s - loss: 0.0887 - mean_absolute_error: 0.3713 - accuracy: 0.7285\n",
      "Epoch 182/200\n",
      " - 0s - loss: 0.0890 - mean_absolute_error: 0.3731 - accuracy: 0.7250\n",
      "Epoch 183/200\n",
      " - 0s - loss: 0.0895 - mean_absolute_error: 0.3761 - accuracy: 0.7221\n",
      "Epoch 184/200\n",
      " - 0s - loss: 0.0870 - mean_absolute_error: 0.3713 - accuracy: 0.7238\n",
      "Epoch 185/200\n",
      " - 0s - loss: 0.0879 - mean_absolute_error: 0.3654 - accuracy: 0.7198\n",
      "Epoch 186/200\n",
      " - 0s - loss: 0.0879 - mean_absolute_error: 0.3703 - accuracy: 0.7291\n",
      "Epoch 187/200\n",
      " - 0s - loss: 0.0872 - mean_absolute_error: 0.3656 - accuracy: 0.7279\n",
      "Epoch 188/200\n",
      " - 0s - loss: 0.0882 - mean_absolute_error: 0.3668 - accuracy: 0.7198\n",
      "Epoch 189/200\n",
      " - 0s - loss: 0.0895 - mean_absolute_error: 0.3755 - accuracy: 0.7140\n",
      "Epoch 190/200\n",
      " - 0s - loss: 0.0865 - mean_absolute_error: 0.3653 - accuracy: 0.7267\n",
      "Epoch 191/200\n",
      " - 0s - loss: 0.0897 - mean_absolute_error: 0.3724 - accuracy: 0.7180\n",
      "Epoch 192/200\n",
      " - 0s - loss: 0.0872 - mean_absolute_error: 0.3671 - accuracy: 0.7203\n",
      "Epoch 193/200\n",
      " - 0s - loss: 0.0873 - mean_absolute_error: 0.3684 - accuracy: 0.7215\n",
      "Epoch 194/200\n",
      " - 0s - loss: 0.0878 - mean_absolute_error: 0.3674 - accuracy: 0.7192\n",
      "Epoch 195/200\n",
      " - 0s - loss: 0.0885 - mean_absolute_error: 0.3736 - accuracy: 0.7238\n",
      "Epoch 196/200\n",
      " - 0s - loss: 0.0885 - mean_absolute_error: 0.3694 - accuracy: 0.7192\n",
      "Epoch 197/200\n",
      " - 0s - loss: 0.0870 - mean_absolute_error: 0.3679 - accuracy: 0.7174\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 198/200\n",
      " - 0s - loss: 0.0884 - mean_absolute_error: 0.3700 - accuracy: 0.7198\n",
      "Epoch 199/200\n",
      " - 0s - loss: 0.0881 - mean_absolute_error: 0.3721 - accuracy: 0.7203\n",
      "Epoch 200/200\n",
      " - 0s - loss: 0.0892 - mean_absolute_error: 0.3731 - accuracy: 0.7215\n"
     ]
    }
   ],
   "source": [
    "model2.compile(optimizer='adam', loss='logcosh',metrics=[metrics.mae, 'accuracy'])\n",
    "history=model2.fit(X1, y1, batch_size=32, epochs=200, verbose=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "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"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['accuracy'])\n",
    "plt.title('Model accuracy')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Valid'], loc='upper left')\n",
    "plt.show()\n",
    "\n",
    "plt.plot(history.history['loss'])\n",
    "plt.title('Model loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Valid'], loc='upper left')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__Accuracy :__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy on Test Data :- \n",
      "\t loss: 0.11469146699616403   mean_absolute_error: 0.47265517711639404   accuracy: 0.6666666865348816\n"
     ]
    }
   ],
   "source": [
    "loss, mean_absolute_error, accuracy =model2.evaluate(X2, y2, batch_size=32, verbose=2)\n",
    "print('Accuracy on Test Data :- ')\n",
    "print('\\t loss:',loss , '  mean_absolute_error:',mean_absolute_error, '  accuracy:',accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__After increasing number of epochs in the model, accuracy increased and loss decreased.__"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Gradient estimation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "___Hyperparameters:___<br>\n",
    "filters=128<br>\n",
    "kernel_size=3<br>\n",
    "Activation functions=  elu, softmax<br>\n",
    "loss= logcosh<br>\n",
    "optimizer= RMSprop, Adagrad<br>\n",
    "epochs=200 <br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__1. optimizers : RMSprop__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      " - 0s - loss: 0.0897 - mean_absolute_error: 0.3732 - accuracy: 0.7192\n",
      "Epoch 2/200\n",
      " - 0s - loss: 0.0894 - mean_absolute_error: 0.3693 - accuracy: 0.7209\n",
      "Epoch 3/200\n",
      " - 0s - loss: 0.0882 - mean_absolute_error: 0.3701 - accuracy: 0.7221\n",
      "Epoch 4/200\n",
      " - 0s - loss: 0.0878 - mean_absolute_error: 0.3638 - accuracy: 0.7198\n",
      "Epoch 5/200\n",
      " - 0s - loss: 0.0871 - mean_absolute_error: 0.3655 - accuracy: 0.7244\n",
      "Epoch 6/200\n",
      " - 0s - loss: 0.0869 - mean_absolute_error: 0.3635 - accuracy: 0.7302\n",
      "Epoch 7/200\n",
      " - 0s - loss: 0.0867 - mean_absolute_error: 0.3619 - accuracy: 0.7221\n",
      "Epoch 8/200\n",
      " - 0s - loss: 0.0868 - mean_absolute_error: 0.3616 - accuracy: 0.7227\n",
      "Epoch 9/200\n",
      " - 0s - loss: 0.0889 - mean_absolute_error: 0.3717 - accuracy: 0.7203\n",
      "Epoch 10/200\n",
      " - 0s - loss: 0.0886 - mean_absolute_error: 0.3684 - accuracy: 0.7186\n",
      "Epoch 11/200\n",
      " - 0s - loss: 0.0890 - mean_absolute_error: 0.3689 - accuracy: 0.7163\n",
      "Epoch 12/200\n",
      " - 0s - loss: 0.0883 - mean_absolute_error: 0.3659 - accuracy: 0.7151\n",
      "Epoch 13/200\n",
      " - 0s - loss: 0.0870 - mean_absolute_error: 0.3661 - accuracy: 0.7262\n",
      "Epoch 14/200\n",
      " - 0s - loss: 0.0880 - mean_absolute_error: 0.3670 - accuracy: 0.7209\n",
      "Epoch 15/200\n",
      " - 0s - loss: 0.0868 - mean_absolute_error: 0.3625 - accuracy: 0.7209\n",
      "Epoch 16/200\n",
      " - 0s - loss: 0.0877 - mean_absolute_error: 0.3603 - accuracy: 0.7198\n",
      "Epoch 17/200\n",
      " - 0s - loss: 0.0878 - mean_absolute_error: 0.3668 - accuracy: 0.7174\n",
      "Epoch 18/200\n",
      " - 0s - loss: 0.0878 - mean_absolute_error: 0.3649 - accuracy: 0.7250\n",
      "Epoch 19/200\n",
      " - 0s - loss: 0.0882 - mean_absolute_error: 0.3641 - accuracy: 0.7250\n",
      "Epoch 20/200\n",
      " - 0s - loss: 0.0876 - mean_absolute_error: 0.3656 - accuracy: 0.7198\n",
      "Epoch 21/200\n",
      " - 0s - loss: 0.0886 - mean_absolute_error: 0.3702 - accuracy: 0.7134\n",
      "Epoch 22/200\n",
      " - 0s - loss: 0.0876 - mean_absolute_error: 0.3637 - accuracy: 0.7227\n",
      "Epoch 23/200\n",
      " - 0s - loss: 0.0875 - mean_absolute_error: 0.3656 - accuracy: 0.7192\n",
      "Epoch 24/200\n",
      " - 0s - loss: 0.0883 - mean_absolute_error: 0.3666 - accuracy: 0.7238\n",
      "Epoch 25/200\n",
      " - 0s - loss: 0.0882 - mean_absolute_error: 0.3658 - accuracy: 0.7238\n",
      "Epoch 26/200\n",
      " - 0s - loss: 0.0881 - mean_absolute_error: 0.3686 - accuracy: 0.7291\n",
      "Epoch 27/200\n",
      " - 0s - loss: 0.0878 - mean_absolute_error: 0.3675 - accuracy: 0.7221\n",
      "Epoch 28/200\n",
      " - 0s - loss: 0.0866 - mean_absolute_error: 0.3620 - accuracy: 0.7262\n",
      "Epoch 29/200\n",
      " - 0s - loss: 0.0866 - mean_absolute_error: 0.3610 - accuracy: 0.7238\n",
      "Epoch 30/200\n",
      " - 0s - loss: 0.0873 - mean_absolute_error: 0.3636 - accuracy: 0.7267\n",
      "Epoch 31/200\n",
      " - 0s - loss: 0.0863 - mean_absolute_error: 0.3581 - accuracy: 0.7314\n",
      "Epoch 32/200\n",
      " - 0s - loss: 0.0877 - mean_absolute_error: 0.3620 - accuracy: 0.7233\n",
      "Epoch 33/200\n",
      " - 0s - loss: 0.0871 - mean_absolute_error: 0.3679 - accuracy: 0.7314\n",
      "Epoch 34/200\n",
      " - 0s - loss: 0.0888 - mean_absolute_error: 0.3684 - accuracy: 0.7116\n",
      "Epoch 35/200\n",
      " - 0s - loss: 0.0877 - mean_absolute_error: 0.3657 - accuracy: 0.7291\n",
      "Epoch 36/200\n",
      " - 0s - loss: 0.0880 - mean_absolute_error: 0.3654 - accuracy: 0.7250\n",
      "Epoch 37/200\n",
      " - 0s - loss: 0.0873 - mean_absolute_error: 0.3648 - accuracy: 0.7279\n",
      "Epoch 38/200\n",
      " - 0s - loss: 0.0857 - mean_absolute_error: 0.3609 - accuracy: 0.7297\n",
      "Epoch 39/200\n",
      " - 0s - loss: 0.0876 - mean_absolute_error: 0.3636 - accuracy: 0.7244\n",
      "Epoch 40/200\n",
      " - 0s - loss: 0.0884 - mean_absolute_error: 0.3646 - accuracy: 0.7262\n",
      "Epoch 41/200\n",
      " - 0s - loss: 0.0871 - mean_absolute_error: 0.3651 - accuracy: 0.7209\n",
      "Epoch 42/200\n",
      " - 0s - loss: 0.0853 - mean_absolute_error: 0.3575 - accuracy: 0.7273\n",
      "Epoch 43/200\n",
      " - 0s - loss: 0.0868 - mean_absolute_error: 0.3611 - accuracy: 0.7273\n",
      "Epoch 44/200\n",
      " - 0s - loss: 0.0884 - mean_absolute_error: 0.3656 - accuracy: 0.7203\n",
      "Epoch 45/200\n",
      " - 0s - loss: 0.0876 - mean_absolute_error: 0.3656 - accuracy: 0.7209\n",
      "Epoch 46/200\n",
      " - 0s - loss: 0.0879 - mean_absolute_error: 0.3625 - accuracy: 0.7186\n",
      "Epoch 47/200\n",
      " - 0s - loss: 0.0874 - mean_absolute_error: 0.3635 - accuracy: 0.7203\n",
      "Epoch 48/200\n",
      " - 0s - loss: 0.0870 - mean_absolute_error: 0.3643 - accuracy: 0.7279\n",
      "Epoch 49/200\n",
      " - 0s - loss: 0.0873 - mean_absolute_error: 0.3626 - accuracy: 0.7250\n",
      "Epoch 50/200\n",
      " - 0s - loss: 0.0874 - mean_absolute_error: 0.3641 - accuracy: 0.7122\n",
      "Epoch 51/200\n",
      " - 0s - loss: 0.0874 - mean_absolute_error: 0.3628 - accuracy: 0.7209\n",
      "Epoch 52/200\n",
      " - 0s - loss: 0.0865 - mean_absolute_error: 0.3603 - accuracy: 0.7262\n",
      "Epoch 53/200\n",
      " - 0s - loss: 0.0875 - mean_absolute_error: 0.3614 - accuracy: 0.7291\n",
      "Epoch 54/200\n",
      " - 0s - loss: 0.0880 - mean_absolute_error: 0.3672 - accuracy: 0.7256\n",
      "Epoch 55/200\n",
      " - 0s - loss: 0.0863 - mean_absolute_error: 0.3623 - accuracy: 0.7250\n",
      "Epoch 56/200\n",
      " - 0s - loss: 0.0876 - mean_absolute_error: 0.3666 - accuracy: 0.7273\n",
      "Epoch 57/200\n",
      " - 0s - loss: 0.0897 - mean_absolute_error: 0.3692 - accuracy: 0.7140\n",
      "Epoch 58/200\n",
      " - 0s - loss: 0.0853 - mean_absolute_error: 0.3599 - accuracy: 0.7291\n",
      "Epoch 59/200\n",
      " - 0s - loss: 0.0876 - mean_absolute_error: 0.3643 - accuracy: 0.7302\n",
      "Epoch 60/200\n",
      " - 0s - loss: 0.0881 - mean_absolute_error: 0.3653 - accuracy: 0.7262\n",
      "Epoch 61/200\n",
      " - 0s - loss: 0.0870 - mean_absolute_error: 0.3618 - accuracy: 0.7169\n",
      "Epoch 62/200\n",
      " - 0s - loss: 0.0858 - mean_absolute_error: 0.3591 - accuracy: 0.7326\n",
      "Epoch 63/200\n",
      " - 0s - loss: 0.0879 - mean_absolute_error: 0.3654 - accuracy: 0.7291\n",
      "Epoch 64/200\n",
      " - 0s - loss: 0.0869 - mean_absolute_error: 0.3609 - accuracy: 0.7244\n",
      "Epoch 65/200\n",
      " - 0s - loss: 0.0868 - mean_absolute_error: 0.3617 - accuracy: 0.7157\n",
      "Epoch 66/200\n",
      " - 0s - loss: 0.0853 - mean_absolute_error: 0.3579 - accuracy: 0.7308\n",
      "Epoch 67/200\n",
      " - 0s - loss: 0.0891 - mean_absolute_error: 0.3687 - accuracy: 0.7227\n",
      "Epoch 68/200\n",
      " - 0s - loss: 0.0852 - mean_absolute_error: 0.3586 - accuracy: 0.7256\n",
      "Epoch 69/200\n",
      " - 0s - loss: 0.0876 - mean_absolute_error: 0.3601 - accuracy: 0.7279\n",
      "Epoch 70/200\n",
      " - 0s - loss: 0.0865 - mean_absolute_error: 0.3614 - accuracy: 0.7256\n",
      "Epoch 71/200\n",
      " - 0s - loss: 0.0883 - mean_absolute_error: 0.3658 - accuracy: 0.7221\n",
      "Epoch 72/200\n",
      " - 0s - loss: 0.0876 - mean_absolute_error: 0.3603 - accuracy: 0.7337\n",
      "Epoch 73/200\n",
      " - 0s - loss: 0.0875 - mean_absolute_error: 0.3644 - accuracy: 0.7297\n",
      "Epoch 74/200\n",
      " - 0s - loss: 0.0865 - mean_absolute_error: 0.3602 - accuracy: 0.7250\n",
      "Epoch 75/200\n",
      " - 0s - loss: 0.0852 - mean_absolute_error: 0.3550 - accuracy: 0.7291\n",
      "Epoch 76/200\n",
      " - 0s - loss: 0.0877 - mean_absolute_error: 0.3632 - accuracy: 0.7227\n",
      "Epoch 77/200\n",
      " - 0s - loss: 0.0872 - mean_absolute_error: 0.3628 - accuracy: 0.7262\n",
      "Epoch 78/200\n",
      " - 0s - loss: 0.0879 - mean_absolute_error: 0.3656 - accuracy: 0.7186\n",
      "Epoch 79/200\n",
      " - 0s - loss: 0.0881 - mean_absolute_error: 0.3667 - accuracy: 0.7198\n",
      "Epoch 80/200\n",
      " - 0s - loss: 0.0861 - mean_absolute_error: 0.3607 - accuracy: 0.7221\n",
      "Epoch 81/200\n",
      " - 0s - loss: 0.0879 - mean_absolute_error: 0.3644 - accuracy: 0.7267\n",
      "Epoch 82/200\n",
      " - 0s - loss: 0.0863 - mean_absolute_error: 0.3590 - accuracy: 0.7279\n",
      "Epoch 83/200\n",
      " - 0s - loss: 0.0881 - mean_absolute_error: 0.3645 - accuracy: 0.7122\n",
      "Epoch 84/200\n",
      " - 0s - loss: 0.0875 - mean_absolute_error: 0.3638 - accuracy: 0.7302\n",
      "Epoch 85/200\n",
      " - 0s - loss: 0.0872 - mean_absolute_error: 0.3613 - accuracy: 0.7349\n",
      "Epoch 86/200\n",
      " - 0s - loss: 0.0875 - mean_absolute_error: 0.3636 - accuracy: 0.7227\n",
      "Epoch 87/200\n",
      " - 0s - loss: 0.0882 - mean_absolute_error: 0.3670 - accuracy: 0.7134\n",
      "Epoch 88/200\n",
      " - 0s - loss: 0.0865 - mean_absolute_error: 0.3616 - accuracy: 0.7215\n",
      "Epoch 89/200\n",
      " - 0s - loss: 0.0871 - mean_absolute_error: 0.3639 - accuracy: 0.7203\n",
      "Epoch 90/200\n",
      " - 0s - loss: 0.0866 - mean_absolute_error: 0.3631 - accuracy: 0.7192\n",
      "Epoch 91/200\n",
      " - 0s - loss: 0.0870 - mean_absolute_error: 0.3599 - accuracy: 0.7174\n",
      "Epoch 92/200\n",
      " - 0s - loss: 0.0887 - mean_absolute_error: 0.3638 - accuracy: 0.7209\n",
      "Epoch 93/200\n",
      " - 0s - loss: 0.0871 - mean_absolute_error: 0.3637 - accuracy: 0.7244\n",
      "Epoch 94/200\n",
      " - 0s - loss: 0.0877 - mean_absolute_error: 0.3661 - accuracy: 0.7227\n",
      "Epoch 95/200\n",
      " - 0s - loss: 0.0880 - mean_absolute_error: 0.3641 - accuracy: 0.7140\n",
      "Epoch 96/200\n",
      " - 0s - loss: 0.0864 - mean_absolute_error: 0.3614 - accuracy: 0.7198\n",
      "Epoch 97/200\n",
      " - 0s - loss: 0.0885 - mean_absolute_error: 0.3683 - accuracy: 0.7215\n",
      "Epoch 98/200\n",
      " - 0s - loss: 0.0888 - mean_absolute_error: 0.3694 - accuracy: 0.7192\n",
      "Epoch 99/200\n",
      " - 0s - loss: 0.0876 - mean_absolute_error: 0.3666 - accuracy: 0.7198\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 100/200\n",
      " - 0s - loss: 0.0879 - mean_absolute_error: 0.3620 - accuracy: 0.7267\n",
      "Epoch 101/200\n",
      " - 0s - loss: 0.0862 - mean_absolute_error: 0.3635 - accuracy: 0.7337\n",
      "Epoch 102/200\n",
      " - 0s - loss: 0.0873 - mean_absolute_error: 0.3635 - accuracy: 0.7256\n",
      "Epoch 103/200\n",
      " - 0s - loss: 0.0878 - mean_absolute_error: 0.3634 - accuracy: 0.7267\n",
      "Epoch 104/200\n",
      " - 0s - loss: 0.0871 - mean_absolute_error: 0.3605 - accuracy: 0.7233\n",
      "Epoch 105/200\n",
      " - 0s - loss: 0.0856 - mean_absolute_error: 0.3615 - accuracy: 0.7331\n",
      "Epoch 106/200\n",
      " - 0s - loss: 0.0858 - mean_absolute_error: 0.3575 - accuracy: 0.7262\n",
      "Epoch 107/200\n",
      " - 0s - loss: 0.0864 - mean_absolute_error: 0.3612 - accuracy: 0.7267\n",
      "Epoch 108/200\n",
      " - 0s - loss: 0.0860 - mean_absolute_error: 0.3595 - accuracy: 0.7291\n",
      "Epoch 109/200\n",
      " - 0s - loss: 0.0871 - mean_absolute_error: 0.3612 - accuracy: 0.7302\n",
      "Epoch 110/200\n",
      " - 0s - loss: 0.0865 - mean_absolute_error: 0.3569 - accuracy: 0.7267\n",
      "Epoch 111/200\n",
      " - 0s - loss: 0.0890 - mean_absolute_error: 0.3654 - accuracy: 0.7221\n",
      "Epoch 112/200\n",
      " - 0s - loss: 0.0865 - mean_absolute_error: 0.3605 - accuracy: 0.7291\n",
      "Epoch 113/200\n",
      " - 0s - loss: 0.0863 - mean_absolute_error: 0.3630 - accuracy: 0.7314\n",
      "Epoch 114/200\n",
      " - 0s - loss: 0.0861 - mean_absolute_error: 0.3599 - accuracy: 0.7331\n",
      "Epoch 115/200\n",
      " - 0s - loss: 0.0863 - mean_absolute_error: 0.3589 - accuracy: 0.7192\n",
      "Epoch 116/200\n",
      " - 0s - loss: 0.0866 - mean_absolute_error: 0.3605 - accuracy: 0.7233\n",
      "Epoch 117/200\n",
      " - 0s - loss: 0.0864 - mean_absolute_error: 0.3589 - accuracy: 0.7314\n",
      "Epoch 118/200\n",
      " - 0s - loss: 0.0877 - mean_absolute_error: 0.3642 - accuracy: 0.7250\n",
      "Epoch 119/200\n",
      " - 0s - loss: 0.0867 - mean_absolute_error: 0.3598 - accuracy: 0.7262\n",
      "Epoch 120/200\n",
      " - 0s - loss: 0.0853 - mean_absolute_error: 0.3547 - accuracy: 0.7238\n",
      "Epoch 121/200\n",
      " - 0s - loss: 0.0847 - mean_absolute_error: 0.3549 - accuracy: 0.7360\n",
      "Epoch 122/200\n",
      " - 0s - loss: 0.0871 - mean_absolute_error: 0.3599 - accuracy: 0.7250\n",
      "Epoch 123/200\n",
      " - 0s - loss: 0.0878 - mean_absolute_error: 0.3618 - accuracy: 0.7192\n",
      "Epoch 124/200\n",
      " - 0s - loss: 0.0877 - mean_absolute_error: 0.3638 - accuracy: 0.7203\n",
      "Epoch 125/200\n",
      " - 0s - loss: 0.0873 - mean_absolute_error: 0.3628 - accuracy: 0.7262\n",
      "Epoch 126/200\n",
      " - 0s - loss: 0.0859 - mean_absolute_error: 0.3587 - accuracy: 0.7302\n",
      "Epoch 127/200\n",
      " - 0s - loss: 0.0864 - mean_absolute_error: 0.3607 - accuracy: 0.7314\n",
      "Epoch 128/200\n",
      " - 0s - loss: 0.0859 - mean_absolute_error: 0.3571 - accuracy: 0.7349\n",
      "Epoch 129/200\n",
      " - 0s - loss: 0.0874 - mean_absolute_error: 0.3614 - accuracy: 0.7279\n",
      "Epoch 130/200\n",
      " - 0s - loss: 0.0876 - mean_absolute_error: 0.3655 - accuracy: 0.7221\n",
      "Epoch 131/200\n",
      " - 0s - loss: 0.0866 - mean_absolute_error: 0.3616 - accuracy: 0.7221\n",
      "Epoch 132/200\n",
      " - 0s - loss: 0.0883 - mean_absolute_error: 0.3575 - accuracy: 0.7157\n",
      "Epoch 133/200\n",
      " - 0s - loss: 0.0870 - mean_absolute_error: 0.3631 - accuracy: 0.7302\n",
      "Epoch 134/200\n",
      " - 0s - loss: 0.0845 - mean_absolute_error: 0.3520 - accuracy: 0.7256\n",
      "Epoch 135/200\n",
      " - 0s - loss: 0.0876 - mean_absolute_error: 0.3616 - accuracy: 0.7233\n",
      "Epoch 136/200\n",
      " - 0s - loss: 0.0866 - mean_absolute_error: 0.3627 - accuracy: 0.7262\n",
      "Epoch 137/200\n",
      " - 0s - loss: 0.0856 - mean_absolute_error: 0.3589 - accuracy: 0.7256\n",
      "Epoch 138/200\n",
      " - 0s - loss: 0.0869 - mean_absolute_error: 0.3596 - accuracy: 0.7267\n",
      "Epoch 139/200\n",
      " - 0s - loss: 0.0858 - mean_absolute_error: 0.3566 - accuracy: 0.7320\n",
      "Epoch 140/200\n",
      " - 0s - loss: 0.0856 - mean_absolute_error: 0.3589 - accuracy: 0.7297\n",
      "Epoch 141/200\n",
      " - 0s - loss: 0.0867 - mean_absolute_error: 0.3605 - accuracy: 0.7256\n",
      "Epoch 142/200\n",
      " - 0s - loss: 0.0873 - mean_absolute_error: 0.3646 - accuracy: 0.7337\n",
      "Epoch 143/200\n",
      " - 0s - loss: 0.0857 - mean_absolute_error: 0.3584 - accuracy: 0.7372\n",
      "Epoch 144/200\n",
      " - 0s - loss: 0.0852 - mean_absolute_error: 0.3582 - accuracy: 0.7349\n",
      "Epoch 145/200\n",
      " - 0s - loss: 0.0855 - mean_absolute_error: 0.3542 - accuracy: 0.7256\n",
      "Epoch 146/200\n",
      " - 0s - loss: 0.0849 - mean_absolute_error: 0.3527 - accuracy: 0.7366\n",
      "Epoch 147/200\n",
      " - 0s - loss: 0.0844 - mean_absolute_error: 0.3533 - accuracy: 0.7337\n",
      "Epoch 148/200\n",
      " - 0s - loss: 0.0869 - mean_absolute_error: 0.3605 - accuracy: 0.7238\n",
      "Epoch 149/200\n",
      " - 0s - loss: 0.0875 - mean_absolute_error: 0.3664 - accuracy: 0.7203\n",
      "Epoch 150/200\n",
      " - 0s - loss: 0.0871 - mean_absolute_error: 0.3617 - accuracy: 0.7372\n",
      "Epoch 151/200\n",
      " - 0s - loss: 0.0866 - mean_absolute_error: 0.3581 - accuracy: 0.7285\n",
      "Epoch 152/200\n",
      " - 0s - loss: 0.0861 - mean_absolute_error: 0.3594 - accuracy: 0.7349\n",
      "Epoch 153/200\n",
      " - 0s - loss: 0.0864 - mean_absolute_error: 0.3606 - accuracy: 0.7233\n",
      "Epoch 154/200\n",
      " - 0s - loss: 0.0853 - mean_absolute_error: 0.3554 - accuracy: 0.7343\n",
      "Epoch 155/200\n",
      " - 0s - loss: 0.0851 - mean_absolute_error: 0.3521 - accuracy: 0.7326\n",
      "Epoch 156/200\n",
      " - 0s - loss: 0.0862 - mean_absolute_error: 0.3580 - accuracy: 0.7238\n",
      "Epoch 157/200\n",
      " - 0s - loss: 0.0851 - mean_absolute_error: 0.3563 - accuracy: 0.7308\n",
      "Epoch 158/200\n",
      " - 0s - loss: 0.0855 - mean_absolute_error: 0.3551 - accuracy: 0.7302\n",
      "Epoch 159/200\n",
      " - 0s - loss: 0.0859 - mean_absolute_error: 0.3574 - accuracy: 0.7360\n",
      "Epoch 160/200\n",
      " - 0s - loss: 0.0849 - mean_absolute_error: 0.3563 - accuracy: 0.7349\n",
      "Epoch 161/200\n",
      " - 0s - loss: 0.0867 - mean_absolute_error: 0.3589 - accuracy: 0.7279\n",
      "Epoch 162/200\n",
      " - 0s - loss: 0.0873 - mean_absolute_error: 0.3657 - accuracy: 0.7198\n",
      "Epoch 163/200\n",
      " - 0s - loss: 0.0869 - mean_absolute_error: 0.3605 - accuracy: 0.7285\n",
      "Epoch 164/200\n",
      " - 0s - loss: 0.0868 - mean_absolute_error: 0.3623 - accuracy: 0.7209\n",
      "Epoch 165/200\n",
      " - 0s - loss: 0.0862 - mean_absolute_error: 0.3577 - accuracy: 0.7244\n",
      "Epoch 166/200\n",
      " - 0s - loss: 0.0837 - mean_absolute_error: 0.3540 - accuracy: 0.7238\n",
      "Epoch 167/200\n",
      " - 0s - loss: 0.0851 - mean_absolute_error: 0.3538 - accuracy: 0.7273\n",
      "Epoch 168/200\n",
      " - 0s - loss: 0.0861 - mean_absolute_error: 0.3599 - accuracy: 0.7320\n",
      "Epoch 169/200\n",
      " - 0s - loss: 0.0867 - mean_absolute_error: 0.3578 - accuracy: 0.7244\n",
      "Epoch 170/200\n",
      " - 0s - loss: 0.0864 - mean_absolute_error: 0.3561 - accuracy: 0.7302\n",
      "Epoch 171/200\n",
      " - 0s - loss: 0.0852 - mean_absolute_error: 0.3538 - accuracy: 0.7256\n",
      "Epoch 172/200\n",
      " - 0s - loss: 0.0870 - mean_absolute_error: 0.3594 - accuracy: 0.7302\n",
      "Epoch 173/200\n",
      " - 0s - loss: 0.0863 - mean_absolute_error: 0.3568 - accuracy: 0.7215\n",
      "Epoch 174/200\n",
      " - 0s - loss: 0.0872 - mean_absolute_error: 0.3633 - accuracy: 0.7256\n",
      "Epoch 175/200\n",
      " - 0s - loss: 0.0884 - mean_absolute_error: 0.3654 - accuracy: 0.7285\n",
      "Epoch 176/200\n",
      " - 0s - loss: 0.0861 - mean_absolute_error: 0.3605 - accuracy: 0.7262\n",
      "Epoch 177/200\n",
      " - 0s - loss: 0.0853 - mean_absolute_error: 0.3580 - accuracy: 0.7273\n",
      "Epoch 178/200\n",
      " - 0s - loss: 0.0855 - mean_absolute_error: 0.3569 - accuracy: 0.7302\n",
      "Epoch 179/200\n",
      " - 0s - loss: 0.0867 - mean_absolute_error: 0.3624 - accuracy: 0.7320\n",
      "Epoch 180/200\n",
      " - 0s - loss: 0.0871 - mean_absolute_error: 0.3608 - accuracy: 0.7267\n",
      "Epoch 181/200\n",
      " - 0s - loss: 0.0869 - mean_absolute_error: 0.3581 - accuracy: 0.7372\n",
      "Epoch 182/200\n",
      " - 0s - loss: 0.0851 - mean_absolute_error: 0.3552 - accuracy: 0.7343\n",
      "Epoch 183/200\n",
      " - 0s - loss: 0.0845 - mean_absolute_error: 0.3503 - accuracy: 0.7384\n",
      "Epoch 184/200\n",
      " - 0s - loss: 0.0868 - mean_absolute_error: 0.3626 - accuracy: 0.7262\n",
      "Epoch 185/200\n",
      " - 0s - loss: 0.0865 - mean_absolute_error: 0.3573 - accuracy: 0.7256\n",
      "Epoch 186/200\n",
      " - 0s - loss: 0.0842 - mean_absolute_error: 0.3547 - accuracy: 0.7279\n",
      "Epoch 187/200\n",
      " - 0s - loss: 0.0855 - mean_absolute_error: 0.3540 - accuracy: 0.7337\n",
      "Epoch 188/200\n",
      " - 0s - loss: 0.0869 - mean_absolute_error: 0.3584 - accuracy: 0.7250\n",
      "Epoch 189/200\n",
      " - 0s - loss: 0.0890 - mean_absolute_error: 0.3653 - accuracy: 0.7215\n",
      "Epoch 190/200\n",
      " - 0s - loss: 0.0858 - mean_absolute_error: 0.3601 - accuracy: 0.7343\n",
      "Epoch 191/200\n",
      " - 0s - loss: 0.0848 - mean_absolute_error: 0.3547 - accuracy: 0.7372\n",
      "Epoch 192/200\n",
      " - 0s - loss: 0.0846 - mean_absolute_error: 0.3517 - accuracy: 0.7343\n",
      "Epoch 193/200\n",
      " - 0s - loss: 0.0866 - mean_absolute_error: 0.3603 - accuracy: 0.7279\n",
      "Epoch 194/200\n",
      " - 0s - loss: 0.0852 - mean_absolute_error: 0.3543 - accuracy: 0.7355\n",
      "Epoch 195/200\n",
      " - 0s - loss: 0.0850 - mean_absolute_error: 0.3543 - accuracy: 0.7238\n",
      "Epoch 196/200\n",
      " - 0s - loss: 0.0872 - mean_absolute_error: 0.3635 - accuracy: 0.7221\n",
      "Epoch 197/200\n",
      " - 0s - loss: 0.0853 - mean_absolute_error: 0.3595 - accuracy: 0.7337\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 198/200\n",
      " - 0s - loss: 0.0853 - mean_absolute_error: 0.3542 - accuracy: 0.7343\n",
      "Epoch 199/200\n",
      " - 0s - loss: 0.0876 - mean_absolute_error: 0.3635 - accuracy: 0.7250\n",
      "Epoch 200/200\n",
      " - 0s - loss: 0.0861 - mean_absolute_error: 0.3556 - accuracy: 0.7308\n"
     ]
    }
   ],
   "source": [
    "RMSprop=optimizers.RMSprop(learning_rate=0.4, rho=0.9)\n",
    "model2.compile(optimizer='RMSprop', loss='logcosh',metrics=[metrics.mae, 'accuracy'])\n",
    "history=model2.fit(X1, y1, batch_size=32, epochs=200, verbose=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "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"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['accuracy'])\n",
    "plt.title('Model accuracy')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Valid'], loc='upper left')\n",
    "plt.show()\n",
    "\n",
    "plt.plot(history.history['loss'])\n",
    "plt.title('Model loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Valid'], loc='upper left')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__Accuracy :__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy on Test Data :- \n",
      "\t loss: 0.10839736822879675   mean_absolute_error: 0.45831698179244995   accuracy: 0.6969696879386902\n"
     ]
    }
   ],
   "source": [
    "loss, mean_absolute_error, accuracy =model2.evaluate(X2, y2, batch_size=32, verbose=2)\n",
    "print('Accuracy on Test Data :- ')\n",
    "print('\\t loss:',loss , '  mean_absolute_error:',mean_absolute_error, '  accuracy:',accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__2. optimizers : Adagrad__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      " - 0s - loss: 0.0898 - mean_absolute_error: 0.3746 - accuracy: 0.7157\n",
      "Epoch 2/200\n",
      " - 0s - loss: 0.0906 - mean_absolute_error: 0.3756 - accuracy: 0.7029\n",
      "Epoch 3/200\n",
      " - 0s - loss: 0.0892 - mean_absolute_error: 0.3695 - accuracy: 0.7221\n",
      "Epoch 4/200\n",
      " - 0s - loss: 0.0868 - mean_absolute_error: 0.3657 - accuracy: 0.7250\n",
      "Epoch 5/200\n",
      " - 0s - loss: 0.0890 - mean_absolute_error: 0.3708 - accuracy: 0.7198\n",
      "Epoch 6/200\n",
      " - 0s - loss: 0.0850 - mean_absolute_error: 0.3605 - accuracy: 0.7273\n",
      "Epoch 7/200\n",
      " - 0s - loss: 0.0853 - mean_absolute_error: 0.3589 - accuracy: 0.7308\n",
      "Epoch 8/200\n",
      " - 0s - loss: 0.0861 - mean_absolute_error: 0.3641 - accuracy: 0.7320\n",
      "Epoch 9/200\n",
      " - 0s - loss: 0.0874 - mean_absolute_error: 0.3645 - accuracy: 0.7267\n",
      "Epoch 10/200\n",
      " - 0s - loss: 0.0871 - mean_absolute_error: 0.3646 - accuracy: 0.7279\n",
      "Epoch 11/200\n",
      " - 0s - loss: 0.0859 - mean_absolute_error: 0.3616 - accuracy: 0.7308\n",
      "Epoch 12/200\n",
      " - 0s - loss: 0.0861 - mean_absolute_error: 0.3634 - accuracy: 0.7302\n",
      "Epoch 13/200\n",
      " - 0s - loss: 0.0851 - mean_absolute_error: 0.3602 - accuracy: 0.7326\n",
      "Epoch 14/200\n",
      " - 0s - loss: 0.0859 - mean_absolute_error: 0.3632 - accuracy: 0.7250\n",
      "Epoch 15/200\n",
      " - 0s - loss: 0.0856 - mean_absolute_error: 0.3602 - accuracy: 0.7326\n",
      "Epoch 16/200\n",
      " - 0s - loss: 0.0852 - mean_absolute_error: 0.3565 - accuracy: 0.7424\n",
      "Epoch 17/200\n",
      " - 0s - loss: 0.0870 - mean_absolute_error: 0.3648 - accuracy: 0.7256\n",
      "Epoch 18/200\n",
      " - 0s - loss: 0.0842 - mean_absolute_error: 0.3572 - accuracy: 0.7372\n",
      "Epoch 19/200\n",
      " - 0s - loss: 0.0849 - mean_absolute_error: 0.3605 - accuracy: 0.7326\n",
      "Epoch 20/200\n",
      " - 0s - loss: 0.0848 - mean_absolute_error: 0.3561 - accuracy: 0.7238\n",
      "Epoch 21/200\n",
      " - 0s - loss: 0.0852 - mean_absolute_error: 0.3612 - accuracy: 0.7314\n",
      "Epoch 22/200\n",
      " - 0s - loss: 0.0865 - mean_absolute_error: 0.3624 - accuracy: 0.7267\n",
      "Epoch 23/200\n",
      " - 0s - loss: 0.0852 - mean_absolute_error: 0.3573 - accuracy: 0.7320\n",
      "Epoch 24/200\n",
      " - 0s - loss: 0.0845 - mean_absolute_error: 0.3580 - accuracy: 0.7326\n",
      "Epoch 25/200\n",
      " - 0s - loss: 0.0861 - mean_absolute_error: 0.3589 - accuracy: 0.7198\n",
      "Epoch 26/200\n",
      " - 0s - loss: 0.0831 - mean_absolute_error: 0.3529 - accuracy: 0.7366\n",
      "Epoch 27/200\n",
      " - 0s - loss: 0.0850 - mean_absolute_error: 0.3592 - accuracy: 0.7262\n",
      "Epoch 28/200\n",
      " - 0s - loss: 0.0841 - mean_absolute_error: 0.3581 - accuracy: 0.7302\n",
      "Epoch 29/200\n",
      " - 0s - loss: 0.0870 - mean_absolute_error: 0.3633 - accuracy: 0.7279\n",
      "Epoch 30/200\n",
      " - 0s - loss: 0.0842 - mean_absolute_error: 0.3579 - accuracy: 0.7349\n",
      "Epoch 31/200\n",
      " - 0s - loss: 0.0835 - mean_absolute_error: 0.3549 - accuracy: 0.7331\n",
      "Epoch 32/200\n",
      " - 0s - loss: 0.0846 - mean_absolute_error: 0.3557 - accuracy: 0.7395\n",
      "Epoch 33/200\n",
      " - 0s - loss: 0.0836 - mean_absolute_error: 0.3551 - accuracy: 0.7419\n",
      "Epoch 34/200\n",
      " - 0s - loss: 0.0849 - mean_absolute_error: 0.3539 - accuracy: 0.7419\n",
      "Epoch 35/200\n",
      " - 0s - loss: 0.0843 - mean_absolute_error: 0.3550 - accuracy: 0.7331\n",
      "Epoch 36/200\n",
      " - 0s - loss: 0.0850 - mean_absolute_error: 0.3558 - accuracy: 0.7262\n",
      "Epoch 37/200\n",
      " - 0s - loss: 0.0846 - mean_absolute_error: 0.3557 - accuracy: 0.7308\n",
      "Epoch 38/200\n",
      " - 0s - loss: 0.0857 - mean_absolute_error: 0.3594 - accuracy: 0.7331\n",
      "Epoch 39/200\n",
      " - 0s - loss: 0.0823 - mean_absolute_error: 0.3499 - accuracy: 0.7390\n",
      "Epoch 40/200\n",
      " - 0s - loss: 0.0846 - mean_absolute_error: 0.3552 - accuracy: 0.7238\n",
      "Epoch 41/200\n",
      " - 0s - loss: 0.0840 - mean_absolute_error: 0.3556 - accuracy: 0.7337\n",
      "Epoch 42/200\n",
      " - 0s - loss: 0.0846 - mean_absolute_error: 0.3572 - accuracy: 0.7302\n",
      "Epoch 43/200\n",
      " - 0s - loss: 0.0840 - mean_absolute_error: 0.3544 - accuracy: 0.7267\n",
      "Epoch 44/200\n",
      " - 0s - loss: 0.0847 - mean_absolute_error: 0.3559 - accuracy: 0.7331\n",
      "Epoch 45/200\n",
      " - 0s - loss: 0.0838 - mean_absolute_error: 0.3543 - accuracy: 0.7407\n",
      "Epoch 46/200\n",
      " - 0s - loss: 0.0840 - mean_absolute_error: 0.3553 - accuracy: 0.7436\n",
      "Epoch 47/200\n",
      " - 0s - loss: 0.0835 - mean_absolute_error: 0.3533 - accuracy: 0.7291\n",
      "Epoch 48/200\n",
      " - 0s - loss: 0.0848 - mean_absolute_error: 0.3564 - accuracy: 0.7297\n",
      "Epoch 49/200\n",
      " - 0s - loss: 0.0836 - mean_absolute_error: 0.3523 - accuracy: 0.7366\n",
      "Epoch 50/200\n",
      " - 0s - loss: 0.0837 - mean_absolute_error: 0.3549 - accuracy: 0.7308\n",
      "Epoch 51/200\n",
      " - 0s - loss: 0.0819 - mean_absolute_error: 0.3464 - accuracy: 0.7407\n",
      "Epoch 52/200\n",
      " - 0s - loss: 0.0850 - mean_absolute_error: 0.3565 - accuracy: 0.7384\n",
      "Epoch 53/200\n",
      " - 0s - loss: 0.0836 - mean_absolute_error: 0.3521 - accuracy: 0.7273\n",
      "Epoch 54/200\n",
      " - 0s - loss: 0.0838 - mean_absolute_error: 0.3523 - accuracy: 0.7390\n",
      "Epoch 55/200\n",
      " - 0s - loss: 0.0849 - mean_absolute_error: 0.3546 - accuracy: 0.7343\n",
      "Epoch 56/200\n",
      " - 0s - loss: 0.0839 - mean_absolute_error: 0.3538 - accuracy: 0.7273\n",
      "Epoch 57/200\n",
      " - 0s - loss: 0.0845 - mean_absolute_error: 0.3568 - accuracy: 0.7349\n",
      "Epoch 58/200\n",
      " - 0s - loss: 0.0846 - mean_absolute_error: 0.3570 - accuracy: 0.7308\n",
      "Epoch 59/200\n",
      " - 0s - loss: 0.0822 - mean_absolute_error: 0.3513 - accuracy: 0.7378\n",
      "Epoch 60/200\n",
      " - 0s - loss: 0.0852 - mean_absolute_error: 0.3573 - accuracy: 0.7285\n",
      "Epoch 61/200\n",
      " - 0s - loss: 0.0841 - mean_absolute_error: 0.3542 - accuracy: 0.7314\n",
      "Epoch 62/200\n",
      " - 0s - loss: 0.0823 - mean_absolute_error: 0.3490 - accuracy: 0.7459\n",
      "Epoch 63/200\n",
      " - 0s - loss: 0.0833 - mean_absolute_error: 0.3519 - accuracy: 0.7355\n",
      "Epoch 64/200\n",
      " - 0s - loss: 0.0833 - mean_absolute_error: 0.3528 - accuracy: 0.7413\n",
      "Epoch 65/200\n",
      " - 0s - loss: 0.0842 - mean_absolute_error: 0.3538 - accuracy: 0.7343\n",
      "Epoch 66/200\n",
      " - 0s - loss: 0.0838 - mean_absolute_error: 0.3518 - accuracy: 0.7331\n",
      "Epoch 67/200\n",
      " - 0s - loss: 0.0850 - mean_absolute_error: 0.3583 - accuracy: 0.7279\n",
      "Epoch 68/200\n",
      " - 0s - loss: 0.0860 - mean_absolute_error: 0.3602 - accuracy: 0.7291\n",
      "Epoch 69/200\n",
      " - 0s - loss: 0.0838 - mean_absolute_error: 0.3541 - accuracy: 0.7360\n",
      "Epoch 70/200\n",
      " - 0s - loss: 0.0853 - mean_absolute_error: 0.3570 - accuracy: 0.7291\n",
      "Epoch 71/200\n",
      " - 0s - loss: 0.0839 - mean_absolute_error: 0.3539 - accuracy: 0.7401\n",
      "Epoch 72/200\n",
      " - 0s - loss: 0.0835 - mean_absolute_error: 0.3538 - accuracy: 0.7314\n",
      "Epoch 73/200\n",
      " - 0s - loss: 0.0821 - mean_absolute_error: 0.3497 - accuracy: 0.7407\n",
      "Epoch 74/200\n",
      " - 0s - loss: 0.0836 - mean_absolute_error: 0.3535 - accuracy: 0.7453\n",
      "Epoch 75/200\n",
      " - 0s - loss: 0.0839 - mean_absolute_error: 0.3538 - accuracy: 0.7314\n",
      "Epoch 76/200\n",
      " - 0s - loss: 0.0817 - mean_absolute_error: 0.3493 - accuracy: 0.7401\n",
      "Epoch 77/200\n",
      " - 0s - loss: 0.0837 - mean_absolute_error: 0.3530 - accuracy: 0.7297\n",
      "Epoch 78/200\n",
      " - 0s - loss: 0.0828 - mean_absolute_error: 0.3506 - accuracy: 0.7401\n",
      "Epoch 79/200\n",
      " - 0s - loss: 0.0823 - mean_absolute_error: 0.3498 - accuracy: 0.7424\n",
      "Epoch 80/200\n",
      " - 0s - loss: 0.0832 - mean_absolute_error: 0.3521 - accuracy: 0.7384\n",
      "Epoch 81/200\n",
      " - 0s - loss: 0.0847 - mean_absolute_error: 0.3561 - accuracy: 0.7326\n",
      "Epoch 82/200\n",
      " - 0s - loss: 0.0844 - mean_absolute_error: 0.3568 - accuracy: 0.7326\n",
      "Epoch 83/200\n",
      " - 0s - loss: 0.0833 - mean_absolute_error: 0.3523 - accuracy: 0.7424\n",
      "Epoch 84/200\n",
      " - 0s - loss: 0.0837 - mean_absolute_error: 0.3517 - accuracy: 0.7419\n",
      "Epoch 85/200\n",
      " - 0s - loss: 0.0851 - mean_absolute_error: 0.3548 - accuracy: 0.7314\n",
      "Epoch 86/200\n",
      " - 0s - loss: 0.0828 - mean_absolute_error: 0.3502 - accuracy: 0.7343\n",
      "Epoch 87/200\n",
      " - 0s - loss: 0.0832 - mean_absolute_error: 0.3499 - accuracy: 0.7349\n",
      "Epoch 88/200\n",
      " - 0s - loss: 0.0842 - mean_absolute_error: 0.3559 - accuracy: 0.7395\n",
      "Epoch 89/200\n",
      " - 0s - loss: 0.0831 - mean_absolute_error: 0.3510 - accuracy: 0.7407\n",
      "Epoch 90/200\n",
      " - 0s - loss: 0.0821 - mean_absolute_error: 0.3507 - accuracy: 0.7401\n",
      "Epoch 91/200\n",
      " - 0s - loss: 0.0830 - mean_absolute_error: 0.3512 - accuracy: 0.7285\n",
      "Epoch 92/200\n",
      " - 0s - loss: 0.0810 - mean_absolute_error: 0.3463 - accuracy: 0.7459\n",
      "Epoch 93/200\n",
      " - 0s - loss: 0.0818 - mean_absolute_error: 0.3441 - accuracy: 0.7395\n",
      "Epoch 94/200\n",
      " - 0s - loss: 0.0821 - mean_absolute_error: 0.3466 - accuracy: 0.7337\n",
      "Epoch 95/200\n",
      " - 0s - loss: 0.0845 - mean_absolute_error: 0.3559 - accuracy: 0.7355\n",
      "Epoch 96/200\n",
      " - 0s - loss: 0.0829 - mean_absolute_error: 0.3511 - accuracy: 0.7477\n",
      "Epoch 97/200\n",
      " - 0s - loss: 0.0831 - mean_absolute_error: 0.3503 - accuracy: 0.7343\n",
      "Epoch 98/200\n",
      " - 0s - loss: 0.0838 - mean_absolute_error: 0.3549 - accuracy: 0.7285\n",
      "Epoch 99/200\n",
      " - 0s - loss: 0.0824 - mean_absolute_error: 0.3486 - accuracy: 0.7378\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 100/200\n",
      " - 0s - loss: 0.0834 - mean_absolute_error: 0.3522 - accuracy: 0.7436\n",
      "Epoch 101/200\n",
      " - 0s - loss: 0.0827 - mean_absolute_error: 0.3520 - accuracy: 0.7390\n",
      "Epoch 102/200\n",
      " - 0s - loss: 0.0829 - mean_absolute_error: 0.3506 - accuracy: 0.7326\n",
      "Epoch 103/200\n",
      " - 0s - loss: 0.0838 - mean_absolute_error: 0.3522 - accuracy: 0.7349\n",
      "Epoch 104/200\n",
      " - 0s - loss: 0.0818 - mean_absolute_error: 0.3477 - accuracy: 0.7384\n",
      "Epoch 105/200\n",
      " - 0s - loss: 0.0831 - mean_absolute_error: 0.3499 - accuracy: 0.7413\n",
      "Epoch 106/200\n",
      " - 0s - loss: 0.0823 - mean_absolute_error: 0.3493 - accuracy: 0.7390\n",
      "Epoch 107/200\n",
      " - 0s - loss: 0.0820 - mean_absolute_error: 0.3493 - accuracy: 0.7343\n",
      "Epoch 108/200\n",
      " - 0s - loss: 0.0829 - mean_absolute_error: 0.3495 - accuracy: 0.7384\n",
      "Epoch 109/200\n",
      " - 0s - loss: 0.0844 - mean_absolute_error: 0.3534 - accuracy: 0.7314\n",
      "Epoch 110/200\n",
      " - 0s - loss: 0.0825 - mean_absolute_error: 0.3498 - accuracy: 0.7314\n",
      "Epoch 111/200\n",
      " - 0s - loss: 0.0815 - mean_absolute_error: 0.3478 - accuracy: 0.7448\n",
      "Epoch 112/200\n",
      " - 0s - loss: 0.0830 - mean_absolute_error: 0.3476 - accuracy: 0.7337\n",
      "Epoch 113/200\n",
      " - 0s - loss: 0.0833 - mean_absolute_error: 0.3518 - accuracy: 0.7279\n",
      "Epoch 114/200\n",
      " - 0s - loss: 0.0826 - mean_absolute_error: 0.3474 - accuracy: 0.7308\n",
      "Epoch 115/200\n",
      " - 0s - loss: 0.0829 - mean_absolute_error: 0.3485 - accuracy: 0.7419\n",
      "Epoch 116/200\n",
      " - 0s - loss: 0.0842 - mean_absolute_error: 0.3536 - accuracy: 0.7262\n",
      "Epoch 117/200\n",
      " - 0s - loss: 0.0827 - mean_absolute_error: 0.3491 - accuracy: 0.7378\n",
      "Epoch 118/200\n",
      " - 0s - loss: 0.0830 - mean_absolute_error: 0.3532 - accuracy: 0.7401\n",
      "Epoch 119/200\n",
      " - 0s - loss: 0.0818 - mean_absolute_error: 0.3463 - accuracy: 0.7424\n",
      "Epoch 120/200\n",
      " - 0s - loss: 0.0823 - mean_absolute_error: 0.3480 - accuracy: 0.7384\n",
      "Epoch 121/200\n",
      " - 0s - loss: 0.0820 - mean_absolute_error: 0.3496 - accuracy: 0.7360\n",
      "Epoch 122/200\n",
      " - 0s - loss: 0.0831 - mean_absolute_error: 0.3497 - accuracy: 0.7395\n",
      "Epoch 123/200\n",
      " - 0s - loss: 0.0809 - mean_absolute_error: 0.3457 - accuracy: 0.7453\n",
      "Epoch 124/200\n",
      " - 0s - loss: 0.0824 - mean_absolute_error: 0.3460 - accuracy: 0.7436\n",
      "Epoch 125/200\n",
      " - 0s - loss: 0.0833 - mean_absolute_error: 0.3502 - accuracy: 0.7337\n",
      "Epoch 126/200\n",
      " - 0s - loss: 0.0817 - mean_absolute_error: 0.3471 - accuracy: 0.7378\n",
      "Epoch 127/200\n",
      " - 0s - loss: 0.0826 - mean_absolute_error: 0.3480 - accuracy: 0.7413\n",
      "Epoch 128/200\n",
      " - 0s - loss: 0.0841 - mean_absolute_error: 0.3531 - accuracy: 0.7320\n",
      "Epoch 129/200\n",
      " - 0s - loss: 0.0830 - mean_absolute_error: 0.3497 - accuracy: 0.7407\n",
      "Epoch 130/200\n",
      " - 0s - loss: 0.0819 - mean_absolute_error: 0.3484 - accuracy: 0.7401\n",
      "Epoch 131/200\n",
      " - 0s - loss: 0.0829 - mean_absolute_error: 0.3515 - accuracy: 0.7390\n",
      "Epoch 132/200\n",
      " - 0s - loss: 0.0826 - mean_absolute_error: 0.3471 - accuracy: 0.7378\n",
      "Epoch 133/200\n",
      " - 0s - loss: 0.0803 - mean_absolute_error: 0.3399 - accuracy: 0.7442\n",
      "Epoch 134/200\n",
      " - 0s - loss: 0.0835 - mean_absolute_error: 0.3486 - accuracy: 0.7349\n",
      "Epoch 135/200\n",
      " - 0s - loss: 0.0819 - mean_absolute_error: 0.3473 - accuracy: 0.7419\n",
      "Epoch 136/200\n",
      " - 0s - loss: 0.0826 - mean_absolute_error: 0.3480 - accuracy: 0.7483\n",
      "Epoch 137/200\n",
      " - 0s - loss: 0.0824 - mean_absolute_error: 0.3492 - accuracy: 0.7314\n",
      "Epoch 138/200\n",
      " - 0s - loss: 0.0819 - mean_absolute_error: 0.3475 - accuracy: 0.7372\n",
      "Epoch 139/200\n",
      " - 0s - loss: 0.0820 - mean_absolute_error: 0.3478 - accuracy: 0.7343\n",
      "Epoch 140/200\n",
      " - 0s - loss: 0.0818 - mean_absolute_error: 0.3482 - accuracy: 0.7419\n",
      "Epoch 141/200\n",
      " - 0s - loss: 0.0824 - mean_absolute_error: 0.3489 - accuracy: 0.7419\n",
      "Epoch 142/200\n",
      " - 0s - loss: 0.0837 - mean_absolute_error: 0.3515 - accuracy: 0.7349\n",
      "Epoch 143/200\n",
      " - 0s - loss: 0.0820 - mean_absolute_error: 0.3481 - accuracy: 0.7378\n",
      "Epoch 144/200\n",
      " - 0s - loss: 0.0831 - mean_absolute_error: 0.3526 - accuracy: 0.7360\n",
      "Epoch 145/200\n",
      " - 0s - loss: 0.0821 - mean_absolute_error: 0.3471 - accuracy: 0.7541\n",
      "Epoch 146/200\n",
      " - 0s - loss: 0.0815 - mean_absolute_error: 0.3475 - accuracy: 0.7419\n",
      "Epoch 147/200\n",
      " - 0s - loss: 0.0848 - mean_absolute_error: 0.3555 - accuracy: 0.7297\n",
      "Epoch 148/200\n",
      " - 0s - loss: 0.0829 - mean_absolute_error: 0.3502 - accuracy: 0.7401\n",
      "Epoch 149/200\n",
      " - 0s - loss: 0.0847 - mean_absolute_error: 0.3532 - accuracy: 0.7378\n",
      "Epoch 150/200\n",
      " - 0s - loss: 0.0817 - mean_absolute_error: 0.3471 - accuracy: 0.7360\n",
      "Epoch 151/200\n",
      " - 0s - loss: 0.0834 - mean_absolute_error: 0.3498 - accuracy: 0.7343\n",
      "Epoch 152/200\n",
      " - 0s - loss: 0.0831 - mean_absolute_error: 0.3496 - accuracy: 0.7314\n",
      "Epoch 153/200\n",
      " - 0s - loss: 0.0838 - mean_absolute_error: 0.3527 - accuracy: 0.7349\n",
      "Epoch 154/200\n",
      " - 0s - loss: 0.0806 - mean_absolute_error: 0.3470 - accuracy: 0.7500\n",
      "Epoch 155/200\n",
      " - 0s - loss: 0.0841 - mean_absolute_error: 0.3524 - accuracy: 0.7378\n",
      "Epoch 156/200\n",
      " - 0s - loss: 0.0830 - mean_absolute_error: 0.3497 - accuracy: 0.7372\n",
      "Epoch 157/200\n",
      " - 0s - loss: 0.0844 - mean_absolute_error: 0.3536 - accuracy: 0.7302\n",
      "Epoch 158/200\n",
      " - 0s - loss: 0.0825 - mean_absolute_error: 0.3503 - accuracy: 0.7326\n",
      "Epoch 159/200\n",
      " - 0s - loss: 0.0814 - mean_absolute_error: 0.3458 - accuracy: 0.7372\n",
      "Epoch 160/200\n",
      " - 0s - loss: 0.0821 - mean_absolute_error: 0.3506 - accuracy: 0.7384\n",
      "Epoch 161/200\n",
      " - 0s - loss: 0.0824 - mean_absolute_error: 0.3487 - accuracy: 0.7390\n",
      "Epoch 162/200\n",
      " - 0s - loss: 0.0838 - mean_absolute_error: 0.3521 - accuracy: 0.7366\n",
      "Epoch 163/200\n",
      " - 0s - loss: 0.0823 - mean_absolute_error: 0.3480 - accuracy: 0.7471\n",
      "Epoch 164/200\n",
      " - 0s - loss: 0.0836 - mean_absolute_error: 0.3503 - accuracy: 0.7331\n",
      "Epoch 165/200\n",
      " - 0s - loss: 0.0809 - mean_absolute_error: 0.3414 - accuracy: 0.7384\n",
      "Epoch 166/200\n",
      " - 0s - loss: 0.0830 - mean_absolute_error: 0.3488 - accuracy: 0.7407\n",
      "Epoch 167/200\n",
      " - 0s - loss: 0.0824 - mean_absolute_error: 0.3444 - accuracy: 0.7517\n",
      "Epoch 168/200\n",
      " - 0s - loss: 0.0823 - mean_absolute_error: 0.3461 - accuracy: 0.7471\n",
      "Epoch 169/200\n",
      " - 0s - loss: 0.0820 - mean_absolute_error: 0.3465 - accuracy: 0.7378\n",
      "Epoch 170/200\n",
      " - 0s - loss: 0.0835 - mean_absolute_error: 0.3511 - accuracy: 0.7291\n",
      "Epoch 171/200\n",
      " - 0s - loss: 0.0812 - mean_absolute_error: 0.3457 - accuracy: 0.7360\n",
      "Epoch 172/200\n",
      " - 0s - loss: 0.0827 - mean_absolute_error: 0.3505 - accuracy: 0.7349\n",
      "Epoch 173/200\n",
      " - 0s - loss: 0.0819 - mean_absolute_error: 0.3480 - accuracy: 0.7378\n",
      "Epoch 174/200\n",
      " - 0s - loss: 0.0803 - mean_absolute_error: 0.3442 - accuracy: 0.7395\n",
      "Epoch 175/200\n",
      " - 0s - loss: 0.0823 - mean_absolute_error: 0.3485 - accuracy: 0.7407\n",
      "Epoch 176/200\n",
      " - 0s - loss: 0.0834 - mean_absolute_error: 0.3497 - accuracy: 0.7326\n",
      "Epoch 177/200\n",
      " - 0s - loss: 0.0817 - mean_absolute_error: 0.3454 - accuracy: 0.7471\n",
      "Epoch 178/200\n",
      " - 0s - loss: 0.0807 - mean_absolute_error: 0.3435 - accuracy: 0.7378\n",
      "Epoch 179/200\n",
      " - 0s - loss: 0.0827 - mean_absolute_error: 0.3450 - accuracy: 0.7360\n",
      "Epoch 180/200\n",
      " - 0s - loss: 0.0825 - mean_absolute_error: 0.3487 - accuracy: 0.7413\n",
      "Epoch 181/200\n",
      " - 0s - loss: 0.0818 - mean_absolute_error: 0.3454 - accuracy: 0.7407\n",
      "Epoch 182/200\n",
      " - 0s - loss: 0.0835 - mean_absolute_error: 0.3515 - accuracy: 0.7366\n",
      "Epoch 183/200\n",
      " - 0s - loss: 0.0828 - mean_absolute_error: 0.3489 - accuracy: 0.7384\n",
      "Epoch 184/200\n",
      " - 0s - loss: 0.0819 - mean_absolute_error: 0.3490 - accuracy: 0.7453\n",
      "Epoch 185/200\n",
      " - 0s - loss: 0.0817 - mean_absolute_error: 0.3481 - accuracy: 0.7424\n",
      "Epoch 186/200\n",
      " - 0s - loss: 0.0833 - mean_absolute_error: 0.3504 - accuracy: 0.7372\n",
      "Epoch 187/200\n",
      " - 0s - loss: 0.0819 - mean_absolute_error: 0.3496 - accuracy: 0.7465\n",
      "Epoch 188/200\n",
      " - 0s - loss: 0.0829 - mean_absolute_error: 0.3506 - accuracy: 0.7378\n",
      "Epoch 189/200\n",
      " - 0s - loss: 0.0825 - mean_absolute_error: 0.3487 - accuracy: 0.7285\n",
      "Epoch 190/200\n",
      " - 0s - loss: 0.0819 - mean_absolute_error: 0.3470 - accuracy: 0.7390\n",
      "Epoch 191/200\n",
      " - 0s - loss: 0.0818 - mean_absolute_error: 0.3470 - accuracy: 0.7465\n",
      "Epoch 192/200\n",
      " - 0s - loss: 0.0807 - mean_absolute_error: 0.3438 - accuracy: 0.7512\n",
      "Epoch 193/200\n",
      " - 0s - loss: 0.0807 - mean_absolute_error: 0.3441 - accuracy: 0.7384\n",
      "Epoch 194/200\n",
      " - 0s - loss: 0.0822 - mean_absolute_error: 0.3464 - accuracy: 0.7453\n",
      "Epoch 195/200\n",
      " - 0s - loss: 0.0806 - mean_absolute_error: 0.3425 - accuracy: 0.7419\n",
      "Epoch 196/200\n",
      " - 0s - loss: 0.0816 - mean_absolute_error: 0.3456 - accuracy: 0.7390\n",
      "Epoch 197/200\n",
      " - 0s - loss: 0.0837 - mean_absolute_error: 0.3514 - accuracy: 0.7378\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 198/200\n",
      " - 0s - loss: 0.0832 - mean_absolute_error: 0.3513 - accuracy: 0.7302\n",
      "Epoch 199/200\n",
      " - 0s - loss: 0.0827 - mean_absolute_error: 0.3504 - accuracy: 0.7320\n",
      "Epoch 200/200\n",
      " - 0s - loss: 0.0828 - mean_absolute_error: 0.3502 - accuracy: 0.7395\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.callbacks.History at 0x17763855cc0>"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Adagrad=optimizers.Adagrad(learning_rate=0.001)\n",
    "model2.compile(optimizer='Adagrad', loss='logcosh',metrics=[metrics.mae, 'accuracy'])\n",
    "model2.fit(X1, y1, batch_size=32, epochs=200, verbose=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "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"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['accuracy'])\n",
    "plt.title('Model accuracy')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Valid'], loc='upper left')\n",
    "plt.show()\n",
    "\n",
    "plt.plot(history.history['loss'])\n",
    "plt.title('Model loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Valid'], loc='upper left')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy on Test Data :- \n",
      "\t loss: 0.10720591111616655   mean_absolute_error: 0.462567538022995   accuracy: 0.6969696879386902\n"
     ]
    }
   ],
   "source": [
    "loss, mean_absolute_error, accuracy =model2.evaluate(X2, y2, batch_size=32, verbose=2)\n",
    "print('Accuracy on Test Data :- ')\n",
    "print('\\t loss:',loss , '  mean_absolute_error:',mean_absolute_error, '  accuracy:',accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__Optimizer: Adagrad is better suit for this model than RMSprop. Although, both are close enough in terms of accuracy and loss.__ "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Network Architecture"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "___Hyperparameters:___<br>\n",
    "filters=256<br>\n",
    "kernel_size=2<br>\n",
    "Activation functions=  relu, elu, softmax<br>\n",
    "loss=logcosh<br>\n",
    "optimizer= adagrad<br>\n",
    "epoch=100<br>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "# define model\n",
    "model3 = Sequential()\n",
    "\n",
    "model3.add(Conv1D(filters=256, kernel_size=3, activation='relu', input_shape=(n_steps, n_features)))\n",
    "model3.add(MaxPooling1D(pool_size=2))\n",
    "model3.add(BatchNormalization())\n",
    "\n",
    "model3.add(Dense(128, activation='elu'))\n",
    "model3.add(Dropout(0.25))\n",
    "\n",
    "model3.add(Conv1D(filters=128, kernel_size=2, activation='elu'))\n",
    "model3.add(MaxPooling1D(pool_size=2))\n",
    "model3.add(BatchNormalization())\n",
    "model3.add(Dropout(0.25))\n",
    "\n",
    "model3.add(Dense(64))\n",
    "model3.add(Dropout(0.25))\n",
    "\n",
    "\n",
    "model3.add(Flatten())\n",
    "\n",
    "model3.add(Dropout(0.25))\n",
    "model3.add(Dense(20, activation='softmax'))\n",
    "model3.add(Dropout(0.25))\n",
    "model3.add(Dense(1))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      " - 1s - loss: 0.1362 - mean_absolute_error: 0.4792 - accuracy: 0.5802\n",
      "Epoch 2/100\n",
      " - 0s - loss: 0.1251 - mean_absolute_error: 0.4585 - accuracy: 0.6192\n",
      "Epoch 3/100\n",
      " - 0s - loss: 0.1205 - mean_absolute_error: 0.4505 - accuracy: 0.6250\n",
      "Epoch 4/100\n",
      " - 0s - loss: 0.1157 - mean_absolute_error: 0.4411 - accuracy: 0.6273\n",
      "Epoch 5/100\n",
      " - 0s - loss: 0.1116 - mean_absolute_error: 0.4265 - accuracy: 0.6465\n",
      "Epoch 6/100\n",
      " - 0s - loss: 0.1092 - mean_absolute_error: 0.4188 - accuracy: 0.6541\n",
      "Epoch 7/100\n",
      " - 0s - loss: 0.1124 - mean_absolute_error: 0.4294 - accuracy: 0.6413\n",
      "Epoch 8/100\n",
      " - 0s - loss: 0.1117 - mean_absolute_error: 0.4319 - accuracy: 0.6349\n",
      "Epoch 9/100\n",
      " - 0s - loss: 0.1092 - mean_absolute_error: 0.4240 - accuracy: 0.6401\n",
      "Epoch 10/100\n",
      " - 0s - loss: 0.1126 - mean_absolute_error: 0.4282 - accuracy: 0.6291\n",
      "Epoch 11/100\n",
      " - 0s - loss: 0.1059 - mean_absolute_error: 0.4152 - accuracy: 0.6576\n",
      "Epoch 12/100\n",
      " - 0s - loss: 0.1081 - mean_absolute_error: 0.4142 - accuracy: 0.6483\n",
      "Epoch 13/100\n",
      " - 0s - loss: 0.1068 - mean_absolute_error: 0.4159 - accuracy: 0.6552\n",
      "Epoch 14/100\n",
      " - 0s - loss: 0.1099 - mean_absolute_error: 0.4269 - accuracy: 0.6256\n",
      "Epoch 15/100\n",
      " - 0s - loss: 0.1055 - mean_absolute_error: 0.4177 - accuracy: 0.6517\n",
      "Epoch 16/100\n",
      " - 0s - loss: 0.1061 - mean_absolute_error: 0.4202 - accuracy: 0.6529\n",
      "Epoch 17/100\n",
      " - 0s - loss: 0.1039 - mean_absolute_error: 0.4150 - accuracy: 0.6628\n",
      "Epoch 18/100\n",
      " - 0s - loss: 0.1029 - mean_absolute_error: 0.4152 - accuracy: 0.6628\n",
      "Epoch 19/100\n",
      " - 0s - loss: 0.1041 - mean_absolute_error: 0.4198 - accuracy: 0.6599\n",
      "Epoch 20/100\n",
      " - 0s - loss: 0.1039 - mean_absolute_error: 0.4179 - accuracy: 0.6471\n",
      "Epoch 21/100\n",
      " - 0s - loss: 0.1034 - mean_absolute_error: 0.4136 - accuracy: 0.6558\n",
      "Epoch 22/100\n",
      " - 0s - loss: 0.1023 - mean_absolute_error: 0.4106 - accuracy: 0.6657\n",
      "Epoch 23/100\n",
      " - 0s - loss: 0.1008 - mean_absolute_error: 0.4057 - accuracy: 0.6703\n",
      "Epoch 24/100\n",
      " - 0s - loss: 0.1028 - mean_absolute_error: 0.4141 - accuracy: 0.6564\n",
      "Epoch 25/100\n",
      " - 0s - loss: 0.0997 - mean_absolute_error: 0.4059 - accuracy: 0.6773\n",
      "Epoch 26/100\n",
      " - 0s - loss: 0.1004 - mean_absolute_error: 0.4032 - accuracy: 0.6750\n",
      "Epoch 27/100\n",
      " - 0s - loss: 0.1010 - mean_absolute_error: 0.4097 - accuracy: 0.6686\n",
      "Epoch 28/100\n",
      " - 0s - loss: 0.0998 - mean_absolute_error: 0.4077 - accuracy: 0.6715\n",
      "Epoch 29/100\n",
      " - 0s - loss: 0.1010 - mean_absolute_error: 0.4076 - accuracy: 0.6651\n",
      "Epoch 30/100\n",
      " - 0s - loss: 0.1006 - mean_absolute_error: 0.4100 - accuracy: 0.6703\n",
      "Epoch 31/100\n",
      " - 0s - loss: 0.1027 - mean_absolute_error: 0.4162 - accuracy: 0.6558\n",
      "Epoch 32/100\n",
      " - 0s - loss: 0.0994 - mean_absolute_error: 0.4090 - accuracy: 0.6727\n",
      "Epoch 33/100\n",
      " - 0s - loss: 0.1025 - mean_absolute_error: 0.4151 - accuracy: 0.6552\n",
      "Epoch 34/100\n",
      " - 0s - loss: 0.1019 - mean_absolute_error: 0.4143 - accuracy: 0.6651\n",
      "Epoch 35/100\n",
      " - 0s - loss: 0.1009 - mean_absolute_error: 0.4107 - accuracy: 0.6744\n",
      "Epoch 36/100\n",
      " - 0s - loss: 0.1014 - mean_absolute_error: 0.4139 - accuracy: 0.6709\n",
      "Epoch 37/100\n",
      " - 0s - loss: 0.0992 - mean_absolute_error: 0.4061 - accuracy: 0.6907\n",
      "Epoch 38/100\n",
      " - 0s - loss: 0.0986 - mean_absolute_error: 0.4064 - accuracy: 0.6831\n",
      "Epoch 39/100\n",
      " - 0s - loss: 0.0996 - mean_absolute_error: 0.4125 - accuracy: 0.6785\n",
      "Epoch 40/100\n",
      " - 0s - loss: 0.0999 - mean_absolute_error: 0.4108 - accuracy: 0.6808\n",
      "Epoch 41/100\n",
      " - 0s - loss: 0.0981 - mean_absolute_error: 0.4075 - accuracy: 0.6837\n",
      "Epoch 42/100\n",
      " - 0s - loss: 0.0997 - mean_absolute_error: 0.4102 - accuracy: 0.6773\n",
      "Epoch 43/100\n",
      " - 0s - loss: 0.0992 - mean_absolute_error: 0.4089 - accuracy: 0.6791\n",
      "Epoch 44/100\n",
      " - 0s - loss: 0.0997 - mean_absolute_error: 0.4113 - accuracy: 0.6744\n",
      "Epoch 45/100\n",
      " - 0s - loss: 0.0992 - mean_absolute_error: 0.4110 - accuracy: 0.6744\n",
      "Epoch 46/100\n",
      " - 0s - loss: 0.0980 - mean_absolute_error: 0.4066 - accuracy: 0.6907\n",
      "Epoch 47/100\n",
      " - 0s - loss: 0.0979 - mean_absolute_error: 0.4075 - accuracy: 0.6872\n",
      "Epoch 48/100\n",
      " - 0s - loss: 0.0994 - mean_absolute_error: 0.4093 - accuracy: 0.6797\n",
      "Epoch 49/100\n",
      " - 0s - loss: 0.0981 - mean_absolute_error: 0.4042 - accuracy: 0.6843\n",
      "Epoch 50/100\n",
      " - 0s - loss: 0.0964 - mean_absolute_error: 0.3997 - accuracy: 0.6866\n",
      "Epoch 51/100\n",
      " - 0s - loss: 0.0982 - mean_absolute_error: 0.4019 - accuracy: 0.6837\n",
      "Epoch 52/100\n",
      " - 0s - loss: 0.0967 - mean_absolute_error: 0.3995 - accuracy: 0.6924\n",
      "Epoch 53/100\n",
      " - 0s - loss: 0.0979 - mean_absolute_error: 0.4020 - accuracy: 0.6872\n",
      "Epoch 54/100\n",
      " - 0s - loss: 0.0973 - mean_absolute_error: 0.3992 - accuracy: 0.6895\n",
      "Epoch 55/100\n",
      " - 0s - loss: 0.0985 - mean_absolute_error: 0.4053 - accuracy: 0.6797\n",
      "Epoch 56/100\n",
      " - 0s - loss: 0.0978 - mean_absolute_error: 0.4049 - accuracy: 0.6837\n",
      "Epoch 57/100\n",
      " - 0s - loss: 0.0987 - mean_absolute_error: 0.4065 - accuracy: 0.6831\n",
      "Epoch 58/100\n",
      " - 0s - loss: 0.0981 - mean_absolute_error: 0.4040 - accuracy: 0.6843\n",
      "Epoch 59/100\n",
      " - 0s - loss: 0.0971 - mean_absolute_error: 0.4019 - accuracy: 0.6802\n",
      "Epoch 60/100\n",
      " - 0s - loss: 0.0988 - mean_absolute_error: 0.4068 - accuracy: 0.6855\n",
      "Epoch 61/100\n",
      " - 0s - loss: 0.0966 - mean_absolute_error: 0.4003 - accuracy: 0.6913\n",
      "Epoch 62/100\n",
      " - 0s - loss: 0.0976 - mean_absolute_error: 0.4026 - accuracy: 0.6930\n",
      "Epoch 63/100\n",
      " - 0s - loss: 0.0968 - mean_absolute_error: 0.4000 - accuracy: 0.6965\n",
      "Epoch 64/100\n",
      " - 0s - loss: 0.0968 - mean_absolute_error: 0.4010 - accuracy: 0.6913\n",
      "Epoch 65/100\n",
      " - 0s - loss: 0.0988 - mean_absolute_error: 0.4082 - accuracy: 0.6977\n",
      "Epoch 66/100\n",
      " - 0s - loss: 0.0952 - mean_absolute_error: 0.3973 - accuracy: 0.7012\n",
      "Epoch 67/100\n",
      " - 0s - loss: 0.0950 - mean_absolute_error: 0.3969 - accuracy: 0.7006\n",
      "Epoch 68/100\n",
      " - 0s - loss: 0.0981 - mean_absolute_error: 0.4044 - accuracy: 0.6913\n",
      "Epoch 69/100\n",
      " - 0s - loss: 0.0943 - mean_absolute_error: 0.3929 - accuracy: 0.7081\n",
      "Epoch 70/100\n",
      " - 0s - loss: 0.0961 - mean_absolute_error: 0.4003 - accuracy: 0.7012\n",
      "Epoch 71/100\n",
      " - 0s - loss: 0.0960 - mean_absolute_error: 0.3990 - accuracy: 0.6924\n",
      "Epoch 72/100\n",
      " - 0s - loss: 0.0961 - mean_absolute_error: 0.3987 - accuracy: 0.6930\n",
      "Epoch 73/100\n",
      " - 0s - loss: 0.0965 - mean_absolute_error: 0.3985 - accuracy: 0.6953\n",
      "Epoch 74/100\n",
      " - 0s - loss: 0.0956 - mean_absolute_error: 0.3956 - accuracy: 0.6983\n",
      "Epoch 75/100\n",
      " - 0s - loss: 0.0948 - mean_absolute_error: 0.3950 - accuracy: 0.6994\n",
      "Epoch 76/100\n",
      " - 0s - loss: 0.0958 - mean_absolute_error: 0.3965 - accuracy: 0.7052\n",
      "Epoch 77/100\n",
      " - 0s - loss: 0.0969 - mean_absolute_error: 0.3966 - accuracy: 0.6977\n",
      "Epoch 78/100\n",
      " - 0s - loss: 0.0945 - mean_absolute_error: 0.3943 - accuracy: 0.6988\n",
      "Epoch 79/100\n",
      " - 0s - loss: 0.0960 - mean_absolute_error: 0.3989 - accuracy: 0.7064\n",
      "Epoch 80/100\n",
      " - 0s - loss: 0.0948 - mean_absolute_error: 0.3953 - accuracy: 0.7064\n",
      "Epoch 81/100\n",
      " - 0s - loss: 0.0962 - mean_absolute_error: 0.3984 - accuracy: 0.6965\n",
      "Epoch 82/100\n",
      " - 0s - loss: 0.0950 - mean_absolute_error: 0.3979 - accuracy: 0.7006\n",
      "Epoch 83/100\n",
      " - 0s - loss: 0.0948 - mean_absolute_error: 0.3963 - accuracy: 0.7000\n",
      "Epoch 84/100\n",
      " - 0s - loss: 0.0959 - mean_absolute_error: 0.3990 - accuracy: 0.6977\n",
      "Epoch 85/100\n",
      " - 0s - loss: 0.0958 - mean_absolute_error: 0.4004 - accuracy: 0.6971\n",
      "Epoch 86/100\n",
      " - 0s - loss: 0.0948 - mean_absolute_error: 0.3960 - accuracy: 0.7029\n",
      "Epoch 87/100\n",
      " - 0s - loss: 0.0969 - mean_absolute_error: 0.4006 - accuracy: 0.6953\n",
      "Epoch 88/100\n",
      " - 0s - loss: 0.0962 - mean_absolute_error: 0.3999 - accuracy: 0.7029\n",
      "Epoch 89/100\n",
      " - 0s - loss: 0.0958 - mean_absolute_error: 0.3991 - accuracy: 0.7006\n",
      "Epoch 90/100\n",
      " - 0s - loss: 0.0965 - mean_absolute_error: 0.4011 - accuracy: 0.6942\n",
      "Epoch 91/100\n",
      " - 0s - loss: 0.0960 - mean_absolute_error: 0.3981 - accuracy: 0.7000\n",
      "Epoch 92/100\n",
      " - 0s - loss: 0.0941 - mean_absolute_error: 0.3962 - accuracy: 0.7076\n",
      "Epoch 93/100\n",
      " - 0s - loss: 0.0974 - mean_absolute_error: 0.4009 - accuracy: 0.6948\n",
      "Epoch 94/100\n",
      " - 0s - loss: 0.0946 - mean_absolute_error: 0.3961 - accuracy: 0.7041\n",
      "Epoch 95/100\n",
      " - 0s - loss: 0.0962 - mean_absolute_error: 0.3993 - accuracy: 0.6924\n",
      "Epoch 96/100\n",
      " - 0s - loss: 0.0935 - mean_absolute_error: 0.3946 - accuracy: 0.7116\n",
      "Epoch 97/100\n",
      " - 0s - loss: 0.0970 - mean_absolute_error: 0.4036 - accuracy: 0.6919\n",
      "Epoch 98/100\n",
      " - 0s - loss: 0.0957 - mean_absolute_error: 0.4002 - accuracy: 0.6965\n",
      "Epoch 99/100\n",
      " - 0s - loss: 0.0955 - mean_absolute_error: 0.3984 - accuracy: 0.7012\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 100/100\n",
      " - 0s - loss: 0.0955 - mean_absolute_error: 0.3990 - accuracy: 0.7035\n"
     ]
    }
   ],
   "source": [
    "Adagrad=optimizers.Adagrad(learning_rate=0.4)\n",
    "model3.compile(optimizer='Adagrad', loss='logcosh',metrics=[metrics.mae, 'accuracy'])\n",
    "history=model3.fit(X1, y1, batch_size=32, epochs=100, verbose=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "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"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['accuracy'])\n",
    "plt.title('Model accuracy')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Valid'], loc='upper left')\n",
    "plt.show()\n",
    "\n",
    "plt.plot(history.history['loss'])\n",
    "plt.title('Model loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Valid'], loc='upper left')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy on Test Data :- \n",
      "\t loss: 0.09973003647544167   mean_absolute_error: 0.4202899932861328   accuracy: 0.6969696879386902\n"
     ]
    }
   ],
   "source": [
    "loss, mean_absolute_error, accuracy =model3.evaluate(X2, y2, batch_size=32, verbose=2)\n",
    "print('Accuracy on Test Data :- ')\n",
    "print('\\t loss:',loss , '  mean_absolute_error:',mean_absolute_error, '  accuracy:',accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__After changing the number of layers and the size of the layers, both accuracy and loss got improve.__"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Network initialization"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "___Hyperparameters:___<br>\n",
    "filters=256<br>\n",
    "kernel_size=3,2<br>\n",
    "Activation functions= relu, elu, softmax<br>\n",
    "kernel_initializer=Ones, RandomNormal <br>\n",
    "loss=mean_squared_error<br>\n",
    "optimizer= adam<br>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "# define model\n",
    "# Default : kernel_initializer='glorot_uniform'\n",
    "\n",
    "model4 = Sequential()\n",
    "\n",
    "model4.add(Conv1D(filters=256, kernel_size=3, activation='relu', kernel_initializer='Ones' , input_shape=(n_steps, n_features)))\n",
    "model4.add(MaxPooling1D(pool_size=2))\n",
    "model4.add(BatchNormalization())\n",
    "\n",
    "\n",
    "model4.add(Conv1D(filters=128, kernel_size=2, activation='elu',kernel_initializer='RandomNormal'))\n",
    "model4.add(MaxPooling1D(pool_size=2))\n",
    "model4.add(BatchNormalization())\n",
    "model4.add(Dropout(0.25))\n",
    "\n",
    "\n",
    "model4.add(Flatten())\n",
    "\n",
    "model4.add(Dropout(0.25))\n",
    "model4.add(Dense(20, activation='softmax'))\n",
    "model4.add(Dropout(0.25))\n",
    "model4.add(Dense(1))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      " - 1s - loss: 0.1322 - mean_absolute_error: 0.4732 - accuracy: 0.5826\n",
      "Epoch 2/200\n",
      " - 0s - loss: 0.1175 - mean_absolute_error: 0.4545 - accuracy: 0.6122\n",
      "Epoch 3/200\n",
      " - 0s - loss: 0.1136 - mean_absolute_error: 0.4429 - accuracy: 0.6227\n",
      "Epoch 4/200\n",
      " - 0s - loss: 0.1130 - mean_absolute_error: 0.4416 - accuracy: 0.6285\n",
      "Epoch 5/200\n",
      " - 0s - loss: 0.1093 - mean_absolute_error: 0.4321 - accuracy: 0.6436\n",
      "Epoch 6/200\n",
      " - 0s - loss: 0.1131 - mean_absolute_error: 0.4369 - accuracy: 0.6291\n",
      "Epoch 7/200\n",
      " - 0s - loss: 0.1115 - mean_absolute_error: 0.4332 - accuracy: 0.6337\n",
      "Epoch 8/200\n",
      " - 0s - loss: 0.1065 - mean_absolute_error: 0.4215 - accuracy: 0.6558\n",
      "Epoch 9/200\n",
      " - 0s - loss: 0.1138 - mean_absolute_error: 0.4393 - accuracy: 0.6250\n",
      "Epoch 10/200\n",
      " - 0s - loss: 0.1084 - mean_absolute_error: 0.4250 - accuracy: 0.6494\n",
      "Epoch 11/200\n",
      " - 0s - loss: 0.1106 - mean_absolute_error: 0.4329 - accuracy: 0.6430\n",
      "Epoch 12/200\n",
      " - 0s - loss: 0.1090 - mean_absolute_error: 0.4283 - accuracy: 0.6384\n",
      "Epoch 13/200\n",
      " - 0s - loss: 0.1090 - mean_absolute_error: 0.4292 - accuracy: 0.6448\n",
      "Epoch 14/200\n",
      " - 0s - loss: 0.1059 - mean_absolute_error: 0.4231 - accuracy: 0.6517\n",
      "Epoch 15/200\n",
      " - 0s - loss: 0.1077 - mean_absolute_error: 0.4260 - accuracy: 0.6471\n",
      "Epoch 16/200\n",
      " - 0s - loss: 0.1086 - mean_absolute_error: 0.4298 - accuracy: 0.6413\n",
      "Epoch 17/200\n",
      " - 0s - loss: 0.1086 - mean_absolute_error: 0.4308 - accuracy: 0.6419\n",
      "Epoch 18/200\n",
      " - 0s - loss: 0.1094 - mean_absolute_error: 0.4320 - accuracy: 0.6343\n",
      "Epoch 19/200\n",
      " - 0s - loss: 0.1048 - mean_absolute_error: 0.4219 - accuracy: 0.6570\n",
      "Epoch 20/200\n",
      " - 0s - loss: 0.1084 - mean_absolute_error: 0.4293 - accuracy: 0.6360\n",
      "Epoch 21/200\n",
      " - 0s - loss: 0.1077 - mean_absolute_error: 0.4286 - accuracy: 0.6494\n",
      "Epoch 22/200\n",
      " - 0s - loss: 0.1074 - mean_absolute_error: 0.4270 - accuracy: 0.6477\n",
      "Epoch 23/200\n",
      " - 0s - loss: 0.1084 - mean_absolute_error: 0.4301 - accuracy: 0.6500\n",
      "Epoch 24/200\n",
      " - 0s - loss: 0.1043 - mean_absolute_error: 0.4194 - accuracy: 0.6610\n",
      "Epoch 25/200\n",
      " - 0s - loss: 0.1075 - mean_absolute_error: 0.4262 - accuracy: 0.6529\n",
      "Epoch 26/200\n",
      " - 0s - loss: 0.1062 - mean_absolute_error: 0.4258 - accuracy: 0.6523\n",
      "Epoch 27/200\n",
      " - 0s - loss: 0.1046 - mean_absolute_error: 0.4190 - accuracy: 0.6564\n",
      "Epoch 28/200\n",
      " - 0s - loss: 0.1082 - mean_absolute_error: 0.4317 - accuracy: 0.6459\n",
      "Epoch 29/200\n",
      " - 0s - loss: 0.1069 - mean_absolute_error: 0.4293 - accuracy: 0.6407\n",
      "Epoch 30/200\n",
      " - 0s - loss: 0.1073 - mean_absolute_error: 0.4296 - accuracy: 0.6407\n",
      "Epoch 31/200\n",
      " - 0s - loss: 0.1067 - mean_absolute_error: 0.4299 - accuracy: 0.6442\n",
      "Epoch 32/200\n",
      " - 0s - loss: 0.1065 - mean_absolute_error: 0.4357 - accuracy: 0.6459\n",
      "Epoch 33/200\n",
      " - 0s - loss: 0.1078 - mean_absolute_error: 0.4361 - accuracy: 0.6343\n",
      "Epoch 34/200\n",
      " - 0s - loss: 0.1041 - mean_absolute_error: 0.4272 - accuracy: 0.6523\n",
      "Epoch 35/200\n",
      " - 0s - loss: 0.1068 - mean_absolute_error: 0.4343 - accuracy: 0.6471\n",
      "Epoch 36/200\n",
      " - 0s - loss: 0.1052 - mean_absolute_error: 0.4336 - accuracy: 0.6512\n",
      "Epoch 37/200\n",
      " - 0s - loss: 0.1036 - mean_absolute_error: 0.4288 - accuracy: 0.6593\n",
      "Epoch 38/200\n",
      " - 0s - loss: 0.1046 - mean_absolute_error: 0.4299 - accuracy: 0.6558\n",
      "Epoch 39/200\n",
      " - 0s - loss: 0.1022 - mean_absolute_error: 0.4250 - accuracy: 0.6640\n",
      "Epoch 40/200\n",
      " - 0s - loss: 0.1040 - mean_absolute_error: 0.4308 - accuracy: 0.6622\n",
      "Epoch 41/200\n",
      " - 0s - loss: 0.1032 - mean_absolute_error: 0.4260 - accuracy: 0.6570\n",
      "Epoch 42/200\n",
      " - 0s - loss: 0.1040 - mean_absolute_error: 0.4319 - accuracy: 0.6640\n",
      "Epoch 43/200\n",
      " - 0s - loss: 0.1036 - mean_absolute_error: 0.4299 - accuracy: 0.6605\n",
      "Epoch 44/200\n",
      " - 0s - loss: 0.1023 - mean_absolute_error: 0.4258 - accuracy: 0.6616\n",
      "Epoch 45/200\n",
      " - 0s - loss: 0.1055 - mean_absolute_error: 0.4342 - accuracy: 0.6506\n",
      "Epoch 46/200\n",
      " - 0s - loss: 0.1012 - mean_absolute_error: 0.4219 - accuracy: 0.6692\n",
      "Epoch 47/200\n",
      " - 0s - loss: 0.1035 - mean_absolute_error: 0.4230 - accuracy: 0.6581\n",
      "Epoch 48/200\n",
      " - 0s - loss: 0.1013 - mean_absolute_error: 0.4214 - accuracy: 0.6680\n",
      "Epoch 49/200\n",
      " - 0s - loss: 0.1020 - mean_absolute_error: 0.4186 - accuracy: 0.6674\n",
      "Epoch 50/200\n",
      " - 0s - loss: 0.1020 - mean_absolute_error: 0.4194 - accuracy: 0.6610\n",
      "Epoch 51/200\n",
      " - 0s - loss: 0.1039 - mean_absolute_error: 0.4252 - accuracy: 0.6669\n",
      "Epoch 52/200\n",
      " - 0s - loss: 0.1013 - mean_absolute_error: 0.4199 - accuracy: 0.6686\n",
      "Epoch 53/200\n",
      " - 0s - loss: 0.1004 - mean_absolute_error: 0.4183 - accuracy: 0.6733\n",
      "Epoch 54/200\n",
      " - 0s - loss: 0.1020 - mean_absolute_error: 0.4180 - accuracy: 0.6669\n",
      "Epoch 55/200\n",
      " - 0s - loss: 0.1011 - mean_absolute_error: 0.4162 - accuracy: 0.6802\n",
      "Epoch 56/200\n",
      " - 0s - loss: 0.1013 - mean_absolute_error: 0.4162 - accuracy: 0.6727\n",
      "Epoch 57/200\n",
      " - 0s - loss: 0.1030 - mean_absolute_error: 0.4241 - accuracy: 0.6587\n",
      "Epoch 58/200\n",
      " - 0s - loss: 0.1022 - mean_absolute_error: 0.4216 - accuracy: 0.6674\n",
      "Epoch 59/200\n",
      " - 0s - loss: 0.1005 - mean_absolute_error: 0.4188 - accuracy: 0.6663\n",
      "Epoch 60/200\n",
      " - 0s - loss: 0.1028 - mean_absolute_error: 0.4225 - accuracy: 0.6616\n",
      "Epoch 61/200\n",
      " - 0s - loss: 0.1016 - mean_absolute_error: 0.4193 - accuracy: 0.6686\n",
      "Epoch 62/200\n",
      " - 0s - loss: 0.1035 - mean_absolute_error: 0.4278 - accuracy: 0.6506\n",
      "Epoch 63/200\n",
      " - 0s - loss: 0.1011 - mean_absolute_error: 0.4214 - accuracy: 0.6651\n",
      "Epoch 64/200\n",
      " - 0s - loss: 0.1009 - mean_absolute_error: 0.4194 - accuracy: 0.6634\n",
      "Epoch 65/200\n",
      " - 0s - loss: 0.1004 - mean_absolute_error: 0.4186 - accuracy: 0.6756\n",
      "Epoch 66/200\n",
      " - 0s - loss: 0.1020 - mean_absolute_error: 0.4246 - accuracy: 0.6715\n",
      "Epoch 67/200\n",
      " - 0s - loss: 0.1009 - mean_absolute_error: 0.4199 - accuracy: 0.6744\n",
      "Epoch 68/200\n",
      " - 0s - loss: 0.1027 - mean_absolute_error: 0.4302 - accuracy: 0.6640\n",
      "Epoch 69/200\n",
      " - 0s - loss: 0.1019 - mean_absolute_error: 0.4264 - accuracy: 0.6721\n",
      "Epoch 70/200\n",
      " - 0s - loss: 0.1011 - mean_absolute_error: 0.4226 - accuracy: 0.6674\n",
      "Epoch 71/200\n",
      " - 0s - loss: 0.0998 - mean_absolute_error: 0.4152 - accuracy: 0.6727\n",
      "Epoch 72/200\n",
      " - 0s - loss: 0.1008 - mean_absolute_error: 0.4200 - accuracy: 0.6814\n",
      "Epoch 73/200\n",
      " - 0s - loss: 0.0998 - mean_absolute_error: 0.4144 - accuracy: 0.6698\n",
      "Epoch 74/200\n",
      " - 0s - loss: 0.1022 - mean_absolute_error: 0.4209 - accuracy: 0.6674\n",
      "Epoch 75/200\n",
      " - 0s - loss: 0.1023 - mean_absolute_error: 0.4235 - accuracy: 0.6669\n",
      "Epoch 76/200\n",
      " - 0s - loss: 0.0999 - mean_absolute_error: 0.4162 - accuracy: 0.6767\n",
      "Epoch 77/200\n",
      " - 0s - loss: 0.1019 - mean_absolute_error: 0.4229 - accuracy: 0.6605\n",
      "Epoch 78/200\n",
      " - 0s - loss: 0.0999 - mean_absolute_error: 0.4186 - accuracy: 0.6733\n",
      "Epoch 79/200\n",
      " - 0s - loss: 0.1011 - mean_absolute_error: 0.4224 - accuracy: 0.6692\n",
      "Epoch 80/200\n",
      " - 0s - loss: 0.1007 - mean_absolute_error: 0.4190 - accuracy: 0.6750\n",
      "Epoch 81/200\n",
      " - 0s - loss: 0.1004 - mean_absolute_error: 0.4180 - accuracy: 0.6727\n",
      "Epoch 82/200\n",
      " - 0s - loss: 0.1005 - mean_absolute_error: 0.4200 - accuracy: 0.6715\n",
      "Epoch 83/200\n",
      " - 0s - loss: 0.1013 - mean_absolute_error: 0.4222 - accuracy: 0.6669\n",
      "Epoch 84/200\n",
      " - 0s - loss: 0.1000 - mean_absolute_error: 0.4194 - accuracy: 0.6709\n",
      "Epoch 85/200\n",
      " - 0s - loss: 0.1013 - mean_absolute_error: 0.4252 - accuracy: 0.6663\n",
      "Epoch 86/200\n",
      " - 0s - loss: 0.0996 - mean_absolute_error: 0.4178 - accuracy: 0.6762\n",
      "Epoch 87/200\n",
      " - 0s - loss: 0.0999 - mean_absolute_error: 0.4170 - accuracy: 0.6767\n",
      "Epoch 88/200\n",
      " - 0s - loss: 0.1009 - mean_absolute_error: 0.4184 - accuracy: 0.6593\n",
      "Epoch 89/200\n",
      " - 0s - loss: 0.0986 - mean_absolute_error: 0.4139 - accuracy: 0.6831\n",
      "Epoch 90/200\n",
      " - 0s - loss: 0.0994 - mean_absolute_error: 0.4170 - accuracy: 0.6802\n",
      "Epoch 91/200\n",
      " - 0s - loss: 0.1004 - mean_absolute_error: 0.4177 - accuracy: 0.6709\n",
      "Epoch 92/200\n",
      " - 0s - loss: 0.1007 - mean_absolute_error: 0.4161 - accuracy: 0.6698\n",
      "Epoch 93/200\n",
      " - 0s - loss: 0.1007 - mean_absolute_error: 0.4165 - accuracy: 0.6645\n",
      "Epoch 94/200\n",
      " - 0s - loss: 0.0988 - mean_absolute_error: 0.4122 - accuracy: 0.6791\n",
      "Epoch 95/200\n",
      " - 0s - loss: 0.1001 - mean_absolute_error: 0.4157 - accuracy: 0.6756\n",
      "Epoch 96/200\n",
      " - 0s - loss: 0.0988 - mean_absolute_error: 0.4152 - accuracy: 0.6872\n",
      "Epoch 97/200\n",
      " - 0s - loss: 0.0996 - mean_absolute_error: 0.4151 - accuracy: 0.6703\n",
      "Epoch 98/200\n",
      " - 0s - loss: 0.1008 - mean_absolute_error: 0.4180 - accuracy: 0.6651\n",
      "Epoch 99/200\n",
      " - 0s - loss: 0.1002 - mean_absolute_error: 0.4170 - accuracy: 0.6773\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 100/200\n",
      " - 0s - loss: 0.1008 - mean_absolute_error: 0.4208 - accuracy: 0.6762\n",
      "Epoch 101/200\n",
      " - 0s - loss: 0.0994 - mean_absolute_error: 0.4159 - accuracy: 0.6785\n",
      "Epoch 102/200\n",
      " - 0s - loss: 0.1005 - mean_absolute_error: 0.4174 - accuracy: 0.6773\n",
      "Epoch 103/200\n",
      " - 0s - loss: 0.0986 - mean_absolute_error: 0.4126 - accuracy: 0.6767\n",
      "Epoch 104/200\n",
      " - 0s - loss: 0.0995 - mean_absolute_error: 0.4150 - accuracy: 0.6814\n",
      "Epoch 105/200\n",
      " - 0s - loss: 0.0990 - mean_absolute_error: 0.4107 - accuracy: 0.6785\n",
      "Epoch 106/200\n",
      " - 0s - loss: 0.0987 - mean_absolute_error: 0.4130 - accuracy: 0.6843\n",
      "Epoch 107/200\n",
      " - 0s - loss: 0.0997 - mean_absolute_error: 0.4153 - accuracy: 0.6802\n",
      "Epoch 108/200\n",
      " - 0s - loss: 0.1000 - mean_absolute_error: 0.4134 - accuracy: 0.6727\n",
      "Epoch 109/200\n",
      " - 0s - loss: 0.0994 - mean_absolute_error: 0.4127 - accuracy: 0.6802\n",
      "Epoch 110/200\n",
      " - 0s - loss: 0.0994 - mean_absolute_error: 0.4138 - accuracy: 0.6727\n",
      "Epoch 111/200\n",
      " - 0s - loss: 0.0989 - mean_absolute_error: 0.4125 - accuracy: 0.6756\n",
      "Epoch 112/200\n",
      " - 0s - loss: 0.0984 - mean_absolute_error: 0.4131 - accuracy: 0.6826\n",
      "Epoch 113/200\n",
      " - 0s - loss: 0.0993 - mean_absolute_error: 0.4141 - accuracy: 0.6698\n",
      "Epoch 114/200\n",
      " - 0s - loss: 0.0978 - mean_absolute_error: 0.4082 - accuracy: 0.6756\n",
      "Epoch 115/200\n",
      " - 0s - loss: 0.0995 - mean_absolute_error: 0.4116 - accuracy: 0.6808\n",
      "Epoch 116/200\n",
      " - 0s - loss: 0.0980 - mean_absolute_error: 0.4096 - accuracy: 0.6785\n",
      "Epoch 117/200\n",
      " - 0s - loss: 0.0981 - mean_absolute_error: 0.4093 - accuracy: 0.6808\n",
      "Epoch 118/200\n",
      " - 0s - loss: 0.0991 - mean_absolute_error: 0.4096 - accuracy: 0.6791\n",
      "Epoch 119/200\n",
      " - 0s - loss: 0.1002 - mean_absolute_error: 0.4109 - accuracy: 0.6785\n",
      "Epoch 120/200\n",
      " - 0s - loss: 0.1001 - mean_absolute_error: 0.4146 - accuracy: 0.6750\n",
      "Epoch 121/200\n",
      " - 0s - loss: 0.0986 - mean_absolute_error: 0.4100 - accuracy: 0.6820\n",
      "Epoch 122/200\n",
      " - 0s - loss: 0.0986 - mean_absolute_error: 0.4101 - accuracy: 0.6797\n",
      "Epoch 123/200\n",
      " - 0s - loss: 0.1001 - mean_absolute_error: 0.4138 - accuracy: 0.6698\n",
      "Epoch 124/200\n",
      " - 0s - loss: 0.0993 - mean_absolute_error: 0.4129 - accuracy: 0.6820\n",
      "Epoch 125/200\n",
      " - 0s - loss: 0.0986 - mean_absolute_error: 0.4111 - accuracy: 0.6750\n",
      "Epoch 126/200\n",
      " - 0s - loss: 0.0986 - mean_absolute_error: 0.4115 - accuracy: 0.6837\n",
      "Epoch 127/200\n",
      " - 0s - loss: 0.0995 - mean_absolute_error: 0.4119 - accuracy: 0.6762\n",
      "Epoch 128/200\n",
      " - 0s - loss: 0.0987 - mean_absolute_error: 0.4123 - accuracy: 0.6901\n",
      "Epoch 129/200\n",
      " - 0s - loss: 0.0978 - mean_absolute_error: 0.4111 - accuracy: 0.6715\n",
      "Epoch 130/200\n",
      " - 0s - loss: 0.1001 - mean_absolute_error: 0.4154 - accuracy: 0.6738\n",
      "Epoch 131/200\n",
      " - 0s - loss: 0.0983 - mean_absolute_error: 0.4125 - accuracy: 0.6826\n",
      "Epoch 132/200\n",
      " - 0s - loss: 0.0996 - mean_absolute_error: 0.4126 - accuracy: 0.6826\n",
      "Epoch 133/200\n",
      " - 0s - loss: 0.0978 - mean_absolute_error: 0.4105 - accuracy: 0.6872\n",
      "Epoch 134/200\n",
      " - 0s - loss: 0.0979 - mean_absolute_error: 0.4069 - accuracy: 0.6791\n",
      "Epoch 135/200\n",
      " - 0s - loss: 0.0996 - mean_absolute_error: 0.4137 - accuracy: 0.6756\n",
      "Epoch 136/200\n",
      " - 0s - loss: 0.1000 - mean_absolute_error: 0.4176 - accuracy: 0.6721\n",
      "Epoch 137/200\n",
      " - 0s - loss: 0.0999 - mean_absolute_error: 0.4135 - accuracy: 0.6767\n",
      "Epoch 138/200\n",
      " - 0s - loss: 0.0994 - mean_absolute_error: 0.4136 - accuracy: 0.6785\n",
      "Epoch 139/200\n",
      " - 0s - loss: 0.0992 - mean_absolute_error: 0.4112 - accuracy: 0.6756\n",
      "Epoch 140/200\n",
      " - 0s - loss: 0.0985 - mean_absolute_error: 0.4106 - accuracy: 0.6756\n",
      "Epoch 141/200\n",
      " - 0s - loss: 0.0982 - mean_absolute_error: 0.4098 - accuracy: 0.6797\n",
      "Epoch 142/200\n",
      " - 0s - loss: 0.0973 - mean_absolute_error: 0.4058 - accuracy: 0.6878\n",
      "Epoch 143/200\n",
      " - 0s - loss: 0.0988 - mean_absolute_error: 0.4106 - accuracy: 0.6831\n",
      "Epoch 144/200\n",
      " - 0s - loss: 0.0969 - mean_absolute_error: 0.4065 - accuracy: 0.6843\n",
      "Epoch 145/200\n",
      " - 0s - loss: 0.0994 - mean_absolute_error: 0.4139 - accuracy: 0.6831\n",
      "Epoch 146/200\n",
      " - 0s - loss: 0.0979 - mean_absolute_error: 0.4107 - accuracy: 0.6831\n",
      "Epoch 147/200\n",
      " - 0s - loss: 0.0985 - mean_absolute_error: 0.4104 - accuracy: 0.6826\n",
      "Epoch 148/200\n",
      " - 0s - loss: 0.0986 - mean_absolute_error: 0.4095 - accuracy: 0.6831\n",
      "Epoch 149/200\n",
      " - 0s - loss: 0.0990 - mean_absolute_error: 0.4159 - accuracy: 0.6727\n",
      "Epoch 150/200\n",
      " - 0s - loss: 0.0983 - mean_absolute_error: 0.4105 - accuracy: 0.6762\n",
      "Epoch 151/200\n",
      " - 0s - loss: 0.0991 - mean_absolute_error: 0.4116 - accuracy: 0.6762\n",
      "Epoch 152/200\n",
      " - 0s - loss: 0.0991 - mean_absolute_error: 0.4118 - accuracy: 0.6826\n",
      "Epoch 153/200\n",
      " - 0s - loss: 0.0967 - mean_absolute_error: 0.4059 - accuracy: 0.6849\n",
      "Epoch 154/200\n",
      " - 0s - loss: 0.0965 - mean_absolute_error: 0.4052 - accuracy: 0.6959\n",
      "Epoch 155/200\n",
      " - 0s - loss: 0.0969 - mean_absolute_error: 0.4074 - accuracy: 0.6814\n",
      "Epoch 156/200\n",
      " - 0s - loss: 0.0990 - mean_absolute_error: 0.4118 - accuracy: 0.6773\n",
      "Epoch 157/200\n",
      " - 0s - loss: 0.0991 - mean_absolute_error: 0.4101 - accuracy: 0.6797\n",
      "Epoch 158/200\n",
      " - 0s - loss: 0.0979 - mean_absolute_error: 0.4080 - accuracy: 0.6837\n",
      "Epoch 159/200\n",
      " - 0s - loss: 0.0979 - mean_absolute_error: 0.4093 - accuracy: 0.6785\n",
      "Epoch 160/200\n",
      " - 0s - loss: 0.0983 - mean_absolute_error: 0.4097 - accuracy: 0.6756\n",
      "Epoch 161/200\n",
      " - 0s - loss: 0.0989 - mean_absolute_error: 0.4082 - accuracy: 0.6814\n",
      "Epoch 162/200\n",
      " - 0s - loss: 0.0990 - mean_absolute_error: 0.4102 - accuracy: 0.6802\n",
      "Epoch 163/200\n",
      " - 0s - loss: 0.0980 - mean_absolute_error: 0.4083 - accuracy: 0.6826\n",
      "Epoch 164/200\n",
      " - 0s - loss: 0.0993 - mean_absolute_error: 0.4090 - accuracy: 0.6843\n",
      "Epoch 165/200\n",
      " - 0s - loss: 0.0976 - mean_absolute_error: 0.4081 - accuracy: 0.6843\n",
      "Epoch 166/200\n",
      " - 0s - loss: 0.0980 - mean_absolute_error: 0.4086 - accuracy: 0.6837\n",
      "Epoch 167/200\n",
      " - 0s - loss: 0.0998 - mean_absolute_error: 0.4123 - accuracy: 0.6703\n",
      "Epoch 168/200\n",
      " - 0s - loss: 0.0980 - mean_absolute_error: 0.4092 - accuracy: 0.6843\n",
      "Epoch 169/200\n",
      " - 0s - loss: 0.0972 - mean_absolute_error: 0.4080 - accuracy: 0.6884\n",
      "Epoch 170/200\n",
      " - 0s - loss: 0.0996 - mean_absolute_error: 0.4106 - accuracy: 0.6791\n",
      "Epoch 171/200\n",
      " - 0s - loss: 0.0988 - mean_absolute_error: 0.4116 - accuracy: 0.6814\n",
      "Epoch 172/200\n",
      " - 0s - loss: 0.0978 - mean_absolute_error: 0.4054 - accuracy: 0.6791\n",
      "Epoch 173/200\n",
      " - 0s - loss: 0.0986 - mean_absolute_error: 0.4082 - accuracy: 0.6890\n",
      "Epoch 174/200\n",
      " - 0s - loss: 0.0983 - mean_absolute_error: 0.4080 - accuracy: 0.6797\n",
      "Epoch 175/200\n",
      " - 0s - loss: 0.0975 - mean_absolute_error: 0.4057 - accuracy: 0.6855\n",
      "Epoch 176/200\n",
      " - 0s - loss: 0.0992 - mean_absolute_error: 0.4093 - accuracy: 0.6762\n",
      "Epoch 177/200\n",
      " - 0s - loss: 0.0993 - mean_absolute_error: 0.4085 - accuracy: 0.6645\n",
      "Epoch 178/200\n",
      " - 0s - loss: 0.0981 - mean_absolute_error: 0.4085 - accuracy: 0.6843\n",
      "Epoch 179/200\n",
      " - 0s - loss: 0.0992 - mean_absolute_error: 0.4112 - accuracy: 0.6808\n",
      "Epoch 180/200\n",
      " - 0s - loss: 0.0981 - mean_absolute_error: 0.4062 - accuracy: 0.6878\n",
      "Epoch 181/200\n",
      " - 0s - loss: 0.0987 - mean_absolute_error: 0.4078 - accuracy: 0.6785\n",
      "Epoch 182/200\n",
      " - 0s - loss: 0.0964 - mean_absolute_error: 0.4047 - accuracy: 0.6808\n",
      "Epoch 183/200\n",
      " - 0s - loss: 0.0979 - mean_absolute_error: 0.4069 - accuracy: 0.6820\n",
      "Epoch 184/200\n",
      " - 0s - loss: 0.0980 - mean_absolute_error: 0.4079 - accuracy: 0.6878\n",
      "Epoch 185/200\n",
      " - 0s - loss: 0.0983 - mean_absolute_error: 0.4058 - accuracy: 0.6773\n",
      "Epoch 186/200\n",
      " - 0s - loss: 0.0975 - mean_absolute_error: 0.4046 - accuracy: 0.6808\n",
      "Epoch 187/200\n",
      " - 0s - loss: 0.0972 - mean_absolute_error: 0.4056 - accuracy: 0.6872\n",
      "Epoch 188/200\n",
      " - 0s - loss: 0.0982 - mean_absolute_error: 0.4068 - accuracy: 0.6872\n",
      "Epoch 189/200\n",
      " - 0s - loss: 0.1002 - mean_absolute_error: 0.4111 - accuracy: 0.6657\n",
      "Epoch 190/200\n",
      " - 0s - loss: 0.0971 - mean_absolute_error: 0.4036 - accuracy: 0.6878\n",
      "Epoch 191/200\n",
      " - 0s - loss: 0.0981 - mean_absolute_error: 0.4076 - accuracy: 0.6843\n",
      "Epoch 192/200\n",
      " - 0s - loss: 0.0978 - mean_absolute_error: 0.4079 - accuracy: 0.6872\n",
      "Epoch 193/200\n",
      " - 0s - loss: 0.0989 - mean_absolute_error: 0.4097 - accuracy: 0.6762\n",
      "Epoch 194/200\n",
      " - 0s - loss: 0.0984 - mean_absolute_error: 0.4066 - accuracy: 0.6814\n",
      "Epoch 195/200\n",
      " - 0s - loss: 0.0972 - mean_absolute_error: 0.4038 - accuracy: 0.6814\n",
      "Epoch 196/200\n",
      " - 0s - loss: 0.0976 - mean_absolute_error: 0.4064 - accuracy: 0.6826\n",
      "Epoch 197/200\n",
      " - 0s - loss: 0.0970 - mean_absolute_error: 0.4049 - accuracy: 0.6878\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 198/200\n",
      " - 0s - loss: 0.0982 - mean_absolute_error: 0.4091 - accuracy: 0.6808\n",
      "Epoch 199/200\n",
      " - 0s - loss: 0.0982 - mean_absolute_error: 0.4086 - accuracy: 0.6791\n",
      "Epoch 200/200\n",
      " - 0s - loss: 0.0982 - mean_absolute_error: 0.4084 - accuracy: 0.6849\n"
     ]
    }
   ],
   "source": [
    "Adagrad=optimizers.Adagrad(learning_rate=0.01)\n",
    "model4.compile(optimizer='Adagrad', loss='logcosh',metrics=[metrics.mae, 'accuracy'])\n",
    "history=model4.fit(X1, y1, batch_size=32, epochs=200, verbose=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "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"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['accuracy'])\n",
    "plt.title('Model accuracy')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Valid'], loc='upper left')\n",
    "plt.show()\n",
    "\n",
    "plt.plot(history.history['loss'])\n",
    "plt.title('Model loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Valid'], loc='upper left')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy on Test Data :- \n",
      "\t loss: 0.10043254404356985   mean_absolute_error: 0.43014082312583923   accuracy: 0.6969696879386902\n"
     ]
    }
   ],
   "source": [
    "loss, mean_absolute_error, accuracy =model4.evaluate(X2, y2, batch_size=32, verbose=2)\n",
    "print('Accuracy on Test Data :- ')\n",
    "print('\\t loss:',loss , '  mean_absolute_error:',mean_absolute_error, '  accuracy:',accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__After changing the kernel initialization, accuracy on validation set got increase.__"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Conclusion:\n",
    "After performing analysis on the CNN model, it is observed that, <br>\n",
    ">1. For the baseline model, Loss is Higher and Accuracy on validation data is 63%.<br> \n",
    ">2. The cost function logcosh is appropriate for this model than hinge.<br> \n",
    ">3. Higher the number of epochs, better the model is.<br>\n",
    ">4. Change in Optimizer can increase the accuracy and it helps to reduce the loss. Adagrad is better suit for this type of model than RMSprop<br>\n",
    ">5. Network architecture is as important as hyperparameters. By adding more layers and with proper use of hyperparamters, we can achieve higher accuracy.<br>\n",
    ">6. After changing the kernel initialization, accuracy on validation set got increase.\n",
    "\n",
    "For Multivariate Tabular data, Convolution Neural Network can produce good results but handling of data For the CNN model is difficult. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Author: \n",
    "Pranav Khiste<br>\n",
    "Information Systems <br>\n",
    "Northeastern University <br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Citation:\n",
    "References: <br>\n",
    "https://keras.io/models<br>\n",
    "https://machinelearningmastery.com/ <br>\n",
    "https://karpathy.github.io/2019/04/25/recipe<br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Licensing"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Copyright 2020 Pranav Sanjay Khiste\n",
    "\n",
    "Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the \"Software\"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:\n",
    "\n",
    "The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.\n",
    "\n",
    "THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}