covid_prediction_model_training.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "# Fetching Dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "_______________________________________________________________________________________________________________________________"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas\n",
    "import numpy\n",
    "import seaborn as sns\n",
    "from sklearn.model_selection import train_test_split\n",
    "from keras.models  import Sequential\n",
    "from keras.layers import Dense\n",
    "from keras.optimizers import Adam"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\gg415\\anaconda3\\lib\\site-packages\\IPython\\core\\interactiveshell.py:3063: DtypeWarning: Columns (7) have mixed types.Specify dtype option on import or set low_memory=False.\n",
      "  interactivity=interactivity, compiler=compiler, result=result)\n"
     ]
    }
   ],
   "source": [
    "dataset = pandas.read_csv('corona_tested_individuals_ver_0083.english.csv/corona_tested_individuals_ver_0083.english.csv')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "pandas.set_option('display.max_columns', None)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>test_date</th>\n",
       "      <th>cough</th>\n",
       "      <th>fever</th>\n",
       "      <th>sore_throat</th>\n",
       "      <th>shortness_of_breath</th>\n",
       "      <th>head_ache</th>\n",
       "      <th>corona_result</th>\n",
       "      <th>age_60_and_above</th>\n",
       "      <th>gender</th>\n",
       "      <th>test_indication</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2020-11-12</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>negative</td>\n",
       "      <td>No</td>\n",
       "      <td>male</td>\n",
       "      <td>Other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2020-11-12</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>negative</td>\n",
       "      <td>No</td>\n",
       "      <td>male</td>\n",
       "      <td>Other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2020-11-12</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>negative</td>\n",
       "      <td>Yes</td>\n",
       "      <td>female</td>\n",
       "      <td>Other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2020-11-12</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>negative</td>\n",
       "      <td>No</td>\n",
       "      <td>male</td>\n",
       "      <td>Other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2020-11-12</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>negative</td>\n",
       "      <td>No</td>\n",
       "      <td>male</td>\n",
       "      <td>Contact with confirmed</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2742591</th>\n",
       "      <td>2020-03-11</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>negative</td>\n",
       "      <td>NaN</td>\n",
       "      <td>female</td>\n",
       "      <td>Other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2742592</th>\n",
       "      <td>2020-03-11</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>negative</td>\n",
       "      <td>NaN</td>\n",
       "      <td>female</td>\n",
       "      <td>Other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2742593</th>\n",
       "      <td>2020-03-11</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>other</td>\n",
       "      <td>NaN</td>\n",
       "      <td>male</td>\n",
       "      <td>Other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2742594</th>\n",
       "      <td>2020-03-11</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>negative</td>\n",
       "      <td>NaN</td>\n",
       "      <td>female</td>\n",
       "      <td>Other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2742595</th>\n",
       "      <td>2020-03-11</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>negative</td>\n",
       "      <td>NaN</td>\n",
       "      <td>male</td>\n",
       "      <td>Other</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2742596 rows × 10 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          test_date  cough  fever  sore_throat  shortness_of_breath  \\\n",
       "0        2020-11-12      0      0            0                    0   \n",
       "1        2020-11-12      0      1            0                    0   \n",
       "2        2020-11-12      0      0            0                    0   \n",
       "3        2020-11-12      0      0            0                    0   \n",
       "4        2020-11-12      0      1            0                    0   \n",
       "...             ...    ...    ...          ...                  ...   \n",
       "2742591  2020-03-11      0      0            0                    0   \n",
       "2742592  2020-03-11      0      0            0                    0   \n",
       "2742593  2020-03-11      0      0            0                    0   \n",
       "2742594  2020-03-11      0      0            0                    0   \n",
       "2742595  2020-03-11      0      1            0                    0   \n",
       "\n",
       "         head_ache corona_result age_60_and_above  gender  \\\n",
       "0                0      negative               No    male   \n",
       "1                0      negative               No    male   \n",
       "2                0      negative              Yes  female   \n",
       "3                0      negative               No    male   \n",
       "4                0      negative               No    male   \n",
       "...            ...           ...              ...     ...   \n",
       "2742591          0      negative              NaN  female   \n",
       "2742592          0      negative              NaN  female   \n",
       "2742593          0         other              NaN    male   \n",
       "2742594          0      negative              NaN  female   \n",
       "2742595          0      negative              NaN    male   \n",
       "\n",
       "                test_indication  \n",
       "0                         Other  \n",
       "1                         Other  \n",
       "2                         Other  \n",
       "3                         Other  \n",
       "4        Contact with confirmed  \n",
       "...                         ...  \n",
       "2742591                   Other  \n",
       "2742592                   Other  \n",
       "2742593                   Other  \n",
       "2742594                   Other  \n",
       "2742595                   Other  \n",
       "\n",
       "[2742596 rows x 10 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['test_date', 'cough', 'fever', 'sore_throat', 'shortness_of_breath',\n",
       "       'head_ache', 'corona_result', 'age_60_and_above', 'gender',\n",
       "       'test_indication'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# EDA"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "________________________________________________________________________________"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 2742596 entries, 0 to 2742595\n",
      "Data columns (total 10 columns):\n",
      " #   Column               Dtype \n",
      "---  ------               ----- \n",
      " 0   test_date            object\n",
      " 1   cough                int64 \n",
      " 2   fever                int64 \n",
      " 3   sore_throat          int64 \n",
      " 4   shortness_of_breath  int64 \n",
      " 5   head_ache            int64 \n",
      " 6   corona_result        object\n",
      " 7   age_60_and_above     object\n",
      " 8   gender               object\n",
      " 9   test_indication      object\n",
      "dtypes: int64(5), object(5)\n",
      "memory usage: 209.2+ MB\n"
     ]
    }
   ],
   "source": [
    "dataset.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "negative    2480403\n",
       "positive     220975\n",
       "other         41218\n",
       "Name: corona_result, dtype: int64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset[\"corona_result\"].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset = dataset[dataset.corona_result != 'other']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "gender = pandas.get_dummies(dataset['gender'])\n",
    "female = gender['female']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "age_60_and_above = pandas.get_dummies(dataset['age_60_and_above'], prefix='age_60_and_above')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "age_60_and_above_Yes = age_60_and_above['age_60_and_above_Yes']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = dataset[['cough', 'fever', 'sore_throat', 'shortness_of_breath','head_ache']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = pandas.concat([X,age_60_and_above_Yes,female],axis =1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "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>cough</th>\n",
       "      <th>fever</th>\n",
       "      <th>sore_throat</th>\n",
       "      <th>shortness_of_breath</th>\n",
       "      <th>head_ache</th>\n",
       "      <th>age_60_and_above_Yes</th>\n",
       "      <th>female</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2742590</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2742591</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2742592</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2742594</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2742595</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2701378 rows × 7 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         cough  fever  sore_throat  shortness_of_breath  head_ache  \\\n",
       "0            0      0            0                    0          0   \n",
       "1            0      1            0                    0          0   \n",
       "2            0      0            0                    0          0   \n",
       "3            0      0            0                    0          0   \n",
       "4            0      1            0                    0          0   \n",
       "...        ...    ...          ...                  ...        ...   \n",
       "2742590      0      0            0                    0          0   \n",
       "2742591      0      0            0                    0          0   \n",
       "2742592      0      0            0                    0          0   \n",
       "2742594      0      0            0                    0          0   \n",
       "2742595      0      1            0                    0          0   \n",
       "\n",
       "         age_60_and_above_Yes  female  \n",
       "0                           0       0  \n",
       "1                           0       0  \n",
       "2                           1       1  \n",
       "3                           0       0  \n",
       "4                           0       0  \n",
       "...                       ...     ...  \n",
       "2742590                     0       0  \n",
       "2742591                     0       1  \n",
       "2742592                     0       1  \n",
       "2742594                     0       1  \n",
       "2742595                     0       0  \n",
       "\n",
       "[2701378 rows x 7 columns]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "corona_result = pandas.get_dummies(dataset['corona_result'], prefix='corona_result')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "y = corona_result['corona_result_positive']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x1296 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize = (18,18))\n",
    "sns.heatmap(X.corr(), annot = True, cmap = \"RdYlGn\")\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Feature Selection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "ExtraTreesRegressor(bootstrap=False, ccp_alpha=0.0, criterion='mse',\n",
       "                    max_depth=None, max_features='auto', max_leaf_nodes=None,\n",
       "                    max_samples=None, min_impurity_decrease=0.0,\n",
       "                    min_impurity_split=None, min_samples_leaf=1,\n",
       "                    min_samples_split=2, min_weight_fraction_leaf=0.0,\n",
       "                    n_estimators=100, n_jobs=None, oob_score=False,\n",
       "                    random_state=None, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.ensemble import ExtraTreesRegressor\n",
    "selection = ExtraTreesRegressor()\n",
    "selection.fit(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.05565895 0.270741   0.09968867 0.02549797 0.53404372 0.0123989\n",
      " 0.0019708 ]\n"
     ]
    }
   ],
   "source": [
    "print(selection.feature_importances_)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#plot graph of feature importances for better visualization\n",
    "\n",
    "plt.figure(figsize = (12,8))\n",
    "feat_importances = pandas.Series(selection.feature_importances_, index=X.columns)\n",
    "feat_importances.nlargest(20).plot(kind='barh')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Model Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.40, random_state=92)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "learning_rate_value = 0.00005\n",
    "        \n",
    "model2 = Sequential()\n",
    "model2.add( Dense(units=50,  \n",
    "                 kernel_initializer=\"zeros\", \n",
    "                 bias_initializer=\"zeros\",\n",
    "                 input_dim=X_train.shape[1],\n",
    "                 activation='relu'\n",
    "                 \n",
    "                )  )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "model2.add(Dense(units=36,activation='relu'))\n",
    "model2.add(Dense(units=12,activation='relu'))\n",
    "model2.add(Dense(units=12,activation='relu'))\n",
    "model2.add(Dense(units=12,activation='relu'))\n",
    "model2.add(Dense( units=1, \n",
    "                 activation='sigmoid',\n",
    "                 kernel_initializer=\"zeros\", \n",
    "                 bias_initializer=\"zeros\"\n",
    "                ))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "model2.compile(optimizer=Adam(learning_rate=learning_rate_value),loss=\"binary_crossentropy\",metrics =['accuracy'])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/20\n",
      "50651/50651 [==============================] - 158s 3ms/step - loss: 0.5154 - accuracy: 0.9180 - val_loss: 0.2864 - val_accuracy: 0.9183\n",
      "Epoch 2/20\n",
      "50651/50651 [==============================] - 144s 3ms/step - loss: 0.2843 - accuracy: 0.9182 - val_loss: 0.2830 - val_accuracy: 0.9183\n",
      "Epoch 3/20\n",
      "50651/50651 [==============================] - 137s 3ms/step - loss: 0.2823 - accuracy: 0.9185 - val_loss: 0.2830 - val_accuracy: 0.9183\n",
      "Epoch 4/20\n",
      "50651/50651 [==============================] - 144s 3ms/step - loss: 0.2841 - accuracy: 0.9178 - val_loss: 0.2830 - val_accuracy: 0.9183\n",
      "Epoch 5/20\n",
      "50651/50651 [==============================] - 145s 3ms/step - loss: 0.2834 - accuracy: 0.9181 - val_loss: 0.2830 - val_accuracy: 0.9183\n",
      "Epoch 6/20\n",
      "50651/50651 [==============================] - 146s 3ms/step - loss: 0.2843 - accuracy: 0.9177 - val_loss: 0.2830 - val_accuracy: 0.9183\n",
      "Epoch 7/20\n",
      "50651/50651 [==============================] - 140s 3ms/step - loss: 0.2823 - accuracy: 0.9185 - val_loss: 0.2830 - val_accuracy: 0.9183\n",
      "Epoch 8/20\n",
      "50651/50651 [==============================] - 177s 4ms/step - loss: 0.2830 - accuracy: 0.9183 - val_loss: 0.2830 - val_accuracy: 0.9183\n",
      "Epoch 9/20\n",
      "50651/50651 [==============================] - 159s 3ms/step - loss: 0.2830 - accuracy: 0.9183 - val_loss: 0.2830 - val_accuracy: 0.9183\n",
      "Epoch 10/20\n",
      "50651/50651 [==============================] - 154s 3ms/step - loss: 0.2836 - accuracy: 0.9180 - val_loss: 0.2830 - val_accuracy: 0.9183\n",
      "Epoch 11/20\n",
      "50651/50651 [==============================] - 159s 3ms/step - loss: 0.2833 - accuracy: 0.9181 - val_loss: 0.2830 - val_accuracy: 0.9183\n",
      "Epoch 12/20\n",
      "50651/50651 [==============================] - 179s 4ms/step - loss: 0.2842 - accuracy: 0.9178 - val_loss: 0.2830 - val_accuracy: 0.9183\n",
      "Epoch 13/20\n",
      "50651/50651 [==============================] - 180s 4ms/step - loss: 0.2834 - accuracy: 0.9181 - val_loss: 0.2830 - val_accuracy: 0.9183\n",
      "Epoch 14/20\n",
      "50651/50651 [==============================] - 181s 4ms/step - loss: 0.2823 - accuracy: 0.9186 - val_loss: 0.2830 - val_accuracy: 0.9183\n",
      "Epoch 15/20\n",
      "50651/50651 [==============================] - 174s 3ms/step - loss: 0.2835 - accuracy: 0.9180 - val_loss: 0.2830 - val_accuracy: 0.9183\n",
      "Epoch 16/20\n",
      "50651/50651 [==============================] - 175s 3ms/step - loss: 0.2826 - accuracy: 0.9184 - val_loss: 0.2830 - val_accuracy: 0.9183\n",
      "Epoch 17/20\n",
      "50651/50651 [==============================] - 150s 3ms/step - loss: 0.2823 - accuracy: 0.9185 - val_loss: 0.2830 - val_accuracy: 0.9183\n",
      "Epoch 18/20\n",
      "50651/50651 [==============================] - 77s 2ms/step - loss: 0.2837 - accuracy: 0.9180 - val_loss: 0.2830 - val_accuracy: 0.9183\n",
      "Epoch 19/20\n",
      "50651/50651 [==============================] - 68s 1ms/step - loss: 0.2836 - accuracy: 0.9180 - val_loss: 0.2830 - val_accuracy: 0.9183\n",
      "Epoch 20/20\n",
      "50651/50651 [==============================] - 72s 1ms/step - loss: 0.2830 - accuracy: 0.9182 - val_loss: 0.2830 - val_accuracy: 0.9183\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x20ec3446cc8>"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#model2.fit(X_train,y_train,\n",
    "#           epochs=20, verbose=1,\n",
    "#           validation_data=(X_test,y_test),\n",
    "#          validation_steps=100,\n",
    "#          shuffle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "scores = model2.evaluate(X_test, y_test, verbose=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.28302785754203796, 0.9182528853416443]"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Saving Model\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "#import joblib\n",
    "#joblib.dump(model2, \"coweb.pk1\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "#import pickle\n",
    "# open a file, where you ant to store the data\n",
    "#file = open('coweb.pk1', 'wb')\n",
    "\n",
    "# dump information to that file\n",
    "#pickle.dump(model2, file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "model2.save('coweb.h5')\n",
    "#tf.keras.models.save_model(model2, file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "loaded_model = tf.keras.models.load_model('coweb.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "scores = loaded_model.evaluate(X_test, y_test, verbose=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.28302785754203796, 0.9182528853416443]"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}