SimpleRNN_Experiment3_Padding_and_Masking.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "python version:  3.7.3 (default, Mar 27 2019, 22:11:17) \n",
      "[GCC 7.3.0]\n",
      "pandas version:  0.24.2\n",
      "numpy version:  1.16.4\n",
      "scipy version:  1.3.0\n",
      "sklearn version:  0.21.2\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "keras version:  2.2.4\n",
      "keras backend:  tensorflow\n",
      "tensorflow version:  1.14.0\n"
     ]
    }
   ],
   "source": [
    "import sys\n",
    "print(\"python version: \", sys.version)\n",
    "\n",
    "import pandas as pd\n",
    "print(\"pandas version: \", pd.__version__)\n",
    "\n",
    "import numpy as np\n",
    "print('numpy version: ', np.__version__)\n",
    "\n",
    "import scipy as sp\n",
    "print('scipy version: ', sp.__version__)\n",
    "\n",
    "import sklearn as skl\n",
    "print('sklearn version: ', skl.__version__)\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "import os\n",
    "os.environ[\"CUDA_DEVICE_ORDER\"] = \"PCI_BUS_ID\"\n",
    "os.environ[\"CUDA_VISIBLE_DEVICES\"] = \"\"\n",
    "\n",
    "import keras as ks\n",
    "print('keras version: ', ks.__version__)\n",
    "print('keras backend: ', ks.backend.backend())\n",
    "\n",
    "import tensorflow as tf\n",
    "print('tensorflow version: ', tf.__version__)\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "\n",
    "from keras import backend as K\n",
    "from sklearn.metrics import confusion_matrix, precision_score, recall_score\n",
    "from sklearn.metrics import roc_curve, auc, f1_score, accuracy_score, precision_recall_curve\n",
    "\n",
    "import random"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(7584, 208)\n"
     ]
    }
   ],
   "source": [
    "excel_file = r'geriomop-data-from-egems-paper-UPDATED-COL-FIX-ATC_2-and-behavior-change-022819.tsv'\n",
    "df = pd.DataFrame(pd.read_csv(excel_file, sep='\\t'))\n",
    "print(df.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Drop features\n",
    "f_drop = [\"Blank-1\", \"ws-sudden-stop\", \"ws-sudden-stop-drugs\", \"ws-prolonged-psychotropics\", \"mds-toilet-prgm-atmptd\"] + \\\n",
    "         ['Deleted-%d' % i for i in range(1, 13)] + ['Psychotropic-%d' % i for i in range(1, 6)] + \\\n",
    "         ['Psychotropic-%d-average-daily-dose' % i for i in range(1, 6)] + [\"Entry-discharge-type\"]\n",
    "\n",
    "f_drop += [\"acute_mental_change\", \"aphasia\", \"balance_toileting\", \"balance_turning_around\", \"bipolar\", \"bph\", \\\n",
    "           \"cam_motor_retardation\", \"cerebralpalsy\", \"cirrhosis\", \"cms-long-stay\", \"comatose\", \"conduct_pain_assmnt\",\\\n",
    "           \"dysrhythmias\", \"embolisms\", \"facility\", \"hemiplegia/hemiparesis\", \"hepatitis\", \"huntingtons\", \"hyperkalemia\",\\\n",
    "           \"hypertension\", \"hyperthyroidism\", \"hyponatremia\", \"hypotension\", \"hypothyroidism\", \\\n",
    "           \"limb_prosthesis_past_7_days\", \"marital-status\", \"mds-conduct-staff-assessment-mental-status\", \\\n",
    "           \"mds-dehydrated\", \"mds-delirium-scale\", \"mds-excess-weight-loss\", \"mds-long-term-memory-ok\", \\\n",
    "           \"mds-pressure-ulcer-stage-1\", \"mds-pressure-ulcer-stage-2\", \"mds-pressure-ulcer-stage-3\", \\\n",
    "           \"mds-pressure-ulcer-stage-4\", \"mds-short-term-memory-ok\", \"mds-staff-assess-pain\", \"multiple-sclerosis\", \\\n",
    "           \"neurogenic-bladder\", \"obstructive-uropathy\", \"paraplegia\", \"pneumonia\", \"ptsd\", \"pvd\", \"quadriplegia\", \\\n",
    "           \"renal-failure\", \"schizophrenia\", \"septicemia\", \"stroke\", \"tourettes\", \"traumatic-brain-injury\", \\\n",
    "           \"tuberculosis\", \"wound\", \"ws-antibiotic-anticoag-coexposure\", \"ws-diuretic-adl\", \\\n",
    "           \"ws-meclizine-psych-coexposure\", \"ws-meclizine-psych-coexposure\", \"ws-meclizine-psych-coexposure\",\\\n",
    "           \"ws-meclizine-psych-coexposure\", \"ws-meclizine-psych-coexposure\", \\\n",
    "           \"ws-psychotropic-with-not-ordered-weight-loss\", \"ws-tramadol-antidepressant-coexposure\"]\n",
    " \n",
    "# Missing value means not relevant\n",
    "f_one_hot = [\"Mds-fall-2-6-months-to-admission\", \"Mds-fall-last-month-to-admission\", \"mds-pain-last-five-days\", \\\n",
    "             \"recent-start-other-fall-risk-rx\"]\n",
    "cate_one_hot = [['Yes', 'No', 'Unable to answer']] * 3 + [['Yes', 'No']]\n",
    "\n",
    "f_period = [(\"Psychotropic-%d-Start-date\" % i, 'Psychotropic-%d-End-date' % i, 'Psychotropic-%d-Period' % i) for i in range(1, 6)]\n",
    "\n",
    "f_label = ['mds-cognitive-scale', 'mds-pain-freq-last-five-days', 'mds-pain-intensity']\n",
    "dic_label = [{'Independent': 0, 'Modified Independence': 1, 'Moderately Impaired': 2, 'Severely Impaired': 3}, \\\n",
    "             {'Continuous': 4, 'Frequent': 3, 'Occasional': 2, 'Rare': 1, 'Unable to respond': 0}, \\\n",
    "             {'None': 0, 'Mild': 1, 'Moderate': 2, 'Severe': 3, 'Very severe, horrible': 4}]\n",
    "\n",
    "f_transform = [('mds-cognitive-scale', 3, 0), ('mds-pain-freq-last-five-days', 4, 0), ('mds-pain-intensity', 4, 0)] + \\\n",
    "              [('Psychotropic-%d-Period' % i, 0, 0) for i in range(1, 6)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Encode features (originally) with no missing values\n",
    "f_drop += [\"predictor-date\", \"projected-date\", \"PPS-assessment\", \"Federal-assessment\"]\n",
    "\n",
    "f_period += [('episode-start-date', 'episode-end-date', 'episode-period'),\\\n",
    "             ('stay-start-date', 'stay-end-date', 'stay-period')]\n",
    "\n",
    "f_label += [\"mds-stay-trans-short-to-long\", \"mds-antianxiety-medication\", \\\n",
    "            \"mds-antidepressant-medication\", \"mds-antipsychotic-medication\", \"mds-antibiotic-medication\",\n",
    "            \"mds-anticoagulant-medication\", \"mds-diuretic-medication\", \\\n",
    "            \"mds-hypnotic-medication\", \"mds-behavioral-symptoms\", \"mds-behavioral-symptoms-to-others\", \\\n",
    "            \"mds-depression\", \"mds-fever\", \"mds-impaired-mobility\", \"mds-impaired-transfer\", \\\n",
    "            \"mds-internal-bleeding\", \"mds-malnutrition\", \"mds-no-problem-conditions\", \"mds-vomiting\", \\\n",
    "            \"mds-impaired-walk-in-room\", \"mds-impaired-walk-in-corridor\", \"mds-impaired-locomot-unit\", \\\n",
    "            \"mds-impaired-locomot-other\"] + [\"mds-pain-non-verbal\"]\n",
    "dic_label += [{'Yes': 1, 'No': 0}] * 22 + [{'None/Mild':0,'Moderate/Severe':1}]\n",
    "\n",
    "f_one_hot += [\"race\", \"gender\", \"mds-bims-summary-ranking\"]\n",
    "cate_one_hot += [['White', 'Black', 'Other'], \\\n",
    "                 ['Female', 'Male'], ['Intact or Moderately Intact', 'Moderate Impairment']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Features need imputation\n",
    "f_one_hot += [\"uti\", \"mdro\", \"anemia\", \"constipation\", \"alzheimers\", \"anxiety\", \"depression\", \"non-alz-dimentia\", \\\n",
    "              \"parkinsons\", \"psychosis\", \"seizure\", \"emphysema\", \"arthritis\", \"ashd\", \"cancer\", \"diabetes\", \"gerd\",\\\n",
    "              \"heart-failure\", \"osteoporosis\", \"thyroid-disorder\", \"transient-ischemic-attack\", \"mds-pain-medication\", \\\n",
    "              \"mds-received-pain-tx-non-pharm\", \"mds-received-prn-pain-medication\", \"mds-pressure-ulcer-prsnt\", \\\n",
    "              \"conduct_bims\", \"cane_or_crutch_past_7_days\", \"walker_past_7_days\", \"wheelchair_past_7_days\", \\\n",
    "              \"no_listed_mobility_device\", \"hyperlipidemia\"] + \\\n",
    "              [\"balance_while_standing\", \"balance_while_walking\", \"balance_bed_to_chair\"] + [\"mds-entered-from\"]\n",
    "cate_one_hot += [['Yes', 'No']] * 31 + \\\n",
    "                [[\"Steady at all times\", \"able to stabilize without assistance\", \\\n",
    "                  \"able to stabilize with assistance\", \"Activity did not occur\"]] * 3 + \\\n",
    "                [[\"Acute Hospital\", \"Psychiatric Hospital\", \"Another nursing home or swing bed\", \"Hospice\", \\\n",
    "                  \"Inpatient rehabilitation facility\",\n",
    "                  \"Community (private home/apt., board/care, assisted living, group home)\", \"Other\"]]\n",
    "\n",
    "f_label += [\"mds-urinary-incontinence\", \"functlimit_rom_upper\", \"functlimit_rom_lower\", \"does_resident_wander\"] + \\\n",
    "           [\"cam_inattention\", \"cam_disorganized_thought\", \"cam_altered_conc\"]\n",
    "dic_label += [{\"Always continent\": 0, \"Occasionally incontinent\": 1, \"Frequently incontinent\": 2,\\\n",
    "               \"Always incontinent\": 3, \"Catheter - not rated\": np.nan}] + \\\n",
    "             [{\"No impairment\": 0, \"Impairment on one side\": 1, \"Impairment on both sides\": 2}] * 2 + \\\n",
    "             [{\"Behavior not exhibited\": 0, \"1 to 3 days\": 1, \"4 to 6 days\": 2, \"occurred daily\": 3}] + \\\n",
    "             [{\"Behavior not present\": 0, \"Behavior present, fluctuates\": 1, \"Behavior continuously present\": 2}] * 3\n",
    "\n",
    "f_transform += [(\"mds-urinary-incontinence\", 3, 0), (\"cam_inattention\", 2, 0), (\"cam_disorganized_thought\", 2, 0), \\\n",
    "                (\"cam_altered_conc\", 2, 0), (\"functlimit_rom_upper\", 2, 0), \\\n",
    "                (\"functlimit_rom_lower\", 2, 0), (\"does_resident_wander\", 3, 0)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Set sdd=0 to NA\n",
    "for i in range(1, 6):\n",
    "    df.loc[df['psych_sdd_%d' % i] == 0, 'psych_sdd_%d' % i] = np.nan\n",
    "# These data were dropped in Experiment 1\n",
    "f_transform += [('psych_sdd_%d' % i, 0, 0) for i in range(1, 6)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Dropping min effective dose and one hot encoding ATC psychotropic drugs\n",
    "# These data were not used in Experiment 1 b/c of missing values\n",
    "f_drop += [\"min_effective_dose_%d\" % i for i in range(1, 6)]\n",
    "\n",
    "atc_list = []\n",
    "for i in range(1, 6):\n",
    "    atc_list.append(df.groupby('atc_%d' % i)['min_effective_dose_%d' % i].first())\n",
    "atc_codes = pd.concat(atc_list)\n",
    "atc_codes = atc_codes[~atc_codes.index.duplicated()].index.tolist()\n",
    "f_one_hot += [\"atc_%d\" % i for i in range(1, 6)]\n",
    "cate_one_hot += [atc_codes] * 5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# generate target\n",
    "target = df['outcome-Mds-fall-since-prior-assessment'].map({'None': 0, 'Yes': 1})\n",
    "dff = df.drop(['outcome-mds-fall-no-injury', 'outcome-mds-fall-minor-injury', 'outcome-mds-fall-major-injury', \\\n",
    "               'outcome-Mds-fall-since-prior-assessment', 'outcome-riskmaster-fall-incident'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# train test split\n",
    "from sklearn.model_selection import GroupShuffleSplit\n",
    "def train_test_split(df, target, groupby, radio, random_state):\n",
    "    train_idx, test_idx = next(GroupShuffleSplit(train_size=radio, test_size=1-radio, random_state=random_state).split(df, target, df[groupby]))\n",
    "    return df.iloc[train_idx], df.iloc[test_idx], target.iloc[train_idx], target.iloc[test_idx]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total number of patients: 3133\n",
      "Number of patients in training set: 2193\n"
     ]
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(dff, target, 'patient-id', 0.7, 0)\n",
    "print('Total number of patients: %d' % len(df.groupby('patient-id')))\n",
    "print('Number of patients in training set: %d' % len(X_train.groupby('patient-id')))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# drop feature\n",
    "def drop_features(df, f):\n",
    "    df.drop(f, axis=1, inplace=True)\n",
    "    \n",
    "# one hot encoding\n",
    "def one_hot_encoding(df, features, categories):\n",
    "    for f, c in zip(features, categories):\n",
    "        df[f] = df[f].astype('category', categories=c)\n",
    "    return pd.get_dummies(df, columns=features, prefix=features)\n",
    "\n",
    "# compute date difference\n",
    "def date_diff(start, end):\n",
    "    return (pd.to_datetime(end) - pd.to_datetime(start)) / np.timedelta64(1, 'D')\n",
    "\n",
    "# generate date diff features\n",
    "def generate_period(df, f):\n",
    "    for start, end, period in f:\n",
    "        df[period] = date_diff(df[start], df[end])\n",
    "        df.drop([start, end], axis=1, inplace=True)\n",
    "        \n",
    "# label encoding\n",
    "def label_encoding(df, features, dictionaries):\n",
    "    for f, dic in zip(features, dictionaries):\n",
    "        df[f] = df[f].map(dic)\n",
    "        \n",
    "# sin&cos transform\n",
    "def col_transform(df, col, mmax, mmin, df_ref=None):\n",
    "    if(mmax <= mmin):\n",
    "        # for test set, if max&min not set, use training set value\n",
    "        if(df_ref is not None): \n",
    "            mmax = df_ref[col].max()\n",
    "            mmin = df_ref[col].max()\n",
    "        else:\n",
    "            mmax = df[col].max()\n",
    "            mmin = df[col].min()\n",
    "    angle = 0.25 * np.pi * (df[col] - mmin) / (mmax - mmin) + 0.125 * np.pi\n",
    "    return (np.cos(angle), np.sin(angle))\n",
    "\n",
    "def df_transform(df, f, df_ref=None):\n",
    "    for col, mmax, mmin in f:\n",
    "        df[col + '_x'], df[col + '_y'] = col_transform(df, col, mmax, mmin, df_ref)\n",
    "        df[col + '_x'].fillna(0, inplace=True)\n",
    "        df[col + '_y'].fillna(0, inplace=True)\n",
    "        df.drop(col, axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 1min 18s, sys: 1.94 s, total: 1min 20s\n",
      "Wall time: 22.4 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "## Training set data reshaping\n",
    "generate_period(X_train, f_period)\n",
    "label_encoding(X_train, f_label, dic_label)\n",
    "X_train_original = X_train.copy() # store value range\n",
    "df_transform(X_train, f_transform)\n",
    "X_train = one_hot_encoding(X_train, f_one_hot, cate_one_hot)\n",
    "drop_features(X_train, f_drop)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_use = X_train.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Index([], dtype='object')\n"
     ]
    }
   ],
   "source": [
    "print(X_train.dtypes[X_train.dtypes == 'object'].index)"
   ]
  },
  {
   "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>count</th>\n",
       "      <th>percentage</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "Empty DataFrame\n",
       "Columns: [count, percentage]\n",
       "Index: []"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# missing data describe\n",
    "def missing_summary(df):\n",
    "    missing_cnt = df.isna().sum()\n",
    "    missing_data = pd.concat([missing_cnt, missing_cnt/df.shape[0]], axis=1, keys=['count', 'percentage'])\n",
    "    return missing_data[missing_data['percentage'] != 0].sort_values(by='percentage', ascending=False)\n",
    "\n",
    "missing_summary(X_train_use)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 19.4 s, sys: 388 ms, total: 19.8 s\n",
      "Wall time: 19.8 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "## Test set data reshaping\n",
    "generate_period(X_test, f_period)\n",
    "label_encoding(X_test, f_label, dic_label)\n",
    "df_transform(X_test, f_transform, X_train_original)\n",
    "X_test = one_hot_encoding(X_test, f_one_hot, cate_one_hot)\n",
    "drop_features(X_test, f_drop)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test_use = X_test.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Index([], dtype='object')\n"
     ]
    }
   ],
   "source": [
    "print(X_test.dtypes[X_test.dtypes == 'object'].index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "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>count</th>\n",
       "      <th>percentage</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "Empty DataFrame\n",
       "Columns: [count, percentage]\n",
       "Index: []"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "missing_summary(X_test_use)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(5301, 398)"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_use.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2283, 398)"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test_use.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "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>patient-id</th>\n",
       "      <th>current-stay-days</th>\n",
       "      <th>cumulative-days-in-facility</th>\n",
       "      <th>mds-stay-trans-short-to-long</th>\n",
       "      <th>age</th>\n",
       "      <th>mds-antianxiety-medication</th>\n",
       "      <th>mds-antidepressant-medication</th>\n",
       "      <th>mds-antipsychotic-medication</th>\n",
       "      <th>mds-antibiotic-medication</th>\n",
       "      <th>mds-anticoagulant-medication</th>\n",
       "      <th>...</th>\n",
       "      <th>atc_5_N06AB10</th>\n",
       "      <th>atc_5_N06AX05</th>\n",
       "      <th>atc_5_N06AX11</th>\n",
       "      <th>atc_5_N06AX12</th>\n",
       "      <th>atc_5_N06AX16</th>\n",
       "      <th>atc_5_N06AX21</th>\n",
       "      <th>atc_5_N06AX24</th>\n",
       "      <th>atc_5_N05AA01</th>\n",
       "      <th>atc_5_N05BA05</th>\n",
       "      <th>atc_5_N06AX23</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>10353</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>67</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</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",
       "      <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>10364</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>1</td>\n",
       "      <td>84</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</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",
       "      <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>10386</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>1</td>\n",
       "      <td>77</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</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",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5000</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>90</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</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",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>5003</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>1</td>\n",
       "      <td>90</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</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",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 398 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   patient-id  current-stay-days  cumulative-days-in-facility  \\\n",
       "0       10353                  7                            7   \n",
       "1       10364                  7                            7   \n",
       "4       10386                  7                            7   \n",
       "5        5000                  7                            7   \n",
       "6        5003                  7                            7   \n",
       "\n",
       "   mds-stay-trans-short-to-long  age  mds-antianxiety-medication  \\\n",
       "0                             0   67                           1   \n",
       "1                             1   84                           0   \n",
       "4                             1   77                           1   \n",
       "5                             0   90                           0   \n",
       "6                             1   90                           0   \n",
       "\n",
       "   mds-antidepressant-medication  mds-antipsychotic-medication  \\\n",
       "0                              1                             0   \n",
       "1                              0                             1   \n",
       "4                              1                             1   \n",
       "5                              1                             1   \n",
       "6                              1                             1   \n",
       "\n",
       "   mds-antibiotic-medication  mds-anticoagulant-medication  ...  \\\n",
       "0                          0                             0  ...   \n",
       "1                          1                             0  ...   \n",
       "4                          0                             0  ...   \n",
       "5                          0                             0  ...   \n",
       "6                          1                             0  ...   \n",
       "\n",
       "   atc_5_N06AB10  atc_5_N06AX05  atc_5_N06AX11  atc_5_N06AX12  atc_5_N06AX16  \\\n",
       "0              0              0              0              0              0   \n",
       "1              0              0              0              0              0   \n",
       "4              0              0              0              0              0   \n",
       "5              0              0              0              0              0   \n",
       "6              0              0              0              0              0   \n",
       "\n",
       "   atc_5_N06AX21  atc_5_N06AX24  atc_5_N05AA01  atc_5_N05BA05  atc_5_N06AX23  \n",
       "0              0              0              0              0              0  \n",
       "1              0              0              0              0              0  \n",
       "4              0              0              0              0              0  \n",
       "5              0              0              0              0              0  \n",
       "6              0              0              0              0              0  \n",
       "\n",
       "[5 rows x 398 columns]"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_use.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "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>patient-id</th>\n",
       "      <th>current-stay-days</th>\n",
       "      <th>cumulative-days-in-facility</th>\n",
       "      <th>mds-stay-trans-short-to-long</th>\n",
       "      <th>age</th>\n",
       "      <th>mds-antianxiety-medication</th>\n",
       "      <th>mds-antidepressant-medication</th>\n",
       "      <th>mds-antipsychotic-medication</th>\n",
       "      <th>mds-antibiotic-medication</th>\n",
       "      <th>mds-anticoagulant-medication</th>\n",
       "      <th>...</th>\n",
       "      <th>atc_5_N06AB10</th>\n",
       "      <th>atc_5_N06AX05</th>\n",
       "      <th>atc_5_N06AX11</th>\n",
       "      <th>atc_5_N06AX12</th>\n",
       "      <th>atc_5_N06AX16</th>\n",
       "      <th>atc_5_N06AX21</th>\n",
       "      <th>atc_5_N06AX24</th>\n",
       "      <th>atc_5_N05AA01</th>\n",
       "      <th>atc_5_N05BA05</th>\n",
       "      <th>atc_5_N06AX23</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>10379</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>84</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>...</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",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>10385</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>1</td>\n",
       "      <td>85</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</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",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>5032</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>85</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</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",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>5032</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>0</td>\n",
       "      <td>86</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</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",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>5032</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>86</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</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",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 398 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    patient-id  current-stay-days  cumulative-days-in-facility  \\\n",
       "2        10379                  7                            7   \n",
       "3        10385                  7                            7   \n",
       "14        5032                  7                            7   \n",
       "15        5032                  8                            8   \n",
       "16        5032                  7                            7   \n",
       "\n",
       "    mds-stay-trans-short-to-long  age  mds-antianxiety-medication  \\\n",
       "2                              0   84                           0   \n",
       "3                              1   85                           0   \n",
       "14                             0   85                           1   \n",
       "15                             0   86                           1   \n",
       "16                             0   86                           1   \n",
       "\n",
       "    mds-antidepressant-medication  mds-antipsychotic-medication  \\\n",
       "2                               1                             0   \n",
       "3                               1                             0   \n",
       "14                              1                             0   \n",
       "15                              1                             0   \n",
       "16                              1                             0   \n",
       "\n",
       "    mds-antibiotic-medication  mds-anticoagulant-medication  ...  \\\n",
       "2                           0                             1  ...   \n",
       "3                           1                             0  ...   \n",
       "14                          1                             0  ...   \n",
       "15                          1                             0  ...   \n",
       "16                          1                             0  ...   \n",
       "\n",
       "    atc_5_N06AB10  atc_5_N06AX05  atc_5_N06AX11  atc_5_N06AX12  atc_5_N06AX16  \\\n",
       "2               0              0              0              0              0   \n",
       "3               0              0              0              0              0   \n",
       "14              0              0              0              0              0   \n",
       "15              0              0              0              0              0   \n",
       "16              0              0              0              0              0   \n",
       "\n",
       "    atc_5_N06AX21  atc_5_N06AX24  atc_5_N05AA01  atc_5_N05BA05  atc_5_N06AX23  \n",
       "2               0              0              0              0              0  \n",
       "3               0              0              0              0              0  \n",
       "14              0              0              0              0              0  \n",
       "15              0              0              0              0              0  \n",
       "16              0              0              0              0              0  \n",
       "\n",
       "[5 rows x 398 columns]"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test_use.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "# The main difference in this experiment from experiment 2 is that it  \n",
    "# includes clincally important variables (based on domain knowledge), \n",
    "# imputing them if neede, and drops a bunch of likely non-relevant features \n",
    "# (TODO:recall if this is based on feature selection + expert knowledge?)\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from keras.models import Sequential\n",
    "from keras.layers import SimpleRNN, Dense, Dropout, Masking, TimeDistributed\n",
    "from numpy.random import seed\n",
    "from tensorflow import set_random_seed\n",
    "from copy import deepcopy\n",
    "from keras.preprocessing.sequence import pad_sequences"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "def scaler_transform(X_train_use, X_test_use):    \n",
    "    scaler = MinMaxScaler(feature_range=(0, 1)).fit(X_train_use.drop('patient-id', axis=1))\n",
    "    X_train_transformed = pd.DataFrame(scaler.transform(X_train_use.drop('patient-id', axis=1)), \\\n",
    "                                       columns=X_train_use.columns[1:], index=X_train_use.index)                                   \n",
    "    X_train_transformed['patient-id'] = X_train_use['patient-id']\n",
    "    X_test_transformed = pd.DataFrame(scaler.transform(X_test_use.drop('patient-id', axis=1)), \\\n",
    "                                       columns=X_test_use.columns[1:], index=X_test_use.index) \n",
    "    X_test_transformed['patient-id'] = X_test_use['patient-id']\n",
    "    return X_train_transformed, X_test_transformed"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "def add_previous_outcome(X, Y):\n",
    "    X = X.copy()\n",
    "    x1, x2 = [], []\n",
    "    pids = X['patient-id'].unique()\n",
    "    for pid in pids:\n",
    "        df = Y[X[X['patient-id']==pid].index]\n",
    "        if(df.shape[0] == 1):\n",
    "            x1.append(pd.Series([0]))\n",
    "            x2.append(pd.Series([0]))\n",
    "        else:\n",
    "            # input 1=(1, 0), 0=(0, 1), NA=(0, 0)\n",
    "            x1.append(pd.Series([0]))\n",
    "            x1.append(df[:-1])\n",
    "            x2.append(pd.Series([0]))\n",
    "            x2.append(df[:-1].map({0: 1, 1: 0}))\n",
    "    X['input_outcome_x'] = pd.Series(pd.concat(x1, ignore_index=True).tolist(), index=X.index)\n",
    "    X['input_outcome_y'] = pd.Series(pd.concat(x2, ignore_index=True).tolist(), index=X.index)\n",
    "    return X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "## This step converts the outcome to a point in x,y coordinates where fall (1) = (1,0); no fall (0) = (0,1); and NA = (0,0) \n",
    "X_train_use = add_previous_outcome(X_train_use, y_train)\n",
    "X_test_use = add_previous_outcome(X_test_use, y_test)\n",
    "\n",
    "## This step transforms the data to range between 0 and 1\n",
    "X_train_transformed, X_test_transformed = scaler_transform(X_train_use, X_test_use)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "def generate_nested_list(X, Y):\n",
    "    x_list, y_list = [], [] \n",
    "    pids = X['patient-id'].unique()\n",
    "    for pid in pids:\n",
    "        df = X[X['patient-id']==pid]\n",
    "        y_list.append(Y[df.index].values)\n",
    "        x_list.append([record for record in df.drop('patient-id', axis=1).values])\n",
    "    return x_list, y_list\n",
    "\n",
    "def mygenerator(x_list, y_list=None):\n",
    "    if(y_list is not None):\n",
    "        while True:\n",
    "            for x, y in zip(x_list, y_list):\n",
    "                yield np.array(x).reshape((len(x), 1, x[0].shape[0])), y\n",
    "    else:\n",
    "        while True:\n",
    "            for x in x_list:\n",
    "                yield np.array(x).reshape((len(x), 1, x[0].shape[0])) \n",
    "                \n",
    "def fit_model(model, epoch, X_train, y_train, X_valid=None, y_valid=None, verbose=0):\n",
    "    history = model.fit(X_train, y_train, epochs=epoch, verbose=verbose)\n",
    "    if(verbose != 0):\n",
    "        plot_history(history)\n",
    "    return history\n",
    "\n",
    "def plot_roc_curve(fpr, tpr):\n",
    "    plt.figure()\n",
    "    plt.plot(fpr, tpr, color='darkorange', label='ROC curve (area = %0.2f)' % auc(fpr, tpr))\n",
    "    plt.plot([0, 1], [0, 1], color='navy', linestyle='--')\n",
    "    plt.xlim([0.0, 1.0])\n",
    "    plt.ylim([0.0, 1.05])\n",
    "    plt.xlabel('False Positive Rate')\n",
    "    plt.ylabel('True Positive Rate')\n",
    "    plt.legend(loc=\"lower right\")\n",
    "    plt.show()\n",
    "    \n",
    "def plot_history(history, loss=None, val_loss=None):\n",
    "    if(history is not None):\n",
    "        loss = history.history['loss']\n",
    "    plt.figure()\n",
    "    plt.plot(loss)\n",
    "    plt.title('model loss')\n",
    "    plt.ylabel('loss')\n",
    "    plt.xlabel('epoch')\n",
    "    if('val_loss' in history.history):\n",
    "        val_loss = history.history['val_loss']\n",
    "        plt.plot(val_loss)\n",
    "    plt.legend(['train', 'test'], loc='upper right')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "def pad_data(X, maxlen):\n",
    "    return pad_sequences(X, maxlen=maxlen, dtype='float64', padding='post')\n",
    "\n",
    "def pad_all(X_train, y_train, X_test, y_test, maxlen):\n",
    "    return pad_data(X_train, maxlen), pad_data(y_train, maxlen), pad_data(X_test, maxlen), pad_data(y_test, maxlen)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "def auroc(y_predict, y_true, plot=False):\n",
    "    fpr, tpr, thresholds = roc_curve(y_true, y_predict, pos_label=1)    \n",
    "    if(plot):\n",
    "        plot_roc_curve(fpr, tpr)\n",
    "    return auc(fpr, tpr)\n",
    "\n",
    "\n",
    "def find_best_f1(y_predict, y_true):\n",
    "    fpr, tpr, thresholds = roc_curve(y_true, y_predict, pos_label=1) \n",
    "    max_f1, th = 0, 0\n",
    "    for threshold in thresholds:\n",
    "        f1 = f1_score(y_true, y_predict > threshold)\n",
    "        if(f1 > max_f1):\n",
    "            max_f1 = f1\n",
    "            th = threshold\n",
    "    return max_f1, th\n",
    "\n",
    "def other_metrics(y_predict, y_true, threshold):\n",
    "    y_threshold = y_predict > threshold\n",
    "    tn, fp, fn, tp = confusion_matrix(y_true, y_threshold).ravel()\n",
    "    return precision_score(y_true, y_threshold), recall_score(y_true, y_threshold),\\\n",
    "            accuracy_score(y_true, y_threshold), float(tn) / (tn + fp)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import GroupKFold\n",
    "def kfold_validation(unit, ratio, epoch, X, y, groupby, n_features, k=5, verbose=0, max_timestamps=20):\n",
    "    auroc_list = []\n",
    "    other_metric_list = []\n",
    "    for train_idx, valid_idx in GroupKFold(n_splits=5).split(X, y, X[groupby]):\n",
    "        X_train, X_valid, y_train, y_valid = X.iloc[train_idx], X.iloc[valid_idx], y.iloc[train_idx], y.iloc[valid_idx]\n",
    "        X_train, y_train = generate_nested_list(X_train, y_train)\n",
    "        X_valid, y_valid = generate_nested_list(X_valid, y_valid)\n",
    "        X_train_pad, y_train_pad, X_valid_pad, y_valid_pad = pad_all(X_train, y_train, X_valid, y_valid, max_timestamps)\n",
    "        y_train_pad = y_train_pad.reshape(y_train_pad.shape[0], y_train_pad.shape[1], 1)\n",
    "        seed(1)\n",
    "        set_random_seed(2)\n",
    "        model = build_model(unit, ratio, max_timestamps, n_features)\n",
    "        fit_model(model, epoch, X_train_pad, y_train_pad, verbose=verbose)\n",
    "        y_valid_len = [len(y) for y in y_valid]\n",
    "        y_predict = model.predict(X_valid_pad)\n",
    "        y_predict = truncate_prediction(y_predict, y_valid_len)\n",
    "        y_true = np.concatenate(y_valid)\n",
    "        auroc_list.append(auroc(y_predict, y_true))\n",
    "        f1, th = find_best_f1(y_predict, y_true)\n",
    "        precision, recall, accuracy, specificity = other_metrics(y_predict, y_true, th)\n",
    "        other_metric_list.append((precision, recall, accuracy, specificity,f1,th))\n",
    "    return (auroc_list, other_metric_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "# shape: (num_of_patients, num_of_records, num_of_features)\n",
    "X_train_list, y_train_list = generate_nested_list(X_train_transformed, y_train)\n",
    "X_test_list, y_test_list = generate_nested_list(X_test_transformed, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "## comparing this to the experiment 1 transformed dataset to confirm that this has \n",
    "## diagnoses and ATC codes while the other does not\n",
    "X_train_transformed.to_csv (r'/tmp/exp3_train_transformed_df.csv', index = False, header=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(5301, 400)"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_transformed.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2283, 400)"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test_transformed.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "19\n"
     ]
    }
   ],
   "source": [
    "max_timestamps = df.groupby(['patient-id']).size().max()\n",
    "print(max_timestamps)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "399\n"
     ]
    }
   ],
   "source": [
    "n_features = X_train_list[0][0].shape[0]\n",
    "print(n_features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "## function to build a SimpleRNN\n",
    "from keras import regularizers\n",
    "from keras.optimizers import Adagrad, Adadelta\n",
    "def build_model(units, dropout_ratio, n_timesteps, n_dimensions):\n",
    "    model = Sequential()\n",
    "    model.add(Masking(input_shape=(n_timesteps, n_dimensions)))\n",
    "    model.add(SimpleRNN(units, return_sequences=True, recurrent_dropout=dropout_ratio, dropout=0.5))\n",
    "    model.add(TimeDistributed(Dense(1, activation='sigmoid')))\n",
    "    model.compile(loss='binary_crossentropy', optimizer='adagrad', metrics=['binary_accuracy'])\n",
    "    return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "def truncate_prediction(y_, y_len):\n",
    "    y_concat = []\n",
    "    for i, y in enumerate(y_):\n",
    "        y_concat.append(np.concatenate(y[:y_len[i], :]))\n",
    "    y_concat = np.concatenate(y_concat)\n",
    "    return y_concat"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "## the data is padded so that the final shape is as follows with the outcome data array having the same\n",
    "## shape but only one value\n",
    "#  array([[patient 1 data record 1],  \n",
    "#         [patient 1 data record 2],\n",
    "#         ...\n",
    "#         [patient 1 data record 19],\n",
    "#        ],\n",
    "#        [[patient 2 data record 1],  \n",
    "#         [patient 2 data record 2],\n",
    "#         ...\n",
    "#         [patient 2 data record 19],\n",
    "#        ],\n",
    "#        ...\n",
    "#       )\n",
    "# \n",
    "X_train_pad, y_train_pad, X_test_pad, y_test_pad = pad_all(X_train_list, y_train_list, X_test_list, y_test_list, max_timestamps)\n",
    "y_train_pad = y_train_pad.reshape((2193, 19, 1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING: Logging before flag parsing goes to stderr.\n",
      "W0424 13:34:05.537689 140632023369536 deprecation_wrapper.py:119] From /usr/local/anaconda3/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py:74: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n",
      "\n",
      "W0424 13:34:05.539671 140632023369536 deprecation_wrapper.py:119] From /usr/local/anaconda3/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py:517: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n",
      "\n",
      "W0424 13:34:05.551449 140632023369536 deprecation_wrapper.py:119] From /usr/local/anaconda3/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py:4138: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n",
      "\n",
      "W0424 13:34:05.592048 140632023369536 deprecation_wrapper.py:119] From /usr/local/anaconda3/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py:133: The name tf.placeholder_with_default is deprecated. Please use tf.compat.v1.placeholder_with_default instead.\n",
      "\n",
      "W0424 13:34:05.600967 140632023369536 deprecation.py:506] From /usr/local/anaconda3/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py:3445: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Please use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n",
      "W0424 13:34:05.717592 140632023369536 deprecation.py:323] From /usr/local/anaconda3/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py:2974: add_dispatch_support.<locals>.wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.where in 2.0, which has the same broadcast rule as np.where\n",
      "W0424 13:34:05.770694 140632023369536 deprecation_wrapper.py:119] From /usr/local/anaconda3/lib/python3.7/site-packages/keras/optimizers.py:790: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n",
      "\n",
      "W0424 13:34:05.788399 140632023369536 deprecation_wrapper.py:119] From /usr/local/anaconda3/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py:3376: The name tf.log is deprecated. Please use tf.math.log instead.\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/175\n",
      "2193/2193 [==============================] - 1s 659us/step - loss: 0.4944 - binary_accuracy: 0.8098\n",
      "Epoch 2/175\n",
      "2193/2193 [==============================] - 1s 362us/step - loss: 0.4469 - binary_accuracy: 0.8243\n",
      "Epoch 3/175\n",
      "2193/2193 [==============================] - 1s 362us/step - loss: 0.4417 - binary_accuracy: 0.8161\n",
      "Epoch 4/175\n",
      "2193/2193 [==============================] - 1s 362us/step - loss: 0.4320 - binary_accuracy: 0.8222\n",
      "Epoch 5/175\n",
      "2193/2193 [==============================] - 1s 366us/step - loss: 0.4382 - binary_accuracy: 0.8255\n",
      "Epoch 6/175\n",
      "2193/2193 [==============================] - 1s 390us/step - loss: 0.4277 - binary_accuracy: 0.8239\n",
      "Epoch 7/175\n",
      "2193/2193 [==============================] - 1s 381us/step - loss: 0.4311 - binary_accuracy: 0.8258\n",
      "Epoch 8/175\n",
      "2193/2193 [==============================] - 1s 395us/step - loss: 0.4265 - binary_accuracy: 0.8247\n",
      "Epoch 9/175\n",
      "2193/2193 [==============================] - 1s 371us/step - loss: 0.4246 - binary_accuracy: 0.8241\n",
      "Epoch 10/175\n",
      "2193/2193 [==============================] - 1s 370us/step - loss: 0.4282 - binary_accuracy: 0.8204\n",
      "Epoch 11/175\n",
      "2193/2193 [==============================] - 1s 367us/step - loss: 0.4289 - binary_accuracy: 0.8217\n",
      "Epoch 12/175\n",
      "2193/2193 [==============================] - 1s 372us/step - loss: 0.4312 - binary_accuracy: 0.8229\n",
      "Epoch 13/175\n",
      "2193/2193 [==============================] - 1s 367us/step - loss: 0.4333 - binary_accuracy: 0.8231\n",
      "Epoch 14/175\n",
      "2193/2193 [==============================] - 1s 368us/step - loss: 0.4267 - binary_accuracy: 0.8231\n",
      "Epoch 15/175\n",
      "2193/2193 [==============================] - 1s 368us/step - loss: 0.4310 - binary_accuracy: 0.8213\n",
      "Epoch 16/175\n",
      "2193/2193 [==============================] - 1s 367us/step - loss: 0.4224 - binary_accuracy: 0.8233\n",
      "Epoch 17/175\n",
      "2193/2193 [==============================] - 1s 364us/step - loss: 0.4273 - binary_accuracy: 0.8202\n",
      "Epoch 18/175\n",
      "2193/2193 [==============================] - 1s 366us/step - loss: 0.4272 - binary_accuracy: 0.8185\n",
      "Epoch 19/175\n",
      "2193/2193 [==============================] - 1s 364us/step - loss: 0.4238 - binary_accuracy: 0.8231\n",
      "Epoch 20/175\n",
      "2193/2193 [==============================] - 1s 365us/step - loss: 0.4220 - binary_accuracy: 0.8276\n",
      "Epoch 21/175\n",
      "2193/2193 [==============================] - 1s 367us/step - loss: 0.4227 - binary_accuracy: 0.8227\n",
      "Epoch 22/175\n",
      "2193/2193 [==============================] - 1s 370us/step - loss: 0.4246 - binary_accuracy: 0.8274\n",
      "Epoch 23/175\n",
      "2193/2193 [==============================] - 1s 363us/step - loss: 0.4177 - binary_accuracy: 0.8300\n",
      "Epoch 24/175\n",
      "2193/2193 [==============================] - 1s 363us/step - loss: 0.4212 - binary_accuracy: 0.8262\n",
      "Epoch 25/175\n",
      "2193/2193 [==============================] - 1s 367us/step - loss: 0.4264 - binary_accuracy: 0.8264\n",
      "Epoch 26/175\n",
      "2193/2193 [==============================] - 1s 375us/step - loss: 0.4209 - binary_accuracy: 0.8255\n",
      "Epoch 27/175\n",
      "2193/2193 [==============================] - 1s 366us/step - loss: 0.4215 - binary_accuracy: 0.8256\n",
      "Epoch 28/175\n",
      "2193/2193 [==============================] - 1s 364us/step - loss: 0.4271 - binary_accuracy: 0.8265\n",
      "Epoch 29/175\n",
      "2193/2193 [==============================] - 1s 373us/step - loss: 0.4247 - binary_accuracy: 0.8239\n",
      "Epoch 30/175\n",
      "2193/2193 [==============================] - 1s 364us/step - loss: 0.4183 - binary_accuracy: 0.8297\n",
      "Epoch 31/175\n",
      "2193/2193 [==============================] - 1s 365us/step - loss: 0.4214 - binary_accuracy: 0.8238\n",
      "Epoch 32/175\n",
      "2193/2193 [==============================] - 1s 362us/step - loss: 0.4255 - binary_accuracy: 0.8266\n",
      "Epoch 33/175\n",
      "2193/2193 [==============================] - 1s 364us/step - loss: 0.4183 - binary_accuracy: 0.8238\n",
      "Epoch 34/175\n",
      "2193/2193 [==============================] - 1s 368us/step - loss: 0.4180 - binary_accuracy: 0.8258\n",
      "Epoch 35/175\n",
      "2193/2193 [==============================] - 1s 394us/step - loss: 0.4209 - binary_accuracy: 0.8262\n",
      "Epoch 36/175\n",
      "2193/2193 [==============================] - 1s 381us/step - loss: 0.4205 - binary_accuracy: 0.8287\n",
      "Epoch 37/175\n",
      "2193/2193 [==============================] - 1s 371us/step - loss: 0.4168 - binary_accuracy: 0.8271\n",
      "Epoch 38/175\n",
      "2193/2193 [==============================] - 1s 363us/step - loss: 0.4220 - binary_accuracy: 0.8243\n",
      "Epoch 39/175\n",
      "2193/2193 [==============================] - 1s 365us/step - loss: 0.4197 - binary_accuracy: 0.8247\n",
      "Epoch 40/175\n",
      "2193/2193 [==============================] - 1s 366us/step - loss: 0.4164 - binary_accuracy: 0.8255\n",
      "Epoch 41/175\n",
      "2193/2193 [==============================] - 1s 363us/step - loss: 0.4183 - binary_accuracy: 0.8223\n",
      "Epoch 42/175\n",
      "2193/2193 [==============================] - 1s 364us/step - loss: 0.4214 - binary_accuracy: 0.8214\n",
      "Epoch 43/175\n",
      "2193/2193 [==============================] - 1s 366us/step - loss: 0.4138 - binary_accuracy: 0.8261\n",
      "Epoch 44/175\n",
      "2193/2193 [==============================] - 1s 366us/step - loss: 0.4136 - binary_accuracy: 0.8287\n",
      "Epoch 45/175\n",
      "2193/2193 [==============================] - 1s 372us/step - loss: 0.4178 - binary_accuracy: 0.8294\n",
      "Epoch 46/175\n",
      "2193/2193 [==============================] - 1s 368us/step - loss: 0.4204 - binary_accuracy: 0.8238\n",
      "Epoch 47/175\n",
      "2193/2193 [==============================] - 1s 367us/step - loss: 0.4222 - binary_accuracy: 0.8261\n",
      "Epoch 48/175\n",
      "2193/2193 [==============================] - 1s 371us/step - loss: 0.4197 - binary_accuracy: 0.8247\n",
      "Epoch 49/175\n",
      "2193/2193 [==============================] - 1s 364us/step - loss: 0.4176 - binary_accuracy: 0.8250\n",
      "Epoch 50/175\n",
      "2193/2193 [==============================] - 1s 365us/step - loss: 0.4189 - binary_accuracy: 0.8255\n",
      "Epoch 51/175\n",
      "2193/2193 [==============================] - 1s 364us/step - loss: 0.4206 - binary_accuracy: 0.8259\n",
      "Epoch 52/175\n",
      "2193/2193 [==============================] - 1s 365us/step - loss: 0.4229 - binary_accuracy: 0.8216\n",
      "Epoch 53/175\n",
      "2193/2193 [==============================] - 1s 368us/step - loss: 0.4149 - binary_accuracy: 0.8297\n",
      "Epoch 54/175\n",
      "2193/2193 [==============================] - 1s 369us/step - loss: 0.4102 - binary_accuracy: 0.8313\n",
      "Epoch 55/175\n",
      "2193/2193 [==============================] - 1s 370us/step - loss: 0.4233 - binary_accuracy: 0.8228\n",
      "Epoch 56/175\n",
      "2193/2193 [==============================] - 1s 365us/step - loss: 0.4200 - binary_accuracy: 0.8236\n",
      "Epoch 57/175\n",
      "2193/2193 [==============================] - 1s 365us/step - loss: 0.4153 - binary_accuracy: 0.8265\n",
      "Epoch 58/175\n",
      "2193/2193 [==============================] - 1s 364us/step - loss: 0.4153 - binary_accuracy: 0.8297\n",
      "Epoch 59/175\n",
      "2193/2193 [==============================] - 1s 363us/step - loss: 0.4209 - binary_accuracy: 0.8240\n",
      "Epoch 60/175\n",
      "2193/2193 [==============================] - 1s 363us/step - loss: 0.4185 - binary_accuracy: 0.8271\n",
      "Epoch 61/175\n",
      "2193/2193 [==============================] - 1s 372us/step - loss: 0.4173 - binary_accuracy: 0.8274\n",
      "Epoch 62/175\n",
      "2193/2193 [==============================] - 1s 396us/step - loss: 0.4147 - binary_accuracy: 0.8283\n",
      "Epoch 63/175\n",
      "2193/2193 [==============================] - 1s 376us/step - loss: 0.4187 - binary_accuracy: 0.8288\n",
      "Epoch 64/175\n",
      "2193/2193 [==============================] - 1s 369us/step - loss: 0.4136 - binary_accuracy: 0.8298\n",
      "Epoch 65/175\n",
      "2193/2193 [==============================] - 1s 366us/step - loss: 0.4168 - binary_accuracy: 0.8264\n",
      "Epoch 66/175\n",
      "2193/2193 [==============================] - 1s 365us/step - loss: 0.4184 - binary_accuracy: 0.8208\n",
      "Epoch 67/175\n",
      "2193/2193 [==============================] - 1s 367us/step - loss: 0.4182 - binary_accuracy: 0.8250\n",
      "Epoch 68/175\n",
      "2193/2193 [==============================] - 1s 367us/step - loss: 0.4149 - binary_accuracy: 0.8269\n",
      "Epoch 69/175\n",
      "2193/2193 [==============================] - 1s 366us/step - loss: 0.4202 - binary_accuracy: 0.8271\n",
      "Epoch 70/175\n",
      "2193/2193 [==============================] - 1s 368us/step - loss: 0.4179 - binary_accuracy: 0.8273\n",
      "Epoch 71/175\n",
      "2193/2193 [==============================] - 1s 365us/step - loss: 0.4202 - binary_accuracy: 0.8263\n",
      "Epoch 72/175\n",
      "2193/2193 [==============================] - 1s 365us/step - loss: 0.4188 - binary_accuracy: 0.8239\n",
      "Epoch 73/175\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2193/2193 [==============================] - 1s 366us/step - loss: 0.4159 - binary_accuracy: 0.8283\n",
      "Epoch 74/175\n",
      "2193/2193 [==============================] - 1s 369us/step - loss: 0.4190 - binary_accuracy: 0.8241\n",
      "Epoch 75/175\n",
      "2193/2193 [==============================] - 1s 365us/step - loss: 0.4131 - binary_accuracy: 0.8289\n",
      "Epoch 76/175\n",
      "2193/2193 [==============================] - 1s 362us/step - loss: 0.4197 - binary_accuracy: 0.8240\n",
      "Epoch 77/175\n",
      "2193/2193 [==============================] - 1s 368us/step - loss: 0.4169 - binary_accuracy: 0.8262\n",
      "Epoch 78/175\n",
      "2193/2193 [==============================] - 1s 368us/step - loss: 0.4216 - binary_accuracy: 0.8244\n",
      "Epoch 79/175\n",
      "2193/2193 [==============================] - 1s 367us/step - loss: 0.4159 - binary_accuracy: 0.8292\n",
      "Epoch 80/175\n",
      "2193/2193 [==============================] - 1s 366us/step - loss: 0.4088 - binary_accuracy: 0.8294\n",
      "Epoch 81/175\n",
      "2193/2193 [==============================] - 1s 364us/step - loss: 0.4156 - binary_accuracy: 0.8268\n",
      "Epoch 82/175\n",
      "2193/2193 [==============================] - 1s 361us/step - loss: 0.4112 - binary_accuracy: 0.8313\n",
      "Epoch 83/175\n",
      "2193/2193 [==============================] - 1s 372us/step - loss: 0.4103 - binary_accuracy: 0.8310\n",
      "Epoch 84/175\n",
      "2193/2193 [==============================] - 1s 370us/step - loss: 0.4169 - binary_accuracy: 0.8271\n",
      "Epoch 85/175\n",
      "2193/2193 [==============================] - 1s 366us/step - loss: 0.4155 - binary_accuracy: 0.8305\n",
      "Epoch 86/175\n",
      "2193/2193 [==============================] - 1s 372us/step - loss: 0.4190 - binary_accuracy: 0.8293\n",
      "Epoch 87/175\n",
      "2193/2193 [==============================] - 1s 364us/step - loss: 0.4155 - binary_accuracy: 0.8287\n",
      "Epoch 88/175\n",
      "2193/2193 [==============================] - 1s 368us/step - loss: 0.4154 - binary_accuracy: 0.8262\n",
      "Epoch 89/175\n",
      "2193/2193 [==============================] - 1s 390us/step - loss: 0.4142 - binary_accuracy: 0.8305\n",
      "Epoch 90/175\n",
      "2193/2193 [==============================] - 1s 380us/step - loss: 0.4128 - binary_accuracy: 0.8290\n",
      "Epoch 91/175\n",
      "2193/2193 [==============================] - 1s 367us/step - loss: 0.4171 - binary_accuracy: 0.8286\n",
      "Epoch 92/175\n",
      "2193/2193 [==============================] - 1s 367us/step - loss: 0.4126 - binary_accuracy: 0.8301\n",
      "Epoch 93/175\n",
      "2193/2193 [==============================] - 1s 366us/step - loss: 0.4189 - binary_accuracy: 0.8244\n",
      "Epoch 94/175\n",
      "2193/2193 [==============================] - 1s 380us/step - loss: 0.4116 - binary_accuracy: 0.8301\n",
      "Epoch 95/175\n",
      "2193/2193 [==============================] - 1s 367us/step - loss: 0.4099 - binary_accuracy: 0.8251\n",
      "Epoch 96/175\n",
      "2193/2193 [==============================] - 1s 366us/step - loss: 0.4111 - binary_accuracy: 0.8277\n",
      "Epoch 97/175\n",
      "2193/2193 [==============================] - 1s 363us/step - loss: 0.4155 - binary_accuracy: 0.8268\n",
      "Epoch 98/175\n",
      "2193/2193 [==============================] - 1s 362us/step - loss: 0.4110 - binary_accuracy: 0.8271\n",
      "Epoch 99/175\n",
      "2193/2193 [==============================] - 1s 363us/step - loss: 0.4220 - binary_accuracy: 0.8281\n",
      "Epoch 100/175\n",
      "2193/2193 [==============================] - 1s 364us/step - loss: 0.4195 - binary_accuracy: 0.8263\n",
      "Epoch 101/175\n",
      "2193/2193 [==============================] - 1s 363us/step - loss: 0.4151 - binary_accuracy: 0.8254\n",
      "Epoch 102/175\n",
      "2193/2193 [==============================] - 1s 364us/step - loss: 0.4156 - binary_accuracy: 0.8268\n",
      "Epoch 103/175\n",
      "2193/2193 [==============================] - 1s 368us/step - loss: 0.4119 - binary_accuracy: 0.8250\n",
      "Epoch 104/175\n",
      "2193/2193 [==============================] - 1s 376us/step - loss: 0.4106 - binary_accuracy: 0.8301\n",
      "Epoch 105/175\n",
      "2193/2193 [==============================] - 1s 375us/step - loss: 0.4125 - binary_accuracy: 0.8325\n",
      "Epoch 106/175\n",
      "2193/2193 [==============================] - 1s 371us/step - loss: 0.4134 - binary_accuracy: 0.8274\n",
      "Epoch 107/175\n",
      "2193/2193 [==============================] - 1s 370us/step - loss: 0.4197 - binary_accuracy: 0.8237\n",
      "Epoch 108/175\n",
      "2193/2193 [==============================] - 1s 368us/step - loss: 0.4104 - binary_accuracy: 0.8304\n",
      "Epoch 109/175\n",
      "2193/2193 [==============================] - 1s 367us/step - loss: 0.4098 - binary_accuracy: 0.8304\n",
      "Epoch 110/175\n",
      "2193/2193 [==============================] - 1s 365us/step - loss: 0.4142 - binary_accuracy: 0.8265\n",
      "Epoch 111/175\n",
      "2193/2193 [==============================] - 1s 366us/step - loss: 0.4146 - binary_accuracy: 0.8277\n",
      "Epoch 112/175\n",
      "2193/2193 [==============================] - 1s 369us/step - loss: 0.4134 - binary_accuracy: 0.8278\n",
      "Epoch 113/175\n",
      "2193/2193 [==============================] - 1s 363us/step - loss: 0.4162 - binary_accuracy: 0.8233\n",
      "Epoch 114/175\n",
      "2193/2193 [==============================] - 1s 365us/step - loss: 0.4170 - binary_accuracy: 0.8259\n",
      "Epoch 115/175\n",
      "2193/2193 [==============================] - 1s 363us/step - loss: 0.4196 - binary_accuracy: 0.8205\n",
      "Epoch 116/175\n",
      "2193/2193 [==============================] - 1s 375us/step - loss: 0.4128 - binary_accuracy: 0.8316\n",
      "Epoch 117/175\n",
      "2193/2193 [==============================] - 1s 390us/step - loss: 0.4194 - binary_accuracy: 0.8275\n",
      "Epoch 118/175\n",
      "2193/2193 [==============================] - 1s 362us/step - loss: 0.4168 - binary_accuracy: 0.8278\n",
      "Epoch 119/175\n",
      "2193/2193 [==============================] - 1s 376us/step - loss: 0.4121 - binary_accuracy: 0.8234\n",
      "Epoch 120/175\n",
      "2193/2193 [==============================] - 1s 371us/step - loss: 0.4163 - binary_accuracy: 0.8266\n",
      "Epoch 121/175\n",
      "2193/2193 [==============================] - 1s 363us/step - loss: 0.4196 - binary_accuracy: 0.8260\n",
      "Epoch 122/175\n",
      "2193/2193 [==============================] - 1s 365us/step - loss: 0.4103 - binary_accuracy: 0.8287\n",
      "Epoch 123/175\n",
      "2193/2193 [==============================] - 1s 364us/step - loss: 0.4177 - binary_accuracy: 0.8247\n",
      "Epoch 124/175\n",
      "2193/2193 [==============================] - 1s 365us/step - loss: 0.4079 - binary_accuracy: 0.8313\n",
      "Epoch 125/175\n",
      "2193/2193 [==============================] - 1s 364us/step - loss: 0.4080 - binary_accuracy: 0.8365\n",
      "Epoch 126/175\n",
      "2193/2193 [==============================] - 1s 365us/step - loss: 0.4151 - binary_accuracy: 0.8277\n",
      "Epoch 127/175\n",
      "2193/2193 [==============================] - 1s 363us/step - loss: 0.4155 - binary_accuracy: 0.8258\n",
      "Epoch 128/175\n",
      "2193/2193 [==============================] - 1s 361us/step - loss: 0.4150 - binary_accuracy: 0.8296\n",
      "Epoch 129/175\n",
      "2193/2193 [==============================] - 1s 363us/step - loss: 0.4142 - binary_accuracy: 0.8292\n",
      "Epoch 130/175\n",
      "2193/2193 [==============================] - 1s 368us/step - loss: 0.4180 - binary_accuracy: 0.8257\n",
      "Epoch 131/175\n",
      "2193/2193 [==============================] - 1s 368us/step - loss: 0.4071 - binary_accuracy: 0.8314\n",
      "Epoch 132/175\n",
      "2193/2193 [==============================] - 1s 361us/step - loss: 0.4166 - binary_accuracy: 0.8270\n",
      "Epoch 133/175\n",
      "2193/2193 [==============================] - 1s 363us/step - loss: 0.4149 - binary_accuracy: 0.8266\n",
      "Epoch 134/175\n",
      "2193/2193 [==============================] - 1s 365us/step - loss: 0.4105 - binary_accuracy: 0.8280\n",
      "Epoch 135/175\n",
      "2193/2193 [==============================] - 1s 365us/step - loss: 0.4036 - binary_accuracy: 0.8331\n",
      "Epoch 136/175\n",
      "2193/2193 [==============================] - 1s 365us/step - loss: 0.4157 - binary_accuracy: 0.8277\n",
      "Epoch 137/175\n",
      "2193/2193 [==============================] - 1s 368us/step - loss: 0.4066 - binary_accuracy: 0.8285\n",
      "Epoch 138/175\n",
      "2193/2193 [==============================] - 1s 365us/step - loss: 0.4086 - binary_accuracy: 0.8305\n",
      "Epoch 139/175\n",
      "2193/2193 [==============================] - 1s 371us/step - loss: 0.4179 - binary_accuracy: 0.8275\n",
      "Epoch 140/175\n",
      "2193/2193 [==============================] - 1s 369us/step - loss: 0.4183 - binary_accuracy: 0.8191\n",
      "Epoch 141/175\n",
      "2193/2193 [==============================] - 1s 364us/step - loss: 0.4176 - binary_accuracy: 0.8239\n",
      "Epoch 142/175\n",
      "2193/2193 [==============================] - 1s 366us/step - loss: 0.4102 - binary_accuracy: 0.8286\n",
      "Epoch 143/175\n",
      "2193/2193 [==============================] - 1s 364us/step - loss: 0.4051 - binary_accuracy: 0.8340\n",
      "Epoch 144/175\n",
      "2193/2193 [==============================] - 1s 366us/step - loss: 0.4072 - binary_accuracy: 0.8287\n",
      "Epoch 145/175\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2193/2193 [==============================] - 1s 376us/step - loss: 0.4100 - binary_accuracy: 0.8291\n",
      "Epoch 146/175\n",
      "2193/2193 [==============================] - 1s 400us/step - loss: 0.4119 - binary_accuracy: 0.8284\n",
      "Epoch 147/175\n",
      "2193/2193 [==============================] - 1s 363us/step - loss: 0.4129 - binary_accuracy: 0.8299\n",
      "Epoch 148/175\n",
      "2193/2193 [==============================] - 1s 363us/step - loss: 0.4088 - binary_accuracy: 0.8303\n",
      "Epoch 149/175\n",
      "2193/2193 [==============================] - 1s 362us/step - loss: 0.4161 - binary_accuracy: 0.8266\n",
      "Epoch 150/175\n",
      "2193/2193 [==============================] - 1s 366us/step - loss: 0.4109 - binary_accuracy: 0.8280\n",
      "Epoch 151/175\n",
      "2193/2193 [==============================] - 1s 364us/step - loss: 0.4096 - binary_accuracy: 0.8304\n",
      "Epoch 152/175\n",
      "2193/2193 [==============================] - 1s 365us/step - loss: 0.4150 - binary_accuracy: 0.8272\n",
      "Epoch 153/175\n",
      "2193/2193 [==============================] - 1s 367us/step - loss: 0.4116 - binary_accuracy: 0.8299\n",
      "Epoch 154/175\n",
      "2193/2193 [==============================] - 1s 360us/step - loss: 0.4117 - binary_accuracy: 0.8265\n",
      "Epoch 155/175\n",
      "2193/2193 [==============================] - 1s 361us/step - loss: 0.4161 - binary_accuracy: 0.8273\n",
      "Epoch 156/175\n",
      "2193/2193 [==============================] - 1s 363us/step - loss: 0.4141 - binary_accuracy: 0.8240\n",
      "Epoch 157/175\n",
      "2193/2193 [==============================] - 1s 367us/step - loss: 0.4070 - binary_accuracy: 0.8317\n",
      "Epoch 158/175\n",
      "2193/2193 [==============================] - 1s 364us/step - loss: 0.4073 - binary_accuracy: 0.8323\n",
      "Epoch 159/175\n",
      "2193/2193 [==============================] - 1s 361us/step - loss: 0.4086 - binary_accuracy: 0.8321\n",
      "Epoch 160/175\n",
      "2193/2193 [==============================] - 1s 364us/step - loss: 0.4100 - binary_accuracy: 0.8329\n",
      "Epoch 161/175\n",
      "2193/2193 [==============================] - 1s 360us/step - loss: 0.4077 - binary_accuracy: 0.8337\n",
      "Epoch 162/175\n",
      "2193/2193 [==============================] - 1s 357us/step - loss: 0.4139 - binary_accuracy: 0.8273\n",
      "Epoch 163/175\n",
      "2193/2193 [==============================] - 1s 365us/step - loss: 0.4150 - binary_accuracy: 0.8314\n",
      "Epoch 164/175\n",
      "2193/2193 [==============================] - 1s 361us/step - loss: 0.4139 - binary_accuracy: 0.8253\n",
      "Epoch 165/175\n",
      "2193/2193 [==============================] - 1s 359us/step - loss: 0.4140 - binary_accuracy: 0.8263\n",
      "Epoch 166/175\n",
      "2193/2193 [==============================] - 1s 361us/step - loss: 0.4112 - binary_accuracy: 0.8311\n",
      "Epoch 167/175\n",
      "2193/2193 [==============================] - 1s 363us/step - loss: 0.4059 - binary_accuracy: 0.8338\n",
      "Epoch 168/175\n",
      "2193/2193 [==============================] - 1s 361us/step - loss: 0.4092 - binary_accuracy: 0.8306\n",
      "Epoch 169/175\n",
      "2193/2193 [==============================] - 1s 366us/step - loss: 0.4194 - binary_accuracy: 0.8241\n",
      "Epoch 170/175\n",
      "2193/2193 [==============================] - 1s 364us/step - loss: 0.4171 - binary_accuracy: 0.8272\n",
      "Epoch 171/175\n",
      "2193/2193 [==============================] - 1s 369us/step - loss: 0.4163 - binary_accuracy: 0.8252\n",
      "Epoch 172/175\n",
      "2193/2193 [==============================] - 1s 372us/step - loss: 0.4104 - binary_accuracy: 0.8315\n",
      "Epoch 173/175\n",
      "2193/2193 [==============================] - 1s 364us/step - loss: 0.4118 - binary_accuracy: 0.8266\n",
      "Epoch 174/175\n",
      "2193/2193 [==============================] - 1s 365us/step - loss: 0.4136 - binary_accuracy: 0.8279\n",
      "Epoch 175/175\n",
      "2193/2193 [==============================] - 1s 368us/step - loss: 0.4144 - binary_accuracy: 0.8270\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "auroc:  0.710944385494039\n",
      "f1:  0.43443443443443447\n",
      "threshold:  0.19736037\n",
      "precision:  0.3833922261484099\n",
      "recall:  0.5011547344110855\n",
      "accuracy:  0.7525186158563294\n",
      "specificity:  0.8113513513513514\n",
      "CPU times: user 7min 25s, sys: 1min 7s, total: 8min 33s\n",
      "Wall time: 2min 24s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "### A single run w/out cross-validation\n",
    "# with previous outcome added to the features\n",
    "# outcome is added before scaling\n",
    "seed(1)\n",
    "set_random_seed(2)\n",
    "model = build_model(32, 0.5, max_timestamps, n_features)\n",
    "fit_model(model, 175, X_train_pad, y_train_pad, verbose=1)\n",
    "y_predict = model.predict(X_test_pad)\n",
    "y_test_len = [len(y) for y in y_test_list]\n",
    "y_predict = truncate_prediction(y_predict, y_test_len)\n",
    "y_true = np.concatenate(y_test_list)\n",
    "print('auroc: ', auroc(y_predict, y_true))\n",
    "f1, th = find_best_f1(y_predict, y_true)\n",
    "print('f1: ', f1);\n",
    "print('threshold: ', th)\n",
    "precision, recall, accuracy, specificity = other_metrics(y_predict, y_true, th)\n",
    "print('precision: ', precision)\n",
    "print('recall: ', recall)\n",
    "print('accuracy: ', accuracy)\n",
    "print('specificity: ', specificity)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/225\n",
      "1755/1755 [==============================] - 1s 736us/step - loss: 0.4764 - binary_accuracy: 0.8125\n",
      "Epoch 2/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4521 - binary_accuracy: 0.8140\n",
      "Epoch 3/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4416 - binary_accuracy: 0.8194\n",
      "Epoch 4/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.4427 - binary_accuracy: 0.8181\n",
      "Epoch 5/225\n",
      "1755/1755 [==============================] - 1s 407us/step - loss: 0.4349 - binary_accuracy: 0.8213\n",
      "Epoch 6/225\n",
      "1755/1755 [==============================] - 1s 409us/step - loss: 0.4362 - binary_accuracy: 0.8192\n",
      "Epoch 7/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4361 - binary_accuracy: 0.8218\n",
      "Epoch 8/225\n",
      "1755/1755 [==============================] - 1s 390us/step - loss: 0.4292 - binary_accuracy: 0.8254\n",
      "Epoch 9/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.4312 - binary_accuracy: 0.8263\n",
      "Epoch 10/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4313 - binary_accuracy: 0.8214\n",
      "Epoch 11/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4231 - binary_accuracy: 0.8215\n",
      "Epoch 12/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4154 - binary_accuracy: 0.8307\n",
      "Epoch 13/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4201 - binary_accuracy: 0.8266\n",
      "Epoch 14/225\n",
      "1755/1755 [==============================] - 1s 388us/step - loss: 0.4264 - binary_accuracy: 0.8225\n",
      "Epoch 15/225\n",
      "1755/1755 [==============================] - 1s 394us/step - loss: 0.4310 - binary_accuracy: 0.8193\n",
      "Epoch 16/225\n",
      "1755/1755 [==============================] - 1s 388us/step - loss: 0.4169 - binary_accuracy: 0.8245\n",
      "Epoch 17/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.4191 - binary_accuracy: 0.8251\n",
      "Epoch 18/225\n",
      "1755/1755 [==============================] - 1s 390us/step - loss: 0.4198 - binary_accuracy: 0.8294\n",
      "Epoch 19/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.4258 - binary_accuracy: 0.8278\n",
      "Epoch 20/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.4238 - binary_accuracy: 0.8235\n",
      "Epoch 21/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.4216 - binary_accuracy: 0.8254\n",
      "Epoch 22/225\n",
      "1755/1755 [==============================] - 1s 404us/step - loss: 0.4288 - binary_accuracy: 0.8201\n",
      "Epoch 23/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.4187 - binary_accuracy: 0.8229\n",
      "Epoch 24/225\n",
      "1755/1755 [==============================] - 1s 406us/step - loss: 0.4177 - binary_accuracy: 0.8256\n",
      "Epoch 25/225\n",
      "1755/1755 [==============================] - 1s 404us/step - loss: 0.4182 - binary_accuracy: 0.8255\n",
      "Epoch 26/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.4181 - binary_accuracy: 0.8271\n",
      "Epoch 27/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.4180 - binary_accuracy: 0.8279\n",
      "Epoch 28/225\n",
      "1755/1755 [==============================] - 1s 393us/step - loss: 0.4204 - binary_accuracy: 0.8254\n",
      "Epoch 29/225\n",
      "1755/1755 [==============================] - 1s 401us/step - loss: 0.4214 - binary_accuracy: 0.8234\n",
      "Epoch 30/225\n",
      "1755/1755 [==============================] - 1s 420us/step - loss: 0.4158 - binary_accuracy: 0.8269\n",
      "Epoch 31/225\n",
      "1755/1755 [==============================] - 1s 389us/step - loss: 0.4147 - binary_accuracy: 0.8213\n",
      "Epoch 32/225\n",
      "1755/1755 [==============================] - 1s 387us/step - loss: 0.4204 - binary_accuracy: 0.8256\n",
      "Epoch 33/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4209 - binary_accuracy: 0.8240\n",
      "Epoch 34/225\n",
      "1755/1755 [==============================] - 1s 391us/step - loss: 0.4252 - binary_accuracy: 0.8211\n",
      "Epoch 35/225\n",
      "1755/1755 [==============================] - 1s 387us/step - loss: 0.4206 - binary_accuracy: 0.8253\n",
      "Epoch 36/225\n",
      "1755/1755 [==============================] - 1s 390us/step - loss: 0.4202 - binary_accuracy: 0.8253\n",
      "Epoch 37/225\n",
      "1755/1755 [==============================] - 1s 388us/step - loss: 0.4196 - binary_accuracy: 0.8282\n",
      "Epoch 38/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4136 - binary_accuracy: 0.8255\n",
      "Epoch 39/225\n",
      "1755/1755 [==============================] - 1s 387us/step - loss: 0.4158 - binary_accuracy: 0.8314\n",
      "Epoch 40/225\n",
      "1755/1755 [==============================] - 1s 387us/step - loss: 0.4180 - binary_accuracy: 0.8261\n",
      "Epoch 41/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4176 - binary_accuracy: 0.8268\n",
      "Epoch 42/225\n",
      "1755/1755 [==============================] - 1s 386us/step - loss: 0.4149 - binary_accuracy: 0.8265\n",
      "Epoch 43/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4245 - binary_accuracy: 0.8237\n",
      "Epoch 44/225\n",
      "1755/1755 [==============================] - 1s 410us/step - loss: 0.4198 - binary_accuracy: 0.8295\n",
      "Epoch 45/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4234 - binary_accuracy: 0.8255\n",
      "Epoch 46/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4141 - binary_accuracy: 0.8281\n",
      "Epoch 47/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4122 - binary_accuracy: 0.8270\n",
      "Epoch 48/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4184 - binary_accuracy: 0.8244\n",
      "Epoch 49/225\n",
      "1755/1755 [==============================] - 1s 395us/step - loss: 0.4083 - binary_accuracy: 0.8266\n",
      "Epoch 50/225\n",
      "1755/1755 [==============================] - 1s 386us/step - loss: 0.4177 - binary_accuracy: 0.8263\n",
      "Epoch 51/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4114 - binary_accuracy: 0.8338\n",
      "Epoch 52/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4204 - binary_accuracy: 0.8277\n",
      "Epoch 53/225\n",
      "1755/1755 [==============================] - 1s 376us/step - loss: 0.4191 - binary_accuracy: 0.8243\n",
      "Epoch 54/225\n",
      "1755/1755 [==============================] - 1s 376us/step - loss: 0.4206 - binary_accuracy: 0.8260\n",
      "Epoch 55/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.4223 - binary_accuracy: 0.8205\n",
      "Epoch 56/225\n",
      "1755/1755 [==============================] - 1s 419us/step - loss: 0.4156 - binary_accuracy: 0.8264\n",
      "Epoch 57/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.4190 - binary_accuracy: 0.8253\n",
      "Epoch 58/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.4160 - binary_accuracy: 0.8262\n",
      "Epoch 59/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4150 - binary_accuracy: 0.8308\n",
      "Epoch 60/225\n",
      "1755/1755 [==============================] - 1s 377us/step - loss: 0.4084 - binary_accuracy: 0.8274\n",
      "Epoch 61/225\n",
      "1755/1755 [==============================] - 1s 395us/step - loss: 0.4129 - binary_accuracy: 0.8293\n",
      "Epoch 62/225\n",
      "1755/1755 [==============================] - 1s 386us/step - loss: 0.4173 - binary_accuracy: 0.8251\n",
      "Epoch 63/225\n",
      "1755/1755 [==============================] - 1s 386us/step - loss: 0.4131 - binary_accuracy: 0.8271\n",
      "Epoch 64/225\n",
      "1755/1755 [==============================] - 1s 510us/step - loss: 0.4072 - binary_accuracy: 0.8304\n",
      "Epoch 65/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4115 - binary_accuracy: 0.8263\n",
      "Epoch 66/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4161 - binary_accuracy: 0.8242\n",
      "Epoch 67/225\n",
      "1755/1755 [==============================] - 1s 389us/step - loss: 0.4170 - binary_accuracy: 0.8252\n",
      "Epoch 68/225\n",
      "1755/1755 [==============================] - 1s 390us/step - loss: 0.4166 - binary_accuracy: 0.8279\n",
      "Epoch 69/225\n",
      "1755/1755 [==============================] - 1s 388us/step - loss: 0.4204 - binary_accuracy: 0.8259\n",
      "Epoch 70/225\n",
      "1755/1755 [==============================] - 1s 389us/step - loss: 0.4184 - binary_accuracy: 0.8275\n",
      "Epoch 71/225\n",
      "1755/1755 [==============================] - 1s 387us/step - loss: 0.4163 - binary_accuracy: 0.8287\n",
      "Epoch 72/225\n",
      "1755/1755 [==============================] - 1s 391us/step - loss: 0.4189 - binary_accuracy: 0.8254\n",
      "Epoch 73/225\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1755/1755 [==============================] - 1s 386us/step - loss: 0.4172 - binary_accuracy: 0.8297\n",
      "Epoch 74/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4126 - binary_accuracy: 0.8255\n",
      "Epoch 75/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4155 - binary_accuracy: 0.8213\n",
      "Epoch 76/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4098 - binary_accuracy: 0.8277\n",
      "Epoch 77/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4139 - binary_accuracy: 0.8238\n",
      "Epoch 78/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4142 - binary_accuracy: 0.8231\n",
      "Epoch 79/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4115 - binary_accuracy: 0.8335\n",
      "Epoch 80/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4173 - binary_accuracy: 0.8270\n",
      "Epoch 81/225\n",
      "1755/1755 [==============================] - 1s 386us/step - loss: 0.4057 - binary_accuracy: 0.8293\n",
      "Epoch 82/225\n",
      "1755/1755 [==============================] - 1s 391us/step - loss: 0.4256 - binary_accuracy: 0.8215\n",
      "Epoch 83/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4204 - binary_accuracy: 0.8239\n",
      "Epoch 84/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4079 - binary_accuracy: 0.8296\n",
      "Epoch 85/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4125 - binary_accuracy: 0.8247\n",
      "Epoch 86/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.4147 - binary_accuracy: 0.8289\n",
      "Epoch 87/225\n",
      "1755/1755 [==============================] - 1s 390us/step - loss: 0.4130 - binary_accuracy: 0.8294\n",
      "Epoch 88/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4140 - binary_accuracy: 0.8304\n",
      "Epoch 89/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4100 - binary_accuracy: 0.8309\n",
      "Epoch 90/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4098 - binary_accuracy: 0.8266\n",
      "Epoch 91/225\n",
      "1755/1755 [==============================] - 1s 387us/step - loss: 0.4129 - binary_accuracy: 0.8297\n",
      "Epoch 92/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4133 - binary_accuracy: 0.8233\n",
      "Epoch 93/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4143 - binary_accuracy: 0.8297\n",
      "Epoch 94/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4073 - binary_accuracy: 0.8276\n",
      "Epoch 95/225\n",
      "1755/1755 [==============================] - 1s 423us/step - loss: 0.4174 - binary_accuracy: 0.8247\n",
      "Epoch 96/225\n",
      "1755/1755 [==============================] - ETA: 0s - loss: 0.4188 - binary_accuracy: 0.821 - 1s 397us/step - loss: 0.4135 - binary_accuracy: 0.8243\n",
      "Epoch 97/225\n",
      "1755/1755 [==============================] - 1s 387us/step - loss: 0.4106 - binary_accuracy: 0.8312\n",
      "Epoch 98/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4118 - binary_accuracy: 0.8294\n",
      "Epoch 99/225\n",
      "1755/1755 [==============================] - 1s 377us/step - loss: 0.4097 - binary_accuracy: 0.8319\n",
      "Epoch 100/225\n",
      "1755/1755 [==============================] - 1s 390us/step - loss: 0.4143 - binary_accuracy: 0.8300\n",
      "Epoch 101/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4157 - binary_accuracy: 0.8256\n",
      "Epoch 102/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4156 - binary_accuracy: 0.8263\n",
      "Epoch 103/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4170 - binary_accuracy: 0.8251\n",
      "Epoch 104/225\n",
      "1755/1755 [==============================] - 1s 377us/step - loss: 0.4101 - binary_accuracy: 0.8297\n",
      "Epoch 105/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4183 - binary_accuracy: 0.8271\n",
      "Epoch 106/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4204 - binary_accuracy: 0.8255\n",
      "Epoch 107/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4174 - binary_accuracy: 0.8301\n",
      "Epoch 108/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4166 - binary_accuracy: 0.8230\n",
      "Epoch 109/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4136 - binary_accuracy: 0.8259\n",
      "Epoch 110/225\n",
      "1755/1755 [==============================] - 1s 391us/step - loss: 0.4070 - binary_accuracy: 0.8293\n",
      "Epoch 111/225\n",
      "1755/1755 [==============================] - 1s 387us/step - loss: 0.4078 - binary_accuracy: 0.8203\n",
      "Epoch 112/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4180 - binary_accuracy: 0.8251\n",
      "Epoch 113/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4151 - binary_accuracy: 0.8272\n",
      "Epoch 114/225\n",
      "1755/1755 [==============================] - 1s 387us/step - loss: 0.4104 - binary_accuracy: 0.8324\n",
      "Epoch 115/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4146 - binary_accuracy: 0.8298\n",
      "Epoch 116/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4120 - binary_accuracy: 0.8285\n",
      "Epoch 117/225\n",
      "1755/1755 [==============================] - 1s 387us/step - loss: 0.4162 - binary_accuracy: 0.8261\n",
      "Epoch 118/225\n",
      "1755/1755 [==============================] - 1s 390us/step - loss: 0.4240 - binary_accuracy: 0.8263\n",
      "Epoch 119/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4093 - binary_accuracy: 0.8265\n",
      "Epoch 120/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4153 - binary_accuracy: 0.8240\n",
      "Epoch 121/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4139 - binary_accuracy: 0.8280\n",
      "Epoch 122/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4114 - binary_accuracy: 0.8293\n",
      "Epoch 123/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4142 - binary_accuracy: 0.8280\n",
      "Epoch 124/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4111 - binary_accuracy: 0.8293\n",
      "Epoch 125/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4051 - binary_accuracy: 0.8347\n",
      "Epoch 126/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4092 - binary_accuracy: 0.8272\n",
      "Epoch 127/225\n",
      "1755/1755 [==============================] - 1s 386us/step - loss: 0.4019 - binary_accuracy: 0.8318\n",
      "Epoch 128/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4209 - binary_accuracy: 0.8242\n",
      "Epoch 129/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4124 - binary_accuracy: 0.8247\n",
      "Epoch 130/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.4117 - binary_accuracy: 0.8301\n",
      "Epoch 131/225\n",
      "1755/1755 [==============================] - 1s 391us/step - loss: 0.4112 - binary_accuracy: 0.8275\n",
      "Epoch 132/225\n",
      "1755/1755 [==============================] - 1s 388us/step - loss: 0.3998 - binary_accuracy: 0.8363\n",
      "Epoch 133/225\n",
      "1755/1755 [==============================] - 1s 401us/step - loss: 0.4104 - binary_accuracy: 0.8249\n",
      "Epoch 134/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4027 - binary_accuracy: 0.8299\n",
      "Epoch 135/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4154 - binary_accuracy: 0.8298\n",
      "Epoch 136/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.4122 - binary_accuracy: 0.8228\n",
      "Epoch 137/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4125 - binary_accuracy: 0.8275\n",
      "Epoch 138/225\n",
      "1755/1755 [==============================] - 1s 387us/step - loss: 0.4150 - binary_accuracy: 0.8244\n",
      "Epoch 139/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4038 - binary_accuracy: 0.8281\n",
      "Epoch 140/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4004 - binary_accuracy: 0.8334\n",
      "Epoch 141/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4096 - binary_accuracy: 0.8271\n",
      "Epoch 142/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4125 - binary_accuracy: 0.8273\n",
      "Epoch 143/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4094 - binary_accuracy: 0.8286\n",
      "Epoch 144/225\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1755/1755 [==============================] - 1s 387us/step - loss: 0.4065 - binary_accuracy: 0.8305\n",
      "Epoch 145/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4112 - binary_accuracy: 0.8302\n",
      "Epoch 146/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4161 - binary_accuracy: 0.8243\n",
      "Epoch 147/225\n",
      "1755/1755 [==============================] - 1s 386us/step - loss: 0.4104 - binary_accuracy: 0.8301\n",
      "Epoch 148/225\n",
      "1755/1755 [==============================] - 1s 387us/step - loss: 0.4050 - binary_accuracy: 0.8322\n",
      "Epoch 149/225\n",
      "1755/1755 [==============================] - 1s 386us/step - loss: 0.4104 - binary_accuracy: 0.8298\n",
      "Epoch 150/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4056 - binary_accuracy: 0.8269\n",
      "Epoch 151/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4055 - binary_accuracy: 0.8308\n",
      "Epoch 152/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4140 - binary_accuracy: 0.8269\n",
      "Epoch 153/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4117 - binary_accuracy: 0.8318\n",
      "Epoch 154/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4093 - binary_accuracy: 0.8288\n",
      "Epoch 155/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4159 - binary_accuracy: 0.8235\n",
      "Epoch 156/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4104 - binary_accuracy: 0.8290\n",
      "Epoch 157/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4061 - binary_accuracy: 0.8299\n",
      "Epoch 158/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4065 - binary_accuracy: 0.8308\n",
      "Epoch 159/225\n",
      "1755/1755 [==============================] - 1s 405us/step - loss: 0.4031 - binary_accuracy: 0.8299\n",
      "Epoch 160/225\n",
      "1755/1755 [==============================] - 1s 417us/step - loss: 0.4101 - binary_accuracy: 0.8280\n",
      "Epoch 161/225\n",
      "1755/1755 [==============================] - 1s 386us/step - loss: 0.4163 - binary_accuracy: 0.8261\n",
      "Epoch 162/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4099 - binary_accuracy: 0.8319\n",
      "Epoch 163/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4042 - binary_accuracy: 0.8306\n",
      "Epoch 164/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.4029 - binary_accuracy: 0.8261\n",
      "Epoch 165/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4058 - binary_accuracy: 0.8333\n",
      "Epoch 166/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4137 - binary_accuracy: 0.8259\n",
      "Epoch 167/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4096 - binary_accuracy: 0.8308\n",
      "Epoch 168/225\n",
      "1755/1755 [==============================] - 1s 387us/step - loss: 0.4098 - binary_accuracy: 0.8312\n",
      "Epoch 169/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4165 - binary_accuracy: 0.8258\n",
      "Epoch 170/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4081 - binary_accuracy: 0.8296\n",
      "Epoch 171/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4098 - binary_accuracy: 0.8319\n",
      "Epoch 172/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4132 - binary_accuracy: 0.8291\n",
      "Epoch 173/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.4073 - binary_accuracy: 0.8289\n",
      "Epoch 174/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4132 - binary_accuracy: 0.8257\n",
      "Epoch 175/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4099 - binary_accuracy: 0.8308\n",
      "Epoch 176/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.3990 - binary_accuracy: 0.8317\n",
      "Epoch 177/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4050 - binary_accuracy: 0.8373\n",
      "Epoch 178/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4130 - binary_accuracy: 0.8279\n",
      "Epoch 179/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4135 - binary_accuracy: 0.8261\n",
      "Epoch 180/225\n",
      "1755/1755 [==============================] - 1s 387us/step - loss: 0.4094 - binary_accuracy: 0.8314\n",
      "Epoch 181/225\n",
      "1755/1755 [==============================] - 1s 386us/step - loss: 0.4079 - binary_accuracy: 0.8281\n",
      "Epoch 182/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.3993 - binary_accuracy: 0.8344\n",
      "Epoch 183/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4131 - binary_accuracy: 0.8262\n",
      "Epoch 184/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4103 - binary_accuracy: 0.8238\n",
      "Epoch 185/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4096 - binary_accuracy: 0.8309\n",
      "Epoch 186/225\n",
      "1755/1755 [==============================] - 1s 387us/step - loss: 0.4110 - binary_accuracy: 0.8306\n",
      "Epoch 187/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4106 - binary_accuracy: 0.8254\n",
      "Epoch 188/225\n",
      "1755/1755 [==============================] - 1s 388us/step - loss: 0.4029 - binary_accuracy: 0.8338\n",
      "Epoch 189/225\n",
      "1755/1755 [==============================] - 1s 387us/step - loss: 0.4077 - binary_accuracy: 0.8260\n",
      "Epoch 190/225\n",
      "1755/1755 [==============================] - 1s 386us/step - loss: 0.4053 - binary_accuracy: 0.8291\n",
      "Epoch 191/225\n",
      "1755/1755 [==============================] - 1s 387us/step - loss: 0.4141 - binary_accuracy: 0.8265\n",
      "Epoch 192/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4137 - binary_accuracy: 0.8298\n",
      "Epoch 193/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4125 - binary_accuracy: 0.8291\n",
      "Epoch 194/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4097 - binary_accuracy: 0.8308\n",
      "Epoch 195/225\n",
      "1755/1755 [==============================] - 1s 387us/step - loss: 0.4110 - binary_accuracy: 0.8283\n",
      "Epoch 196/225\n",
      "1755/1755 [==============================] - 1s 387us/step - loss: 0.4074 - binary_accuracy: 0.8289\n",
      "Epoch 197/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4177 - binary_accuracy: 0.8224\n",
      "Epoch 198/225\n",
      "1755/1755 [==============================] - 1s 406us/step - loss: 0.4079 - binary_accuracy: 0.8307\n",
      "Epoch 199/225\n",
      "1755/1755 [==============================] - 1s 411us/step - loss: 0.4039 - binary_accuracy: 0.8334\n",
      "Epoch 200/225\n",
      "1755/1755 [==============================] - 1s 388us/step - loss: 0.4118 - binary_accuracy: 0.8299\n",
      "Epoch 201/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4044 - binary_accuracy: 0.8312\n",
      "Epoch 202/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4101 - binary_accuracy: 0.8321\n",
      "Epoch 203/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4084 - binary_accuracy: 0.8301\n",
      "Epoch 204/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.3971 - binary_accuracy: 0.8373\n",
      "Epoch 205/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4072 - binary_accuracy: 0.8342\n",
      "Epoch 206/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4157 - binary_accuracy: 0.8246\n",
      "Epoch 207/225\n",
      "1755/1755 [==============================] - 1s 387us/step - loss: 0.4036 - binary_accuracy: 0.8318\n",
      "Epoch 208/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4101 - binary_accuracy: 0.8298\n",
      "Epoch 209/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4122 - binary_accuracy: 0.8279\n",
      "Epoch 210/225\n",
      "1755/1755 [==============================] - 1s 390us/step - loss: 0.4076 - binary_accuracy: 0.8322\n",
      "Epoch 211/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4066 - binary_accuracy: 0.8308\n",
      "Epoch 212/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4107 - binary_accuracy: 0.8304\n",
      "Epoch 213/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4072 - binary_accuracy: 0.8269\n",
      "Epoch 214/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4081 - binary_accuracy: 0.8293\n",
      "Epoch 215/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4106 - binary_accuracy: 0.8297\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 216/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4091 - binary_accuracy: 0.8310\n",
      "Epoch 217/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4063 - binary_accuracy: 0.8311\n",
      "Epoch 218/225\n",
      "1755/1755 [==============================] - 1s 386us/step - loss: 0.4116 - binary_accuracy: 0.8279\n",
      "Epoch 219/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4078 - binary_accuracy: 0.8301\n",
      "Epoch 220/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.3999 - binary_accuracy: 0.8340\n",
      "Epoch 221/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4108 - binary_accuracy: 0.8269\n",
      "Epoch 222/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4065 - binary_accuracy: 0.8290\n",
      "Epoch 223/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4104 - binary_accuracy: 0.8312\n",
      "Epoch 224/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4118 - binary_accuracy: 0.8241\n",
      "Epoch 225/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4045 - binary_accuracy: 0.8292\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/225\n",
      "1755/1755 [==============================] - 1s 774us/step - loss: 0.4823 - binary_accuracy: 0.8145\n",
      "Epoch 2/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4408 - binary_accuracy: 0.8193\n",
      "Epoch 3/225\n",
      "1755/1755 [==============================] - 1s 377us/step - loss: 0.4379 - binary_accuracy: 0.8215\n",
      "Epoch 4/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4375 - binary_accuracy: 0.8214\n",
      "Epoch 5/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4407 - binary_accuracy: 0.8236\n",
      "Epoch 6/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4289 - binary_accuracy: 0.8271\n",
      "Epoch 7/225\n",
      "1755/1755 [==============================] - 1s 377us/step - loss: 0.4345 - binary_accuracy: 0.8225\n",
      "Epoch 8/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4262 - binary_accuracy: 0.8264\n",
      "Epoch 9/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4224 - binary_accuracy: 0.8306\n",
      "Epoch 10/225\n",
      "1755/1755 [==============================] - 1s 375us/step - loss: 0.4298 - binary_accuracy: 0.8215\n",
      "Epoch 11/225\n",
      "1755/1755 [==============================] - 1s 376us/step - loss: 0.4205 - binary_accuracy: 0.8269\n",
      "Epoch 12/225\n",
      "1755/1755 [==============================] - 1s 376us/step - loss: 0.4196 - binary_accuracy: 0.8275\n",
      "Epoch 13/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4279 - binary_accuracy: 0.8255\n",
      "Epoch 14/225\n",
      "1755/1755 [==============================] - 1s 374us/step - loss: 0.4312 - binary_accuracy: 0.8226\n",
      "Epoch 15/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4230 - binary_accuracy: 0.8264\n",
      "Epoch 16/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4303 - binary_accuracy: 0.8263\n",
      "Epoch 17/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4249 - binary_accuracy: 0.8265\n",
      "Epoch 18/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4252 - binary_accuracy: 0.8236\n",
      "Epoch 19/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4192 - binary_accuracy: 0.8238\n",
      "Epoch 20/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4196 - binary_accuracy: 0.8253\n",
      "Epoch 21/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4170 - binary_accuracy: 0.8252\n",
      "Epoch 22/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4287 - binary_accuracy: 0.8217\n",
      "Epoch 23/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.4283 - binary_accuracy: 0.8251\n",
      "Epoch 24/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4185 - binary_accuracy: 0.8253\n",
      "Epoch 25/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4230 - binary_accuracy: 0.8257\n",
      "Epoch 26/225\n",
      "1755/1755 [==============================] - 1s 391us/step - loss: 0.4120 - binary_accuracy: 0.8317\n",
      "Epoch 27/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4231 - binary_accuracy: 0.8250\n",
      "Epoch 28/225\n",
      "1755/1755 [==============================] - 1s 390us/step - loss: 0.4220 - binary_accuracy: 0.8248\n",
      "Epoch 29/225\n",
      "1755/1755 [==============================] - 1s 406us/step - loss: 0.4219 - binary_accuracy: 0.8296\n",
      "Epoch 30/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4197 - binary_accuracy: 0.8259\n",
      "Epoch 31/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4207 - binary_accuracy: 0.8258\n",
      "Epoch 32/225\n",
      "1755/1755 [==============================] - 1s 386us/step - loss: 0.4264 - binary_accuracy: 0.8202\n",
      "Epoch 33/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4190 - binary_accuracy: 0.8282\n",
      "Epoch 34/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.4112 - binary_accuracy: 0.8327\n",
      "Epoch 35/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4217 - binary_accuracy: 0.8279\n",
      "Epoch 36/225\n",
      "1755/1755 [==============================] - 1s 377us/step - loss: 0.4235 - binary_accuracy: 0.8267\n",
      "Epoch 37/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.4154 - binary_accuracy: 0.8315\n",
      "Epoch 38/225\n",
      "1755/1755 [==============================] - 1s 386us/step - loss: 0.4188 - binary_accuracy: 0.8234\n",
      "Epoch 39/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4163 - binary_accuracy: 0.8251\n",
      "Epoch 40/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4115 - binary_accuracy: 0.8311\n",
      "Epoch 41/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4141 - binary_accuracy: 0.8276\n",
      "Epoch 42/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4214 - binary_accuracy: 0.8220\n",
      "Epoch 43/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.4119 - binary_accuracy: 0.8303\n",
      "Epoch 44/225\n",
      "1755/1755 [==============================] - 1s 377us/step - loss: 0.4091 - binary_accuracy: 0.8293\n",
      "Epoch 45/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4133 - binary_accuracy: 0.8317\n",
      "Epoch 46/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4118 - binary_accuracy: 0.8279\n",
      "Epoch 47/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4197 - binary_accuracy: 0.8257\n",
      "Epoch 48/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4134 - binary_accuracy: 0.8254\n",
      "Epoch 49/225\n",
      "1755/1755 [==============================] - 1s 377us/step - loss: 0.4093 - binary_accuracy: 0.8336\n",
      "Epoch 50/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4168 - binary_accuracy: 0.8270\n",
      "Epoch 51/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4162 - binary_accuracy: 0.8308\n",
      "Epoch 52/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4152 - binary_accuracy: 0.8284\n",
      "Epoch 53/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4135 - binary_accuracy: 0.8294\n",
      "Epoch 54/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.4187 - binary_accuracy: 0.8230\n",
      "Epoch 55/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4174 - binary_accuracy: 0.8269\n",
      "Epoch 56/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4125 - binary_accuracy: 0.8294\n",
      "Epoch 57/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4042 - binary_accuracy: 0.8337\n",
      "Epoch 58/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.4116 - binary_accuracy: 0.8285\n",
      "Epoch 59/225\n",
      "1755/1755 [==============================] - 1s 377us/step - loss: 0.4134 - binary_accuracy: 0.8250\n",
      "Epoch 60/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4127 - binary_accuracy: 0.8288\n",
      "Epoch 61/225\n",
      "1755/1755 [==============================] - 1s 387us/step - loss: 0.4171 - binary_accuracy: 0.8269\n",
      "Epoch 62/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4128 - binary_accuracy: 0.8323\n",
      "Epoch 63/225\n",
      "1755/1755 [==============================] - 1s 374us/step - loss: 0.4042 - binary_accuracy: 0.8352\n",
      "Epoch 64/225\n",
      "1755/1755 [==============================] - 1s 374us/step - loss: 0.4104 - binary_accuracy: 0.8319\n",
      "Epoch 65/225\n",
      "1755/1755 [==============================] - 1s 377us/step - loss: 0.4136 - binary_accuracy: 0.8251\n",
      "Epoch 66/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4162 - binary_accuracy: 0.8340\n",
      "Epoch 67/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4119 - binary_accuracy: 0.8274\n",
      "Epoch 68/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4135 - binary_accuracy: 0.8285\n",
      "Epoch 69/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.4194 - binary_accuracy: 0.8197\n",
      "Epoch 70/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4104 - binary_accuracy: 0.8325\n",
      "Epoch 71/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4072 - binary_accuracy: 0.8313\n",
      "Epoch 72/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4173 - binary_accuracy: 0.8285\n",
      "Epoch 73/225\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4073 - binary_accuracy: 0.8313\n",
      "Epoch 74/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4107 - binary_accuracy: 0.8301\n",
      "Epoch 75/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4088 - binary_accuracy: 0.8313\n",
      "Epoch 76/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4153 - binary_accuracy: 0.8280\n",
      "Epoch 77/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4065 - binary_accuracy: 0.8312\n",
      "Epoch 78/225\n",
      "1755/1755 [==============================] - 1s 391us/step - loss: 0.4147 - binary_accuracy: 0.8320\n",
      "Epoch 79/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4100 - binary_accuracy: 0.8358\n",
      "Epoch 80/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.4108 - binary_accuracy: 0.8316\n",
      "Epoch 81/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4050 - binary_accuracy: 0.8321\n",
      "Epoch 82/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4141 - binary_accuracy: 0.8253\n",
      "Epoch 83/225\n",
      "1755/1755 [==============================] - 1s 415us/step - loss: 0.4102 - binary_accuracy: 0.8269\n",
      "Epoch 84/225\n",
      "1755/1755 [==============================] - 1s 386us/step - loss: 0.4161 - binary_accuracy: 0.8234\n",
      "Epoch 85/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4136 - binary_accuracy: 0.8308\n",
      "Epoch 86/225\n",
      "1755/1755 [==============================] - 1s 374us/step - loss: 0.4136 - binary_accuracy: 0.8276\n",
      "Epoch 87/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4143 - binary_accuracy: 0.8278\n",
      "Epoch 88/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.4133 - binary_accuracy: 0.8257\n",
      "Epoch 89/225\n",
      "1755/1755 [==============================] - 1s 373us/step - loss: 0.4082 - binary_accuracy: 0.8330\n",
      "Epoch 90/225\n",
      "1755/1755 [==============================] - 1s 386us/step - loss: 0.4167 - binary_accuracy: 0.8268\n",
      "Epoch 91/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4060 - binary_accuracy: 0.8272\n",
      "Epoch 92/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.4124 - binary_accuracy: 0.8278\n",
      "Epoch 93/225\n",
      "1755/1755 [==============================] - 1s 376us/step - loss: 0.4132 - binary_accuracy: 0.8289\n",
      "Epoch 94/225\n",
      "1755/1755 [==============================] - 1s 377us/step - loss: 0.4074 - binary_accuracy: 0.8357\n",
      "Epoch 95/225\n",
      "1755/1755 [==============================] - 1s 372us/step - loss: 0.4111 - binary_accuracy: 0.8306\n",
      "Epoch 96/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.4119 - binary_accuracy: 0.8298\n",
      "Epoch 97/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4092 - binary_accuracy: 0.8288\n",
      "Epoch 98/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4094 - binary_accuracy: 0.8299\n",
      "Epoch 99/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4152 - binary_accuracy: 0.8275\n",
      "Epoch 100/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4074 - binary_accuracy: 0.8337\n",
      "Epoch 101/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4109 - binary_accuracy: 0.8292\n",
      "Epoch 102/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4077 - binary_accuracy: 0.8308\n",
      "Epoch 103/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4114 - binary_accuracy: 0.8260\n",
      "Epoch 104/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4122 - binary_accuracy: 0.8284\n",
      "Epoch 105/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4132 - binary_accuracy: 0.8307\n",
      "Epoch 106/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4147 - binary_accuracy: 0.8290\n",
      "Epoch 107/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4070 - binary_accuracy: 0.8350\n",
      "Epoch 108/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.4088 - binary_accuracy: 0.8312\n",
      "Epoch 109/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.4073 - binary_accuracy: 0.8357\n",
      "Epoch 110/225\n",
      "1755/1755 [==============================] - 1s 376us/step - loss: 0.4146 - binary_accuracy: 0.8301\n",
      "Epoch 111/225\n",
      "1755/1755 [==============================] - 1s 391us/step - loss: 0.4090 - binary_accuracy: 0.8282\n",
      "Epoch 112/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.4085 - binary_accuracy: 0.8312\n",
      "Epoch 113/225\n",
      "1755/1755 [==============================] - 1s 386us/step - loss: 0.4081 - binary_accuracy: 0.8347\n",
      "Epoch 114/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4103 - binary_accuracy: 0.8298\n",
      "Epoch 115/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4084 - binary_accuracy: 0.8328\n",
      "Epoch 116/225\n",
      "1755/1755 [==============================] - 1s 370us/step - loss: 0.4131 - binary_accuracy: 0.8308\n",
      "Epoch 117/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.4067 - binary_accuracy: 0.8277\n",
      "Epoch 118/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4066 - binary_accuracy: 0.8361\n",
      "Epoch 119/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.4107 - binary_accuracy: 0.8358\n",
      "Epoch 120/225\n",
      "1755/1755 [==============================] - 1s 388us/step - loss: 0.4110 - binary_accuracy: 0.8287\n",
      "Epoch 121/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4111 - binary_accuracy: 0.8261\n",
      "Epoch 122/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4101 - binary_accuracy: 0.8245\n",
      "Epoch 123/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.4121 - binary_accuracy: 0.8281\n",
      "Epoch 124/225\n",
      "1755/1755 [==============================] - 1s 394us/step - loss: 0.4090 - binary_accuracy: 0.8312\n",
      "Epoch 125/225\n",
      "1755/1755 [==============================] - 1s 409us/step - loss: 0.4110 - binary_accuracy: 0.8279\n",
      "Epoch 126/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4015 - binary_accuracy: 0.8316\n",
      "Epoch 127/225\n",
      "1755/1755 [==============================] - 1s 386us/step - loss: 0.4078 - binary_accuracy: 0.8298\n",
      "Epoch 128/225\n",
      "1755/1755 [==============================] - 1s 377us/step - loss: 0.4027 - binary_accuracy: 0.8313\n",
      "Epoch 129/225\n",
      "1755/1755 [==============================] - 1s 377us/step - loss: 0.3990 - binary_accuracy: 0.8346\n",
      "Epoch 130/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4056 - binary_accuracy: 0.8318\n",
      "Epoch 131/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4089 - binary_accuracy: 0.8327\n",
      "Epoch 132/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4021 - binary_accuracy: 0.8330\n",
      "Epoch 133/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.4026 - binary_accuracy: 0.8333\n",
      "Epoch 134/225\n",
      "1755/1755 [==============================] - 1s 377us/step - loss: 0.4066 - binary_accuracy: 0.8292\n",
      "Epoch 135/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4035 - binary_accuracy: 0.8306\n",
      "Epoch 136/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4128 - binary_accuracy: 0.8281\n",
      "Epoch 137/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.4040 - binary_accuracy: 0.8344\n",
      "Epoch 138/225\n",
      "1755/1755 [==============================] - 1s 376us/step - loss: 0.4131 - binary_accuracy: 0.8306\n",
      "Epoch 139/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4087 - binary_accuracy: 0.8295\n",
      "Epoch 140/225\n",
      "1755/1755 [==============================] - 1s 389us/step - loss: 0.4073 - binary_accuracy: 0.8311\n",
      "Epoch 141/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4103 - binary_accuracy: 0.8341\n",
      "Epoch 142/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4082 - binary_accuracy: 0.8313\n",
      "Epoch 143/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4038 - binary_accuracy: 0.8332\n",
      "Epoch 144/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4054 - binary_accuracy: 0.8303\n",
      "Epoch 145/225\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4035 - binary_accuracy: 0.8309\n",
      "Epoch 146/225\n",
      "1755/1755 [==============================] - 1s 375us/step - loss: 0.4079 - binary_accuracy: 0.8359\n",
      "Epoch 147/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4151 - binary_accuracy: 0.8268\n",
      "Epoch 148/225\n",
      "1755/1755 [==============================] - 1s 377us/step - loss: 0.4086 - binary_accuracy: 0.8325\n",
      "Epoch 149/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4022 - binary_accuracy: 0.8327\n",
      "Epoch 150/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4087 - binary_accuracy: 0.8312\n",
      "Epoch 151/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4100 - binary_accuracy: 0.8273\n",
      "Epoch 152/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.4117 - binary_accuracy: 0.8282\n",
      "Epoch 153/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4086 - binary_accuracy: 0.8340\n",
      "Epoch 154/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4064 - binary_accuracy: 0.8303\n",
      "Epoch 155/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4113 - binary_accuracy: 0.8297\n",
      "Epoch 156/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4016 - binary_accuracy: 0.8367\n",
      "Epoch 157/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4066 - binary_accuracy: 0.8315\n",
      "Epoch 158/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4059 - binary_accuracy: 0.8313\n",
      "Epoch 159/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.3999 - binary_accuracy: 0.8353\n",
      "Epoch 160/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4123 - binary_accuracy: 0.8264\n",
      "Epoch 161/225\n",
      "1755/1755 [==============================] - 1s 391us/step - loss: 0.4019 - binary_accuracy: 0.8353\n",
      "Epoch 162/225\n",
      "1755/1755 [==============================] - 1s 415us/step - loss: 0.4032 - binary_accuracy: 0.8333\n",
      "Epoch 163/225\n",
      "1755/1755 [==============================] - 1s 375us/step - loss: 0.4051 - binary_accuracy: 0.8318\n",
      "Epoch 164/225\n",
      "1755/1755 [==============================] - 1s 377us/step - loss: 0.4040 - binary_accuracy: 0.8328\n",
      "Epoch 165/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4058 - binary_accuracy: 0.8287\n",
      "Epoch 166/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.4022 - binary_accuracy: 0.8314\n",
      "Epoch 167/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4019 - binary_accuracy: 0.8284\n",
      "Epoch 168/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4097 - binary_accuracy: 0.8280\n",
      "Epoch 169/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.3989 - binary_accuracy: 0.8334\n",
      "Epoch 170/225\n",
      "1755/1755 [==============================] - 1s 377us/step - loss: 0.3963 - binary_accuracy: 0.8356\n",
      "Epoch 171/225\n",
      "1755/1755 [==============================] - 1s 376us/step - loss: 0.4103 - binary_accuracy: 0.8279\n",
      "Epoch 172/225\n",
      "1755/1755 [==============================] - 1s 373us/step - loss: 0.4020 - binary_accuracy: 0.8364\n",
      "Epoch 173/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4003 - binary_accuracy: 0.8309\n",
      "Epoch 174/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4058 - binary_accuracy: 0.8294\n",
      "Epoch 175/225\n",
      "1755/1755 [==============================] - 1s 390us/step - loss: 0.4100 - binary_accuracy: 0.8299\n",
      "Epoch 176/225\n",
      "1755/1755 [==============================] - 1s 386us/step - loss: 0.4016 - binary_accuracy: 0.8323\n",
      "Epoch 177/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4116 - binary_accuracy: 0.8311\n",
      "Epoch 178/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4067 - binary_accuracy: 0.8301\n",
      "Epoch 179/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4063 - binary_accuracy: 0.8365\n",
      "Epoch 180/225\n",
      "1755/1755 [==============================] - 1s 377us/step - loss: 0.4060 - binary_accuracy: 0.8317\n",
      "Epoch 181/225\n",
      "1755/1755 [==============================] - 1s 377us/step - loss: 0.4076 - binary_accuracy: 0.8364\n",
      "Epoch 182/225\n",
      "1755/1755 [==============================] - 1s 377us/step - loss: 0.4044 - binary_accuracy: 0.8314\n",
      "Epoch 183/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.4060 - binary_accuracy: 0.8311\n",
      "Epoch 184/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4111 - binary_accuracy: 0.8297\n",
      "Epoch 185/225\n",
      "1755/1755 [==============================] - 1s 374us/step - loss: 0.3983 - binary_accuracy: 0.8346\n",
      "Epoch 186/225\n",
      "1755/1755 [==============================] - 1s 373us/step - loss: 0.4092 - binary_accuracy: 0.8274\n",
      "Epoch 187/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.4154 - binary_accuracy: 0.8297\n",
      "Epoch 188/225\n",
      "1755/1755 [==============================] - 1s 377us/step - loss: 0.4082 - binary_accuracy: 0.8321\n",
      "Epoch 189/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4055 - binary_accuracy: 0.8314\n",
      "Epoch 190/225\n",
      "1755/1755 [==============================] - 1s 387us/step - loss: 0.4000 - binary_accuracy: 0.8374\n",
      "Epoch 191/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4095 - binary_accuracy: 0.8298\n",
      "Epoch 192/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4017 - binary_accuracy: 0.8344\n",
      "Epoch 193/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.4034 - binary_accuracy: 0.8324\n",
      "Epoch 194/225\n",
      "1755/1755 [==============================] - 1s 386us/step - loss: 0.4055 - binary_accuracy: 0.8337\n",
      "Epoch 195/225\n",
      "1755/1755 [==============================] - 1s 377us/step - loss: 0.4073 - binary_accuracy: 0.8288\n",
      "Epoch 196/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4039 - binary_accuracy: 0.8316\n",
      "Epoch 197/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4042 - binary_accuracy: 0.8292\n",
      "Epoch 198/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4061 - binary_accuracy: 0.8293\n",
      "Epoch 199/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.4048 - binary_accuracy: 0.8320\n",
      "Epoch 200/225\n",
      "1755/1755 [==============================] - 1s 401us/step - loss: 0.4027 - binary_accuracy: 0.8372\n",
      "Epoch 201/225\n",
      "1755/1755 [==============================] - 1s 380us/step - loss: 0.3946 - binary_accuracy: 0.8380\n",
      "Epoch 202/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4077 - binary_accuracy: 0.8283\n",
      "Epoch 203/225\n",
      "1755/1755 [==============================] - 1s 384us/step - loss: 0.4074 - binary_accuracy: 0.8354\n",
      "Epoch 204/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4093 - binary_accuracy: 0.8303\n",
      "Epoch 205/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4009 - binary_accuracy: 0.8303\n",
      "Epoch 206/225\n",
      "1755/1755 [==============================] - 1s 383us/step - loss: 0.4038 - binary_accuracy: 0.8328\n",
      "Epoch 207/225\n",
      "1755/1755 [==============================] - 1s 377us/step - loss: 0.4065 - binary_accuracy: 0.8275\n",
      "Epoch 208/225\n",
      "1755/1755 [==============================] - 1s 376us/step - loss: 0.4023 - binary_accuracy: 0.8347\n",
      "Epoch 209/225\n",
      "1755/1755 [==============================] - 1s 376us/step - loss: 0.4040 - binary_accuracy: 0.8355\n",
      "Epoch 210/225\n",
      "1755/1755 [==============================] - 1s 382us/step - loss: 0.4021 - binary_accuracy: 0.8324\n",
      "Epoch 211/225\n",
      "1755/1755 [==============================] - 1s 377us/step - loss: 0.4049 - binary_accuracy: 0.8311\n",
      "Epoch 212/225\n",
      "1755/1755 [==============================] - 1s 381us/step - loss: 0.4004 - binary_accuracy: 0.8325\n",
      "Epoch 213/225\n",
      "1755/1755 [==============================] - 1s 376us/step - loss: 0.3967 - binary_accuracy: 0.8326\n",
      "Epoch 214/225\n",
      "1755/1755 [==============================] - 1s 376us/step - loss: 0.4044 - binary_accuracy: 0.8331\n",
      "Epoch 215/225\n",
      "1755/1755 [==============================] - 1s 385us/step - loss: 0.4014 - binary_accuracy: 0.8323\n",
      "Epoch 216/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.4087 - binary_accuracy: 0.8343\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 217/225\n",
      "1755/1755 [==============================] - 1s 376us/step - loss: 0.4029 - binary_accuracy: 0.8299\n",
      "Epoch 218/225\n",
      "1755/1755 [==============================] - 1s 375us/step - loss: 0.4080 - binary_accuracy: 0.8273\n",
      "Epoch 219/225\n",
      "1755/1755 [==============================] - 1s 374us/step - loss: 0.4022 - binary_accuracy: 0.8364\n",
      "Epoch 220/225\n",
      "1755/1755 [==============================] - 1s 387us/step - loss: 0.4053 - binary_accuracy: 0.8343\n",
      "Epoch 221/225\n",
      "1755/1755 [==============================] - 1s 386us/step - loss: 0.4000 - binary_accuracy: 0.8308\n",
      "Epoch 222/225\n",
      "1755/1755 [==============================] - 1s 379us/step - loss: 0.4027 - binary_accuracy: 0.8326\n",
      "Epoch 223/225\n",
      "1755/1755 [==============================] - 1s 373us/step - loss: 0.4042 - binary_accuracy: 0.8274\n",
      "Epoch 224/225\n",
      "1755/1755 [==============================] - 1s 378us/step - loss: 0.3972 - binary_accuracy: 0.8377\n",
      "Epoch 225/225\n",
      "1755/1755 [==============================] - 1s 376us/step - loss: 0.4024 - binary_accuracy: 0.8327\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/225\n",
      "1754/1754 [==============================] - 1s 829us/step - loss: 0.4754 - binary_accuracy: 0.8126\n",
      "Epoch 2/225\n",
      "1754/1754 [==============================] - 1s 391us/step - loss: 0.4408 - binary_accuracy: 0.8247\n",
      "Epoch 3/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4391 - binary_accuracy: 0.8198\n",
      "Epoch 4/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4279 - binary_accuracy: 0.8296\n",
      "Epoch 5/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4319 - binary_accuracy: 0.8249\n",
      "Epoch 6/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4257 - binary_accuracy: 0.8266\n",
      "Epoch 7/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4253 - binary_accuracy: 0.8232\n",
      "Epoch 8/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4283 - binary_accuracy: 0.8217\n",
      "Epoch 9/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4227 - binary_accuracy: 0.8236\n",
      "Epoch 10/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4296 - binary_accuracy: 0.8249\n",
      "Epoch 11/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4188 - binary_accuracy: 0.8273\n",
      "Epoch 12/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4270 - binary_accuracy: 0.8276\n",
      "Epoch 13/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4169 - binary_accuracy: 0.8298\n",
      "Epoch 14/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4167 - binary_accuracy: 0.8297\n",
      "Epoch 15/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4198 - binary_accuracy: 0.8263\n",
      "Epoch 16/225\n",
      "1754/1754 [==============================] - 1s 376us/step - loss: 0.4277 - binary_accuracy: 0.8243\n",
      "Epoch 17/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4108 - binary_accuracy: 0.8334\n",
      "Epoch 18/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4199 - binary_accuracy: 0.8273\n",
      "Epoch 19/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4143 - binary_accuracy: 0.8309\n",
      "Epoch 20/225\n",
      "1754/1754 [==============================] - 1s 385us/step - loss: 0.4169 - binary_accuracy: 0.8281\n",
      "Epoch 21/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4182 - binary_accuracy: 0.8270\n",
      "Epoch 22/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4176 - binary_accuracy: 0.8254\n",
      "Epoch 23/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4222 - binary_accuracy: 0.8211\n",
      "Epoch 24/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4247 - binary_accuracy: 0.8231\n",
      "Epoch 25/225\n",
      "1754/1754 [==============================] - 1s 376us/step - loss: 0.4142 - binary_accuracy: 0.8304\n",
      "Epoch 26/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4140 - binary_accuracy: 0.8291\n",
      "Epoch 27/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4164 - binary_accuracy: 0.8317\n",
      "Epoch 28/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4166 - binary_accuracy: 0.8291\n",
      "Epoch 29/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4149 - binary_accuracy: 0.8244\n",
      "Epoch 30/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4106 - binary_accuracy: 0.8293\n",
      "Epoch 31/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4139 - binary_accuracy: 0.8266\n",
      "Epoch 32/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4234 - binary_accuracy: 0.8261\n",
      "Epoch 33/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4175 - binary_accuracy: 0.8269\n",
      "Epoch 34/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4141 - binary_accuracy: 0.8247\n",
      "Epoch 35/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4169 - binary_accuracy: 0.8322\n",
      "Epoch 36/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4165 - binary_accuracy: 0.8257\n",
      "Epoch 37/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4118 - binary_accuracy: 0.8290\n",
      "Epoch 38/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4118 - binary_accuracy: 0.8316\n",
      "Epoch 39/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.4184 - binary_accuracy: 0.8240\n",
      "Epoch 40/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4080 - binary_accuracy: 0.8331\n",
      "Epoch 41/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4115 - binary_accuracy: 0.8309\n",
      "Epoch 42/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4153 - binary_accuracy: 0.8276\n",
      "Epoch 43/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4159 - binary_accuracy: 0.8264\n",
      "Epoch 44/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4153 - binary_accuracy: 0.8271\n",
      "Epoch 45/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4132 - binary_accuracy: 0.8295\n",
      "Epoch 46/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4099 - binary_accuracy: 0.8300\n",
      "Epoch 47/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4085 - binary_accuracy: 0.8282\n",
      "Epoch 48/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4155 - binary_accuracy: 0.8292\n",
      "Epoch 49/225\n",
      "1754/1754 [==============================] - 1s 387us/step - loss: 0.4114 - binary_accuracy: 0.8296\n",
      "Epoch 50/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4121 - binary_accuracy: 0.8266\n",
      "Epoch 51/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4123 - binary_accuracy: 0.8291\n",
      "Epoch 52/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4117 - binary_accuracy: 0.8258\n",
      "Epoch 53/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4080 - binary_accuracy: 0.8336\n",
      "Epoch 54/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4124 - binary_accuracy: 0.8285\n",
      "Epoch 55/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4048 - binary_accuracy: 0.8290\n",
      "Epoch 56/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.4095 - binary_accuracy: 0.8298\n",
      "Epoch 57/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4088 - binary_accuracy: 0.8304\n",
      "Epoch 58/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4134 - binary_accuracy: 0.8310\n",
      "Epoch 59/225\n",
      "1754/1754 [==============================] - 1s 408us/step - loss: 0.4077 - binary_accuracy: 0.8323\n",
      "Epoch 60/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4125 - binary_accuracy: 0.8253\n",
      "Epoch 61/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4058 - binary_accuracy: 0.8336\n",
      "Epoch 62/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4074 - binary_accuracy: 0.8314\n",
      "Epoch 63/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4044 - binary_accuracy: 0.8296\n",
      "Epoch 64/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4048 - binary_accuracy: 0.8280\n",
      "Epoch 65/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4163 - binary_accuracy: 0.8256\n",
      "Epoch 66/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4170 - binary_accuracy: 0.8288\n",
      "Epoch 67/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4087 - binary_accuracy: 0.8313\n",
      "Epoch 68/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4148 - binary_accuracy: 0.8324\n",
      "Epoch 69/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4143 - binary_accuracy: 0.8303\n",
      "Epoch 70/225\n",
      "1754/1754 [==============================] - 1s 385us/step - loss: 0.4150 - binary_accuracy: 0.8295\n",
      "Epoch 71/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4105 - binary_accuracy: 0.8353\n",
      "Epoch 72/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4089 - binary_accuracy: 0.8280\n",
      "Epoch 73/225\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4117 - binary_accuracy: 0.8305\n",
      "Epoch 74/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4150 - binary_accuracy: 0.8272\n",
      "Epoch 75/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4088 - binary_accuracy: 0.8295\n",
      "Epoch 76/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4154 - binary_accuracy: 0.8258\n",
      "Epoch 77/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4088 - binary_accuracy: 0.8297\n",
      "Epoch 78/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4142 - binary_accuracy: 0.8271\n",
      "Epoch 79/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4092 - binary_accuracy: 0.8325\n",
      "Epoch 80/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4120 - binary_accuracy: 0.8281\n",
      "Epoch 81/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4094 - binary_accuracy: 0.8295\n",
      "Epoch 82/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4090 - binary_accuracy: 0.8279\n",
      "Epoch 83/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4060 - binary_accuracy: 0.8304\n",
      "Epoch 84/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4050 - binary_accuracy: 0.8362\n",
      "Epoch 85/225\n",
      "1754/1754 [==============================] - 1s 388us/step - loss: 0.4162 - binary_accuracy: 0.8244\n",
      "Epoch 86/225\n",
      "1754/1754 [==============================] - 1s 387us/step - loss: 0.4126 - binary_accuracy: 0.8222\n",
      "Epoch 87/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4038 - binary_accuracy: 0.8319\n",
      "Epoch 88/225\n",
      "1754/1754 [==============================] - 1s 385us/step - loss: 0.4071 - binary_accuracy: 0.8322\n",
      "Epoch 89/225\n",
      "1754/1754 [==============================] - 1s 388us/step - loss: 0.4059 - binary_accuracy: 0.8297\n",
      "Epoch 90/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4096 - binary_accuracy: 0.8299\n",
      "Epoch 91/225\n",
      "1754/1754 [==============================] - 1s 389us/step - loss: 0.4067 - binary_accuracy: 0.8308\n",
      "Epoch 92/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.3987 - binary_accuracy: 0.8346\n",
      "Epoch 93/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4077 - binary_accuracy: 0.8327\n",
      "Epoch 94/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4130 - binary_accuracy: 0.8282\n",
      "Epoch 95/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.4100 - binary_accuracy: 0.8298\n",
      "Epoch 96/225\n",
      "1754/1754 [==============================] - 1s 387us/step - loss: 0.4080 - binary_accuracy: 0.8324\n",
      "Epoch 97/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4039 - binary_accuracy: 0.8343\n",
      "Epoch 98/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4061 - binary_accuracy: 0.8292\n",
      "Epoch 99/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4018 - binary_accuracy: 0.8319\n",
      "Epoch 100/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4043 - binary_accuracy: 0.8335\n",
      "Epoch 101/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4041 - binary_accuracy: 0.8295\n",
      "Epoch 102/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4095 - binary_accuracy: 0.8293\n",
      "Epoch 103/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4061 - binary_accuracy: 0.8274\n",
      "Epoch 104/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4058 - binary_accuracy: 0.8313\n",
      "Epoch 105/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4035 - binary_accuracy: 0.8306\n",
      "Epoch 106/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4109 - binary_accuracy: 0.8269\n",
      "Epoch 107/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4059 - binary_accuracy: 0.8314\n",
      "Epoch 108/225\n",
      "1754/1754 [==============================] - 1s 385us/step - loss: 0.4027 - binary_accuracy: 0.8371\n",
      "Epoch 109/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4092 - binary_accuracy: 0.8307\n",
      "Epoch 110/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4033 - binary_accuracy: 0.8291\n",
      "Epoch 111/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4048 - binary_accuracy: 0.8343\n",
      "Epoch 112/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4075 - binary_accuracy: 0.8295\n",
      "Epoch 113/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4102 - binary_accuracy: 0.8302\n",
      "Epoch 114/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4017 - binary_accuracy: 0.8311\n",
      "Epoch 115/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4128 - binary_accuracy: 0.8279\n",
      "Epoch 116/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4114 - binary_accuracy: 0.8304\n",
      "Epoch 117/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4065 - binary_accuracy: 0.8301\n",
      "Epoch 118/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4042 - binary_accuracy: 0.8264\n",
      "Epoch 119/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4080 - binary_accuracy: 0.8297\n",
      "Epoch 120/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4152 - binary_accuracy: 0.8253\n",
      "Epoch 121/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4053 - binary_accuracy: 0.8305\n",
      "Epoch 122/225\n",
      "1754/1754 [==============================] - 1s 389us/step - loss: 0.4056 - binary_accuracy: 0.8254\n",
      "Epoch 123/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4084 - binary_accuracy: 0.8281\n",
      "Epoch 124/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.3978 - binary_accuracy: 0.8325\n",
      "Epoch 125/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4073 - binary_accuracy: 0.8280\n",
      "Epoch 126/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.3996 - binary_accuracy: 0.8368\n",
      "Epoch 127/225\n",
      "1754/1754 [==============================] - 1s 387us/step - loss: 0.4085 - binary_accuracy: 0.8299\n",
      "Epoch 128/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4062 - binary_accuracy: 0.8306\n",
      "Epoch 129/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4092 - binary_accuracy: 0.8256\n",
      "Epoch 130/225\n",
      "1754/1754 [==============================] - 1s 375us/step - loss: 0.4022 - binary_accuracy: 0.8341\n",
      "Epoch 131/225\n",
      "1754/1754 [==============================] - 1s 390us/step - loss: 0.4097 - binary_accuracy: 0.8289\n",
      "Epoch 132/225\n",
      "1754/1754 [==============================] - 1s 388us/step - loss: 0.3982 - binary_accuracy: 0.8367\n",
      "Epoch 133/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4018 - binary_accuracy: 0.8312\n",
      "Epoch 134/225\n",
      "1754/1754 [==============================] - 1s 385us/step - loss: 0.4042 - binary_accuracy: 0.8328\n",
      "Epoch 135/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4053 - binary_accuracy: 0.8321\n",
      "Epoch 136/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4052 - binary_accuracy: 0.8333\n",
      "Epoch 137/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4086 - binary_accuracy: 0.8298\n",
      "Epoch 138/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4118 - binary_accuracy: 0.8266\n",
      "Epoch 139/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4059 - binary_accuracy: 0.8286\n",
      "Epoch 140/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4078 - binary_accuracy: 0.8313\n",
      "Epoch 141/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4076 - binary_accuracy: 0.8304\n",
      "Epoch 142/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4103 - binary_accuracy: 0.8248\n",
      "Epoch 143/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4089 - binary_accuracy: 0.8270\n",
      "Epoch 144/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4070 - binary_accuracy: 0.8257\n",
      "Epoch 145/225\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4005 - binary_accuracy: 0.8334\n",
      "Epoch 146/225\n",
      "1754/1754 [==============================] - 1s 405us/step - loss: 0.4062 - binary_accuracy: 0.8286\n",
      "Epoch 147/225\n",
      "1754/1754 [==============================] - 1s 417us/step - loss: 0.4030 - binary_accuracy: 0.8334\n",
      "Epoch 148/225\n",
      "1754/1754 [==============================] - 1s 376us/step - loss: 0.4066 - binary_accuracy: 0.8286\n",
      "Epoch 149/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4056 - binary_accuracy: 0.8287\n",
      "Epoch 150/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.4068 - binary_accuracy: 0.8296\n",
      "Epoch 151/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.4054 - binary_accuracy: 0.8333\n",
      "Epoch 152/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4082 - binary_accuracy: 0.8347\n",
      "Epoch 153/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4023 - binary_accuracy: 0.8307\n",
      "Epoch 154/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4035 - binary_accuracy: 0.8303\n",
      "Epoch 155/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.3968 - binary_accuracy: 0.8361\n",
      "Epoch 156/225\n",
      "1754/1754 [==============================] - 1s 376us/step - loss: 0.4010 - binary_accuracy: 0.8315\n",
      "Epoch 157/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4005 - binary_accuracy: 0.8363\n",
      "Epoch 158/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.4040 - binary_accuracy: 0.8313\n",
      "Epoch 159/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4023 - binary_accuracy: 0.8326\n",
      "Epoch 160/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4055 - binary_accuracy: 0.8348\n",
      "Epoch 161/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4081 - binary_accuracy: 0.8317\n",
      "Epoch 162/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4028 - binary_accuracy: 0.8345\n",
      "Epoch 163/225\n",
      "1754/1754 [==============================] - 1s 385us/step - loss: 0.4014 - binary_accuracy: 0.8347\n",
      "Epoch 164/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.3986 - binary_accuracy: 0.8362\n",
      "Epoch 165/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.3991 - binary_accuracy: 0.8353\n",
      "Epoch 166/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4016 - binary_accuracy: 0.8355\n",
      "Epoch 167/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.3985 - binary_accuracy: 0.8336\n",
      "Epoch 168/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.3948 - binary_accuracy: 0.8398\n",
      "Epoch 169/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4009 - binary_accuracy: 0.8347\n",
      "Epoch 170/225\n",
      "1754/1754 [==============================] - 1s 389us/step - loss: 0.3901 - binary_accuracy: 0.8329\n",
      "Epoch 171/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4003 - binary_accuracy: 0.8318\n",
      "Epoch 172/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4035 - binary_accuracy: 0.8365\n",
      "Epoch 173/225\n",
      "1754/1754 [==============================] - 1s 385us/step - loss: 0.4053 - binary_accuracy: 0.8299\n",
      "Epoch 174/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.3976 - binary_accuracy: 0.8374\n",
      "Epoch 175/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.3996 - binary_accuracy: 0.8368\n",
      "Epoch 176/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4035 - binary_accuracy: 0.8303\n",
      "Epoch 177/225\n",
      "1754/1754 [==============================] - 1s 387us/step - loss: 0.4026 - binary_accuracy: 0.8345\n",
      "Epoch 178/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4056 - binary_accuracy: 0.8327\n",
      "Epoch 179/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4045 - binary_accuracy: 0.8291\n",
      "Epoch 180/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4003 - binary_accuracy: 0.8339\n",
      "Epoch 181/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.3991 - binary_accuracy: 0.8374\n",
      "Epoch 182/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.3968 - binary_accuracy: 0.8331\n",
      "Epoch 183/225\n",
      "1754/1754 [==============================] - 1s 390us/step - loss: 0.3958 - binary_accuracy: 0.8339\n",
      "Epoch 184/225\n",
      "1754/1754 [==============================] - 1s 388us/step - loss: 0.4046 - binary_accuracy: 0.8304\n",
      "Epoch 185/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.3985 - binary_accuracy: 0.8365\n",
      "Epoch 186/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4054 - binary_accuracy: 0.8337\n",
      "Epoch 187/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4031 - binary_accuracy: 0.8351\n",
      "Epoch 188/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.3945 - binary_accuracy: 0.8355\n",
      "Epoch 189/225\n",
      "1754/1754 [==============================] - 1s 388us/step - loss: 0.4111 - binary_accuracy: 0.8244\n",
      "Epoch 190/225\n",
      "1754/1754 [==============================] - 1s 412us/step - loss: 0.4000 - binary_accuracy: 0.8293\n",
      "Epoch 191/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4081 - binary_accuracy: 0.8284\n",
      "Epoch 192/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4078 - binary_accuracy: 0.8285\n",
      "Epoch 193/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4177 - binary_accuracy: 0.8286\n",
      "Epoch 194/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.3998 - binary_accuracy: 0.8329\n",
      "Epoch 195/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4035 - binary_accuracy: 0.8293\n",
      "Epoch 196/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4044 - binary_accuracy: 0.8316\n",
      "Epoch 197/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4094 - binary_accuracy: 0.8289\n",
      "Epoch 198/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4015 - binary_accuracy: 0.8297\n",
      "Epoch 199/225\n",
      "1754/1754 [==============================] - 1s 385us/step - loss: 0.4093 - binary_accuracy: 0.8266\n",
      "Epoch 200/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.3955 - binary_accuracy: 0.8309\n",
      "Epoch 201/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4061 - binary_accuracy: 0.8313\n",
      "Epoch 202/225\n",
      "1754/1754 [==============================] - 1s 385us/step - loss: 0.4024 - binary_accuracy: 0.8347\n",
      "Epoch 203/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4006 - binary_accuracy: 0.8371\n",
      "Epoch 204/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4012 - binary_accuracy: 0.8368\n",
      "Epoch 205/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.3942 - binary_accuracy: 0.8382\n",
      "Epoch 206/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.3995 - binary_accuracy: 0.8358\n",
      "Epoch 207/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.3975 - binary_accuracy: 0.8379\n",
      "Epoch 208/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.3929 - binary_accuracy: 0.8355\n",
      "Epoch 209/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.3963 - binary_accuracy: 0.8330\n",
      "Epoch 210/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.4051 - binary_accuracy: 0.8270\n",
      "Epoch 211/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4050 - binary_accuracy: 0.8315\n",
      "Epoch 212/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.3963 - binary_accuracy: 0.8309\n",
      "Epoch 213/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4006 - binary_accuracy: 0.8340\n",
      "Epoch 214/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.3996 - binary_accuracy: 0.8353\n",
      "Epoch 215/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.3997 - binary_accuracy: 0.8312\n",
      "Epoch 216/225\n",
      "1754/1754 [==============================] - 1s 375us/step - loss: 0.4038 - binary_accuracy: 0.8314\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 217/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4053 - binary_accuracy: 0.8294\n",
      "Epoch 218/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4039 - binary_accuracy: 0.8317\n",
      "Epoch 219/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.3940 - binary_accuracy: 0.8358\n",
      "Epoch 220/225\n",
      "1754/1754 [==============================] - 1s 375us/step - loss: 0.4002 - binary_accuracy: 0.8365\n",
      "Epoch 221/225\n",
      "1754/1754 [==============================] - 1s 376us/step - loss: 0.3948 - binary_accuracy: 0.8399\n",
      "Epoch 222/225\n",
      "1754/1754 [==============================] - 1s 376us/step - loss: 0.3989 - binary_accuracy: 0.8314\n",
      "Epoch 223/225\n",
      "1754/1754 [==============================] - 1s 387us/step - loss: 0.4029 - binary_accuracy: 0.8337\n",
      "Epoch 224/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4036 - binary_accuracy: 0.8289\n",
      "Epoch 225/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4017 - binary_accuracy: 0.8297\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/225\n",
      "1754/1754 [==============================] - 2s 863us/step - loss: 0.4886 - binary_accuracy: 0.8101\n",
      "Epoch 2/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4550 - binary_accuracy: 0.8183\n",
      "Epoch 3/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4482 - binary_accuracy: 0.8168\n",
      "Epoch 4/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4410 - binary_accuracy: 0.8241\n",
      "Epoch 5/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4413 - binary_accuracy: 0.8189\n",
      "Epoch 6/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4329 - binary_accuracy: 0.8174\n",
      "Epoch 7/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4209 - binary_accuracy: 0.8256\n",
      "Epoch 8/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4247 - binary_accuracy: 0.8254\n",
      "Epoch 9/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4326 - binary_accuracy: 0.8241\n",
      "Epoch 10/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.4282 - binary_accuracy: 0.8269\n",
      "Epoch 11/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4320 - binary_accuracy: 0.8184\n",
      "Epoch 12/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4224 - binary_accuracy: 0.8272\n",
      "Epoch 13/225\n",
      "1754/1754 [==============================] - 1s 385us/step - loss: 0.4363 - binary_accuracy: 0.8224\n",
      "Epoch 14/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4251 - binary_accuracy: 0.8226\n",
      "Epoch 15/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4282 - binary_accuracy: 0.8232\n",
      "Epoch 16/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4318 - binary_accuracy: 0.8230\n",
      "Epoch 17/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.4244 - binary_accuracy: 0.8202\n",
      "Epoch 18/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4334 - binary_accuracy: 0.8212\n",
      "Epoch 19/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4306 - binary_accuracy: 0.8218\n",
      "Epoch 20/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.4286 - binary_accuracy: 0.8204\n",
      "Epoch 21/225\n",
      "1754/1754 [==============================] - 1s 376us/step - loss: 0.4250 - binary_accuracy: 0.8259\n",
      "Epoch 22/225\n",
      "1754/1754 [==============================] - 1s 385us/step - loss: 0.4274 - binary_accuracy: 0.8264\n",
      "Epoch 23/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4273 - binary_accuracy: 0.8206\n",
      "Epoch 24/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4248 - binary_accuracy: 0.8212\n",
      "Epoch 25/225\n",
      "1754/1754 [==============================] - 1s 404us/step - loss: 0.4225 - binary_accuracy: 0.8236\n",
      "Epoch 26/225\n",
      "1754/1754 [==============================] - 1s 407us/step - loss: 0.4241 - binary_accuracy: 0.8235\n",
      "Epoch 27/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4194 - binary_accuracy: 0.8249\n",
      "Epoch 28/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4236 - binary_accuracy: 0.8217\n",
      "Epoch 29/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4244 - binary_accuracy: 0.8239\n",
      "Epoch 30/225\n",
      "1754/1754 [==============================] - 1s 385us/step - loss: 0.4220 - binary_accuracy: 0.8246\n",
      "Epoch 31/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4192 - binary_accuracy: 0.8270\n",
      "Epoch 32/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4144 - binary_accuracy: 0.8314\n",
      "Epoch 33/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4214 - binary_accuracy: 0.8257\n",
      "Epoch 34/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4174 - binary_accuracy: 0.8239\n",
      "Epoch 35/225\n",
      "1754/1754 [==============================] - 1s 389us/step - loss: 0.4209 - binary_accuracy: 0.8248\n",
      "Epoch 36/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4313 - binary_accuracy: 0.8237\n",
      "Epoch 37/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4247 - binary_accuracy: 0.8212\n",
      "Epoch 38/225\n",
      "1754/1754 [==============================] - 1s 388us/step - loss: 0.4216 - binary_accuracy: 0.8246\n",
      "Epoch 39/225\n",
      "1754/1754 [==============================] - 1s 387us/step - loss: 0.4213 - binary_accuracy: 0.8207\n",
      "Epoch 40/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4220 - binary_accuracy: 0.8233\n",
      "Epoch 41/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4168 - binary_accuracy: 0.8272\n",
      "Epoch 42/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4205 - binary_accuracy: 0.8238\n",
      "Epoch 43/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4254 - binary_accuracy: 0.8204\n",
      "Epoch 44/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4306 - binary_accuracy: 0.8204\n",
      "Epoch 45/225\n",
      "1754/1754 [==============================] - 1s 385us/step - loss: 0.4283 - binary_accuracy: 0.8252\n",
      "Epoch 46/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4333 - binary_accuracy: 0.8170\n",
      "Epoch 47/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4224 - binary_accuracy: 0.8300\n",
      "Epoch 48/225\n",
      "1754/1754 [==============================] - 1s 374us/step - loss: 0.4152 - binary_accuracy: 0.8286\n",
      "Epoch 49/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4194 - binary_accuracy: 0.8235\n",
      "Epoch 50/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4188 - binary_accuracy: 0.8235\n",
      "Epoch 51/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4214 - binary_accuracy: 0.8266\n",
      "Epoch 52/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.4201 - binary_accuracy: 0.8260\n",
      "Epoch 53/225\n",
      "1754/1754 [==============================] - 1s 376us/step - loss: 0.4273 - binary_accuracy: 0.8210\n",
      "Epoch 54/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4224 - binary_accuracy: 0.8230\n",
      "Epoch 55/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4143 - binary_accuracy: 0.8290\n",
      "Epoch 56/225\n",
      "1754/1754 [==============================] - 1s 385us/step - loss: 0.4235 - binary_accuracy: 0.8251\n",
      "Epoch 57/225\n",
      "1754/1754 [==============================] - 1s 369us/step - loss: 0.4164 - binary_accuracy: 0.8271\n",
      "Epoch 58/225\n",
      "1754/1754 [==============================] - 1s 369us/step - loss: 0.4258 - binary_accuracy: 0.8193\n",
      "Epoch 59/225\n",
      "1754/1754 [==============================] - 1s 367us/step - loss: 0.4222 - binary_accuracy: 0.8251\n",
      "Epoch 60/225\n",
      "1754/1754 [==============================] - 1s 374us/step - loss: 0.4177 - binary_accuracy: 0.8272\n",
      "Epoch 61/225\n",
      "1754/1754 [==============================] - 1s 372us/step - loss: 0.4213 - binary_accuracy: 0.8273\n",
      "Epoch 62/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4152 - binary_accuracy: 0.8283\n",
      "Epoch 63/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4261 - binary_accuracy: 0.8255\n",
      "Epoch 64/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4243 - binary_accuracy: 0.8230\n",
      "Epoch 65/225\n",
      "1754/1754 [==============================] - 1s 389us/step - loss: 0.4224 - binary_accuracy: 0.8259\n",
      "Epoch 66/225\n",
      "1754/1754 [==============================] - 1s 372us/step - loss: 0.4229 - binary_accuracy: 0.8219\n",
      "Epoch 67/225\n",
      "1754/1754 [==============================] - 1s 374us/step - loss: 0.4190 - binary_accuracy: 0.8242\n",
      "Epoch 68/225\n",
      "1754/1754 [==============================] - 1s 372us/step - loss: 0.4204 - binary_accuracy: 0.8267\n",
      "Epoch 69/225\n",
      "1754/1754 [==============================] - 1s 376us/step - loss: 0.4187 - binary_accuracy: 0.8256\n",
      "Epoch 70/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.4184 - binary_accuracy: 0.8297\n",
      "Epoch 71/225\n",
      "1754/1754 [==============================] - 1s 370us/step - loss: 0.4150 - binary_accuracy: 0.8252\n",
      "Epoch 72/225\n",
      "1754/1754 [==============================] - 1s 372us/step - loss: 0.4166 - binary_accuracy: 0.8245\n",
      "Epoch 73/225\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1754/1754 [==============================] - 1s 375us/step - loss: 0.4147 - binary_accuracy: 0.8259\n",
      "Epoch 74/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4237 - binary_accuracy: 0.8200\n",
      "Epoch 75/225\n",
      "1754/1754 [==============================] - 1s 373us/step - loss: 0.4202 - binary_accuracy: 0.8279\n",
      "Epoch 76/225\n",
      "1754/1754 [==============================] - 1s 375us/step - loss: 0.4165 - binary_accuracy: 0.8235\n",
      "Epoch 77/225\n",
      "1754/1754 [==============================] - 1s 374us/step - loss: 0.4125 - binary_accuracy: 0.8299\n",
      "Epoch 78/225\n",
      "1754/1754 [==============================] - 1s 371us/step - loss: 0.4215 - binary_accuracy: 0.8234\n",
      "Epoch 79/225\n",
      "1754/1754 [==============================] - 1s 371us/step - loss: 0.4155 - binary_accuracy: 0.8245\n",
      "Epoch 80/225\n",
      "1754/1754 [==============================] - 1s 385us/step - loss: 0.4102 - binary_accuracy: 0.8288\n",
      "Epoch 81/225\n",
      "1754/1754 [==============================] - 1s 372us/step - loss: 0.4185 - binary_accuracy: 0.8236\n",
      "Epoch 82/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4107 - binary_accuracy: 0.8284\n",
      "Epoch 83/225\n",
      "1754/1754 [==============================] - 1s 370us/step - loss: 0.4183 - binary_accuracy: 0.8324\n",
      "Epoch 84/225\n",
      "1754/1754 [==============================] - 1s 369us/step - loss: 0.4101 - binary_accuracy: 0.8323\n",
      "Epoch 85/225\n",
      "1754/1754 [==============================] - 1s 371us/step - loss: 0.4172 - binary_accuracy: 0.8277\n",
      "Epoch 86/225\n",
      "1754/1754 [==============================] - 1s 369us/step - loss: 0.4141 - binary_accuracy: 0.8240\n",
      "Epoch 87/225\n",
      "1754/1754 [==============================] - 1s 373us/step - loss: 0.4186 - binary_accuracy: 0.8251\n",
      "Epoch 88/225\n",
      "1754/1754 [==============================] - 1s 375us/step - loss: 0.4195 - binary_accuracy: 0.8242\n",
      "Epoch 89/225\n",
      "1754/1754 [==============================] - 1s 371us/step - loss: 0.4230 - binary_accuracy: 0.8238\n",
      "Epoch 90/225\n",
      "1754/1754 [==============================] - 1s 375us/step - loss: 0.4197 - binary_accuracy: 0.8236\n",
      "Epoch 91/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.4168 - binary_accuracy: 0.8279\n",
      "Epoch 92/225\n",
      "1754/1754 [==============================] - 1s 373us/step - loss: 0.4181 - binary_accuracy: 0.8281\n",
      "Epoch 93/225\n",
      "1754/1754 [==============================] - 1s 374us/step - loss: 0.4186 - binary_accuracy: 0.8293\n",
      "Epoch 94/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4080 - binary_accuracy: 0.8291\n",
      "Epoch 95/225\n",
      "1754/1754 [==============================] - 1s 372us/step - loss: 0.4117 - binary_accuracy: 0.8301\n",
      "Epoch 96/225\n",
      "1754/1754 [==============================] - 1s 373us/step - loss: 0.4161 - binary_accuracy: 0.8268\n",
      "Epoch 97/225\n",
      "1754/1754 [==============================] - 1s 374us/step - loss: 0.4167 - binary_accuracy: 0.8276\n",
      "Epoch 98/225\n",
      "1754/1754 [==============================] - 1s 372us/step - loss: 0.4170 - binary_accuracy: 0.8250\n",
      "Epoch 99/225\n",
      "1754/1754 [==============================] - 1s 375us/step - loss: 0.4099 - binary_accuracy: 0.8342\n",
      "Epoch 100/225\n",
      "1754/1754 [==============================] - 1s 370us/step - loss: 0.4106 - binary_accuracy: 0.8257\n",
      "Epoch 101/225\n",
      "1754/1754 [==============================] - 1s 373us/step - loss: 0.4140 - binary_accuracy: 0.8275\n",
      "Epoch 102/225\n",
      "1754/1754 [==============================] - 1s 374us/step - loss: 0.4222 - binary_accuracy: 0.8250\n",
      "Epoch 103/225\n",
      "1754/1754 [==============================] - 1s 388us/step - loss: 0.4138 - binary_accuracy: 0.8260\n",
      "Epoch 104/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.4142 - binary_accuracy: 0.8290\n",
      "Epoch 105/225\n",
      "1754/1754 [==============================] - 1s 369us/step - loss: 0.4168 - binary_accuracy: 0.8276\n",
      "Epoch 106/225\n",
      "1754/1754 [==============================] - 1s 368us/step - loss: 0.4140 - binary_accuracy: 0.8248\n",
      "Epoch 107/225\n",
      "1754/1754 [==============================] - 1s 376us/step - loss: 0.4128 - binary_accuracy: 0.8304\n",
      "Epoch 108/225\n",
      "1754/1754 [==============================] - 1s 409us/step - loss: 0.4105 - binary_accuracy: 0.8288\n",
      "Epoch 109/225\n",
      "1754/1754 [==============================] - 1s 376us/step - loss: 0.4111 - binary_accuracy: 0.8272\n",
      "Epoch 110/225\n",
      "1754/1754 [==============================] - 1s 374us/step - loss: 0.4078 - binary_accuracy: 0.8354\n",
      "Epoch 111/225\n",
      "1754/1754 [==============================] - 1s 374us/step - loss: 0.4148 - binary_accuracy: 0.8281\n",
      "Epoch 112/225\n",
      "1754/1754 [==============================] - 1s 372us/step - loss: 0.4157 - binary_accuracy: 0.8276\n",
      "Epoch 113/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4150 - binary_accuracy: 0.8271\n",
      "Epoch 114/225\n",
      "1754/1754 [==============================] - 1s 374us/step - loss: 0.4194 - binary_accuracy: 0.8279\n",
      "Epoch 115/225\n",
      "1754/1754 [==============================] - 1s 375us/step - loss: 0.4120 - binary_accuracy: 0.8293\n",
      "Epoch 116/225\n",
      "1754/1754 [==============================] - 1s 373us/step - loss: 0.4094 - binary_accuracy: 0.8310\n",
      "Epoch 117/225\n",
      "1754/1754 [==============================] - 1s 372us/step - loss: 0.4115 - binary_accuracy: 0.8294\n",
      "Epoch 118/225\n",
      "1754/1754 [==============================] - 1s 373us/step - loss: 0.4181 - binary_accuracy: 0.8261\n",
      "Epoch 119/225\n",
      "1754/1754 [==============================] - 1s 375us/step - loss: 0.4169 - binary_accuracy: 0.8273\n",
      "Epoch 120/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4107 - binary_accuracy: 0.8286\n",
      "Epoch 121/225\n",
      "1754/1754 [==============================] - 1s 374us/step - loss: 0.4152 - binary_accuracy: 0.8251\n",
      "Epoch 122/225\n",
      "1754/1754 [==============================] - 1s 370us/step - loss: 0.4105 - binary_accuracy: 0.8319\n",
      "Epoch 123/225\n",
      "1754/1754 [==============================] - 1s 367us/step - loss: 0.4116 - binary_accuracy: 0.8287\n",
      "Epoch 124/225\n",
      "1754/1754 [==============================] - 1s 373us/step - loss: 0.4141 - binary_accuracy: 0.8292\n",
      "Epoch 125/225\n",
      "1754/1754 [==============================] - 1s 368us/step - loss: 0.4173 - binary_accuracy: 0.8274\n",
      "Epoch 126/225\n",
      "1754/1754 [==============================] - 1s 374us/step - loss: 0.4068 - binary_accuracy: 0.8296\n",
      "Epoch 127/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4132 - binary_accuracy: 0.8257\n",
      "Epoch 128/225\n",
      "1754/1754 [==============================] - 1s 376us/step - loss: 0.4104 - binary_accuracy: 0.8257\n",
      "Epoch 129/225\n",
      "1754/1754 [==============================] - 1s 368us/step - loss: 0.4151 - binary_accuracy: 0.8286\n",
      "Epoch 130/225\n",
      "1754/1754 [==============================] - 1s 363us/step - loss: 0.4120 - binary_accuracy: 0.8265\n",
      "Epoch 131/225\n",
      "1754/1754 [==============================] - 1s 370us/step - loss: 0.4187 - binary_accuracy: 0.8259\n",
      "Epoch 132/225\n",
      "1754/1754 [==============================] - 1s 367us/step - loss: 0.4128 - binary_accuracy: 0.8287\n",
      "Epoch 133/225\n",
      "1754/1754 [==============================] - 1s 368us/step - loss: 0.4116 - binary_accuracy: 0.8297\n",
      "Epoch 134/225\n",
      "1754/1754 [==============================] - 1s 369us/step - loss: 0.4224 - binary_accuracy: 0.8217\n",
      "Epoch 135/225\n",
      "1754/1754 [==============================] - 1s 368us/step - loss: 0.4119 - binary_accuracy: 0.8277\n",
      "Epoch 136/225\n",
      "1754/1754 [==============================] - 1s 367us/step - loss: 0.4104 - binary_accuracy: 0.8289\n",
      "Epoch 137/225\n",
      "1754/1754 [==============================] - 1s 371us/step - loss: 0.4135 - binary_accuracy: 0.8269\n",
      "Epoch 138/225\n",
      "1754/1754 [==============================] - 1s 363us/step - loss: 0.4147 - binary_accuracy: 0.8269\n",
      "Epoch 139/225\n",
      "1754/1754 [==============================] - 1s 364us/step - loss: 0.4124 - binary_accuracy: 0.8301\n",
      "Epoch 140/225\n",
      "1754/1754 [==============================] - 1s 367us/step - loss: 0.4127 - binary_accuracy: 0.8240\n",
      "Epoch 141/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4164 - binary_accuracy: 0.8248\n",
      "Epoch 142/225\n",
      "1754/1754 [==============================] - 1s 369us/step - loss: 0.4089 - binary_accuracy: 0.8312\n",
      "Epoch 143/225\n",
      "1754/1754 [==============================] - 1s 369us/step - loss: 0.4074 - binary_accuracy: 0.8257\n",
      "Epoch 144/225\n",
      "1754/1754 [==============================] - 1s 367us/step - loss: 0.4113 - binary_accuracy: 0.8295\n",
      "Epoch 145/225\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1754/1754 [==============================] - 1s 365us/step - loss: 0.4170 - binary_accuracy: 0.8248\n",
      "Epoch 146/225\n",
      "1754/1754 [==============================] - 1s 364us/step - loss: 0.4203 - binary_accuracy: 0.8227\n",
      "Epoch 147/225\n",
      "1754/1754 [==============================] - 1s 368us/step - loss: 0.4064 - binary_accuracy: 0.8315\n",
      "Epoch 148/225\n",
      "1754/1754 [==============================] - 1s 363us/step - loss: 0.4046 - binary_accuracy: 0.8287\n",
      "Epoch 149/225\n",
      "1754/1754 [==============================] - 1s 363us/step - loss: 0.4169 - binary_accuracy: 0.8258\n",
      "Epoch 150/225\n",
      "1754/1754 [==============================] - 1s 373us/step - loss: 0.4093 - binary_accuracy: 0.8304\n",
      "Epoch 151/225\n",
      "1754/1754 [==============================] - 1s 365us/step - loss: 0.4147 - binary_accuracy: 0.8261\n",
      "Epoch 152/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4102 - binary_accuracy: 0.8324\n",
      "Epoch 153/225\n",
      "1754/1754 [==============================] - 1s 374us/step - loss: 0.4176 - binary_accuracy: 0.8268\n",
      "Epoch 154/225\n",
      "1754/1754 [==============================] - 1s 366us/step - loss: 0.4147 - binary_accuracy: 0.8291\n",
      "Epoch 155/225\n",
      "1754/1754 [==============================] - 1s 371us/step - loss: 0.4101 - binary_accuracy: 0.8281\n",
      "Epoch 156/225\n",
      "1754/1754 [==============================] - 1s 367us/step - loss: 0.4217 - binary_accuracy: 0.8232\n",
      "Epoch 157/225\n",
      "1754/1754 [==============================] - 1s 366us/step - loss: 0.4058 - binary_accuracy: 0.8327\n",
      "Epoch 158/225\n",
      "1754/1754 [==============================] - 1s 364us/step - loss: 0.4149 - binary_accuracy: 0.8237\n",
      "Epoch 159/225\n",
      "1754/1754 [==============================] - 1s 366us/step - loss: 0.4142 - binary_accuracy: 0.8254\n",
      "Epoch 160/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4115 - binary_accuracy: 0.8280\n",
      "Epoch 161/225\n",
      "1754/1754 [==============================] - 1s 369us/step - loss: 0.4116 - binary_accuracy: 0.8295\n",
      "Epoch 162/225\n",
      "1754/1754 [==============================] - 1s 373us/step - loss: 0.4076 - binary_accuracy: 0.8283\n",
      "Epoch 163/225\n",
      "1754/1754 [==============================] - 1s 370us/step - loss: 0.4101 - binary_accuracy: 0.8266\n",
      "Epoch 164/225\n",
      "1754/1754 [==============================] - 1s 373us/step - loss: 0.4162 - binary_accuracy: 0.8246\n",
      "Epoch 165/225\n",
      "1754/1754 [==============================] - 1s 372us/step - loss: 0.4111 - binary_accuracy: 0.8311\n",
      "Epoch 166/225\n",
      "1754/1754 [==============================] - 1s 388us/step - loss: 0.4068 - binary_accuracy: 0.8286\n",
      "Epoch 167/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4058 - binary_accuracy: 0.8328\n",
      "Epoch 168/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4075 - binary_accuracy: 0.8320\n",
      "Epoch 169/225\n",
      "1754/1754 [==============================] - 1s 371us/step - loss: 0.4130 - binary_accuracy: 0.8278\n",
      "Epoch 170/225\n",
      "1754/1754 [==============================] - 1s 376us/step - loss: 0.4068 - binary_accuracy: 0.8311\n",
      "Epoch 171/225\n",
      "1754/1754 [==============================] - 1s 374us/step - loss: 0.4172 - binary_accuracy: 0.8258\n",
      "Epoch 172/225\n",
      "1754/1754 [==============================] - 1s 374us/step - loss: 0.4070 - binary_accuracy: 0.8332\n",
      "Epoch 173/225\n",
      "1754/1754 [==============================] - 1s 374us/step - loss: 0.4155 - binary_accuracy: 0.8292\n",
      "Epoch 174/225\n",
      "1754/1754 [==============================] - 1s 373us/step - loss: 0.4063 - binary_accuracy: 0.8295\n",
      "Epoch 175/225\n",
      "1754/1754 [==============================] - 1s 374us/step - loss: 0.4107 - binary_accuracy: 0.8282\n",
      "Epoch 176/225\n",
      "1754/1754 [==============================] - 1s 372us/step - loss: 0.4067 - binary_accuracy: 0.8292\n",
      "Epoch 177/225\n",
      "1754/1754 [==============================] - 1s 373us/step - loss: 0.4169 - binary_accuracy: 0.8256\n",
      "Epoch 178/225\n",
      "1754/1754 [==============================] - 1s 371us/step - loss: 0.4096 - binary_accuracy: 0.8319\n",
      "Epoch 179/225\n",
      "1754/1754 [==============================] - 1s 375us/step - loss: 0.4080 - binary_accuracy: 0.8316\n",
      "Epoch 180/225\n",
      "1754/1754 [==============================] - 1s 373us/step - loss: 0.4100 - binary_accuracy: 0.8287\n",
      "Epoch 181/225\n",
      "1754/1754 [==============================] - 1s 373us/step - loss: 0.4102 - binary_accuracy: 0.8307\n",
      "Epoch 182/225\n",
      "1754/1754 [==============================] - 1s 370us/step - loss: 0.4104 - binary_accuracy: 0.8274\n",
      "Epoch 183/225\n",
      "1754/1754 [==============================] - 1s 370us/step - loss: 0.4049 - binary_accuracy: 0.8290\n",
      "Epoch 184/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4120 - binary_accuracy: 0.8304\n",
      "Epoch 185/225\n",
      "1754/1754 [==============================] - 1s 374us/step - loss: 0.4071 - binary_accuracy: 0.8272\n",
      "Epoch 186/225\n",
      "1754/1754 [==============================] - 1s 374us/step - loss: 0.4082 - binary_accuracy: 0.8285\n",
      "Epoch 187/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.4089 - binary_accuracy: 0.8292\n",
      "Epoch 188/225\n",
      "1754/1754 [==============================] - 1s 376us/step - loss: 0.4109 - binary_accuracy: 0.8301\n",
      "Epoch 189/225\n",
      "1754/1754 [==============================] - 1s 375us/step - loss: 0.4049 - binary_accuracy: 0.8321\n",
      "Epoch 190/225\n",
      "1754/1754 [==============================] - 1s 373us/step - loss: 0.4098 - binary_accuracy: 0.8265\n",
      "Epoch 191/225\n",
      "1754/1754 [==============================] - 1s 370us/step - loss: 0.4069 - binary_accuracy: 0.8295\n",
      "Epoch 192/225\n",
      "1754/1754 [==============================] - 1s 373us/step - loss: 0.4154 - binary_accuracy: 0.8234\n",
      "Epoch 193/225\n",
      "1754/1754 [==============================] - 1s 373us/step - loss: 0.4077 - binary_accuracy: 0.8244\n",
      "Epoch 194/225\n",
      "1754/1754 [==============================] - 1s 372us/step - loss: 0.4077 - binary_accuracy: 0.8291\n",
      "Epoch 195/225\n",
      "1754/1754 [==============================] - 1s 374us/step - loss: 0.4122 - binary_accuracy: 0.8243\n",
      "Epoch 196/225\n",
      "1754/1754 [==============================] - 1s 373us/step - loss: 0.4125 - binary_accuracy: 0.8270\n",
      "Epoch 197/225\n",
      "1754/1754 [==============================] - 1s 372us/step - loss: 0.4124 - binary_accuracy: 0.8316\n",
      "Epoch 198/225\n",
      "1754/1754 [==============================] - 1s 373us/step - loss: 0.4129 - binary_accuracy: 0.8258\n",
      "Epoch 199/225\n",
      "1754/1754 [==============================] - 1s 372us/step - loss: 0.4100 - binary_accuracy: 0.8310\n",
      "Epoch 200/225\n",
      "1754/1754 [==============================] - 1s 375us/step - loss: 0.4165 - binary_accuracy: 0.8170\n",
      "Epoch 201/225\n",
      "1754/1754 [==============================] - 1s 374us/step - loss: 0.4083 - binary_accuracy: 0.8300\n",
      "Epoch 202/225\n",
      "1754/1754 [==============================] - 1s 373us/step - loss: 0.4059 - binary_accuracy: 0.8297\n",
      "Epoch 203/225\n",
      "1754/1754 [==============================] - 1s 375us/step - loss: 0.4120 - binary_accuracy: 0.8262\n",
      "Epoch 204/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4108 - binary_accuracy: 0.8286\n",
      "Epoch 205/225\n",
      "1754/1754 [==============================] - 1s 371us/step - loss: 0.4069 - binary_accuracy: 0.8253\n",
      "Epoch 206/225\n",
      "1754/1754 [==============================] - 1s 368us/step - loss: 0.4130 - binary_accuracy: 0.8269\n",
      "Epoch 207/225\n",
      "1754/1754 [==============================] - 1s 373us/step - loss: 0.4084 - binary_accuracy: 0.8267\n",
      "Epoch 208/225\n",
      "1754/1754 [==============================] - 1s 372us/step - loss: 0.4140 - binary_accuracy: 0.8274\n",
      "Epoch 209/225\n",
      "1754/1754 [==============================] - 1s 371us/step - loss: 0.4077 - binary_accuracy: 0.8278\n",
      "Epoch 210/225\n",
      "1754/1754 [==============================] - 1s 371us/step - loss: 0.4020 - binary_accuracy: 0.8349\n",
      "Epoch 211/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4070 - binary_accuracy: 0.8285\n",
      "Epoch 212/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4010 - binary_accuracy: 0.8324\n",
      "Epoch 213/225\n",
      "1754/1754 [==============================] - 1s 391us/step - loss: 0.4044 - binary_accuracy: 0.8327\n",
      "Epoch 214/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4159 - binary_accuracy: 0.8240\n",
      "Epoch 215/225\n",
      "1754/1754 [==============================] - 1s 374us/step - loss: 0.4124 - binary_accuracy: 0.8276\n",
      "Epoch 216/225\n",
      "1754/1754 [==============================] - 1s 376us/step - loss: 0.4053 - binary_accuracy: 0.8343\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 217/225\n",
      "1754/1754 [==============================] - 1s 371us/step - loss: 0.4107 - binary_accuracy: 0.8257\n",
      "Epoch 218/225\n",
      "1754/1754 [==============================] - 1s 371us/step - loss: 0.4075 - binary_accuracy: 0.8317\n",
      "Epoch 219/225\n",
      "1754/1754 [==============================] - 1s 373us/step - loss: 0.4097 - binary_accuracy: 0.8315\n",
      "Epoch 220/225\n",
      "1754/1754 [==============================] - 1s 375us/step - loss: 0.4128 - binary_accuracy: 0.8273\n",
      "Epoch 221/225\n",
      "1754/1754 [==============================] - 1s 372us/step - loss: 0.4028 - binary_accuracy: 0.8324\n",
      "Epoch 222/225\n",
      "1754/1754 [==============================] - 1s 369us/step - loss: 0.4056 - binary_accuracy: 0.8301\n",
      "Epoch 223/225\n",
      "1754/1754 [==============================] - 1s 372us/step - loss: 0.4066 - binary_accuracy: 0.8284\n",
      "Epoch 224/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.4069 - binary_accuracy: 0.8305\n",
      "Epoch 225/225\n",
      "1754/1754 [==============================] - 1s 374us/step - loss: 0.4055 - binary_accuracy: 0.8266\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/225\n",
      "1754/1754 [==============================] - 2s 902us/step - loss: 0.4785 - binary_accuracy: 0.8110\n",
      "Epoch 2/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4566 - binary_accuracy: 0.8167\n",
      "Epoch 3/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4444 - binary_accuracy: 0.8171\n",
      "Epoch 4/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4471 - binary_accuracy: 0.8163\n",
      "Epoch 5/225\n",
      "1754/1754 [==============================] - 1s 391us/step - loss: 0.4387 - binary_accuracy: 0.8199\n",
      "Epoch 6/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4420 - binary_accuracy: 0.8177\n",
      "Epoch 7/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4356 - binary_accuracy: 0.8220\n",
      "Epoch 8/225\n",
      "1754/1754 [==============================] - 1s 385us/step - loss: 0.4324 - binary_accuracy: 0.8165\n",
      "Epoch 9/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4315 - binary_accuracy: 0.8198\n",
      "Epoch 10/225\n",
      "1754/1754 [==============================] - 1s 385us/step - loss: 0.4256 - binary_accuracy: 0.8198\n",
      "Epoch 11/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4361 - binary_accuracy: 0.8181\n",
      "Epoch 12/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4238 - binary_accuracy: 0.8233\n",
      "Epoch 13/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4260 - binary_accuracy: 0.8243\n",
      "Epoch 14/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4276 - binary_accuracy: 0.8246\n",
      "Epoch 15/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4250 - binary_accuracy: 0.8227\n",
      "Epoch 16/225\n",
      "1754/1754 [==============================] - 1s 388us/step - loss: 0.4272 - binary_accuracy: 0.8228\n",
      "Epoch 17/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.4362 - binary_accuracy: 0.8199\n",
      "Epoch 18/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4331 - binary_accuracy: 0.8186\n",
      "Epoch 19/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4259 - binary_accuracy: 0.8235\n",
      "Epoch 20/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4245 - binary_accuracy: 0.8221\n",
      "Epoch 21/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4224 - binary_accuracy: 0.8259\n",
      "Epoch 22/225\n",
      "1754/1754 [==============================] - 1s 385us/step - loss: 0.4204 - binary_accuracy: 0.8210\n",
      "Epoch 23/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4297 - binary_accuracy: 0.8193\n",
      "Epoch 24/225\n",
      "1754/1754 [==============================] - 1s 375us/step - loss: 0.4198 - binary_accuracy: 0.8221\n",
      "Epoch 25/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4230 - binary_accuracy: 0.8193\n",
      "Epoch 26/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.4163 - binary_accuracy: 0.8237\n",
      "Epoch 27/225\n",
      "1754/1754 [==============================] - 1s 374us/step - loss: 0.4265 - binary_accuracy: 0.8189\n",
      "Epoch 28/225\n",
      "1754/1754 [==============================] - 1s 374us/step - loss: 0.4260 - binary_accuracy: 0.8241\n",
      "Epoch 29/225\n",
      "1754/1754 [==============================] - 1s 375us/step - loss: 0.4302 - binary_accuracy: 0.8187\n",
      "Epoch 30/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4239 - binary_accuracy: 0.8207\n",
      "Epoch 31/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4255 - binary_accuracy: 0.8248\n",
      "Epoch 32/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4150 - binary_accuracy: 0.8246\n",
      "Epoch 33/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4215 - binary_accuracy: 0.8229\n",
      "Epoch 34/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4164 - binary_accuracy: 0.8249\n",
      "Epoch 35/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4190 - binary_accuracy: 0.8215\n",
      "Epoch 36/225\n",
      "1754/1754 [==============================] - 1s 388us/step - loss: 0.4198 - binary_accuracy: 0.8233\n",
      "Epoch 37/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4147 - binary_accuracy: 0.8278\n",
      "Epoch 38/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4213 - binary_accuracy: 0.8192\n",
      "Epoch 39/225\n",
      "1754/1754 [==============================] - 1s 422us/step - loss: 0.4208 - binary_accuracy: 0.8244\n",
      "Epoch 40/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4238 - binary_accuracy: 0.8167\n",
      "Epoch 41/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4214 - binary_accuracy: 0.8209\n",
      "Epoch 42/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4209 - binary_accuracy: 0.8235\n",
      "Epoch 43/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4167 - binary_accuracy: 0.8253\n",
      "Epoch 44/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4144 - binary_accuracy: 0.8276\n",
      "Epoch 45/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4160 - binary_accuracy: 0.8262\n",
      "Epoch 46/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4196 - binary_accuracy: 0.8236\n",
      "Epoch 47/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.4208 - binary_accuracy: 0.8251\n",
      "Epoch 48/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4286 - binary_accuracy: 0.8215\n",
      "Epoch 49/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4222 - binary_accuracy: 0.8192\n",
      "Epoch 50/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4198 - binary_accuracy: 0.8255\n",
      "Epoch 51/225\n",
      "1754/1754 [==============================] - 1s 385us/step - loss: 0.4247 - binary_accuracy: 0.8231\n",
      "Epoch 52/225\n",
      "1754/1754 [==============================] - 1s 385us/step - loss: 0.4204 - binary_accuracy: 0.8225\n",
      "Epoch 53/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.4224 - binary_accuracy: 0.8226\n",
      "Epoch 54/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4176 - binary_accuracy: 0.8200\n",
      "Epoch 55/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.4096 - binary_accuracy: 0.8248\n",
      "Epoch 56/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4166 - binary_accuracy: 0.8249\n",
      "Epoch 57/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4269 - binary_accuracy: 0.8202\n",
      "Epoch 58/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4198 - binary_accuracy: 0.8298\n",
      "Epoch 59/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4200 - binary_accuracy: 0.8252\n",
      "Epoch 60/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4136 - binary_accuracy: 0.8290\n",
      "Epoch 61/225\n",
      "1754/1754 [==============================] - 1s 388us/step - loss: 0.4200 - binary_accuracy: 0.8253\n",
      "Epoch 62/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4203 - binary_accuracy: 0.8236\n",
      "Epoch 63/225\n",
      "1754/1754 [==============================] - 1s 375us/step - loss: 0.4136 - binary_accuracy: 0.8291\n",
      "Epoch 64/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4121 - binary_accuracy: 0.8209\n",
      "Epoch 65/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4174 - binary_accuracy: 0.8288\n",
      "Epoch 66/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4194 - binary_accuracy: 0.8258\n",
      "Epoch 67/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4226 - binary_accuracy: 0.8229\n",
      "Epoch 68/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4154 - binary_accuracy: 0.8294\n",
      "Epoch 69/225\n",
      "1754/1754 [==============================] - 1s 374us/step - loss: 0.4170 - binary_accuracy: 0.8214\n",
      "Epoch 70/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4206 - binary_accuracy: 0.8241\n",
      "Epoch 71/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4155 - binary_accuracy: 0.8269\n",
      "Epoch 72/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4112 - binary_accuracy: 0.8293\n",
      "Epoch 73/225\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4154 - binary_accuracy: 0.8253\n",
      "Epoch 74/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4166 - binary_accuracy: 0.8273\n",
      "Epoch 75/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4140 - binary_accuracy: 0.8245\n",
      "Epoch 76/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4146 - binary_accuracy: 0.8262\n",
      "Epoch 77/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4183 - binary_accuracy: 0.8249\n",
      "Epoch 78/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4133 - binary_accuracy: 0.8322\n",
      "Epoch 79/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4154 - binary_accuracy: 0.8254\n",
      "Epoch 80/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4146 - binary_accuracy: 0.8268\n",
      "Epoch 81/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4219 - binary_accuracy: 0.8272\n",
      "Epoch 82/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4106 - binary_accuracy: 0.8307\n",
      "Epoch 83/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4079 - binary_accuracy: 0.8276\n",
      "Epoch 84/225\n",
      "1754/1754 [==============================] - 1s 385us/step - loss: 0.4214 - binary_accuracy: 0.8249\n",
      "Epoch 85/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4175 - binary_accuracy: 0.8267\n",
      "Epoch 86/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4168 - binary_accuracy: 0.8250\n",
      "Epoch 87/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4126 - binary_accuracy: 0.8312\n",
      "Epoch 88/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4126 - binary_accuracy: 0.8265\n",
      "Epoch 89/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4113 - binary_accuracy: 0.8305\n",
      "Epoch 90/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4145 - binary_accuracy: 0.8277\n",
      "Epoch 91/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4121 - binary_accuracy: 0.8310\n",
      "Epoch 92/225\n",
      "1754/1754 [==============================] - 1s 387us/step - loss: 0.4103 - binary_accuracy: 0.8283\n",
      "Epoch 93/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4205 - binary_accuracy: 0.8240\n",
      "Epoch 94/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4136 - binary_accuracy: 0.8266\n",
      "Epoch 95/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4055 - binary_accuracy: 0.8299\n",
      "Epoch 96/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4068 - binary_accuracy: 0.8320\n",
      "Epoch 97/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4222 - binary_accuracy: 0.8270\n",
      "Epoch 98/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4143 - binary_accuracy: 0.8246\n",
      "Epoch 99/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4147 - binary_accuracy: 0.8263\n",
      "Epoch 100/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4117 - binary_accuracy: 0.8300\n",
      "Epoch 101/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4204 - binary_accuracy: 0.8215\n",
      "Epoch 102/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4131 - binary_accuracy: 0.8233\n",
      "Epoch 103/225\n",
      "1754/1754 [==============================] - 1s 411us/step - loss: 0.4132 - binary_accuracy: 0.8269\n",
      "Epoch 104/225\n",
      "1754/1754 [==============================] - 1s 411us/step - loss: 0.4084 - binary_accuracy: 0.8277\n",
      "Epoch 105/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4128 - binary_accuracy: 0.8288\n",
      "Epoch 106/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4096 - binary_accuracy: 0.8282\n",
      "Epoch 107/225\n",
      "1754/1754 [==============================] - 1s 387us/step - loss: 0.4137 - binary_accuracy: 0.8291\n",
      "Epoch 108/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4068 - binary_accuracy: 0.8297\n",
      "Epoch 109/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4154 - binary_accuracy: 0.8244\n",
      "Epoch 110/225\n",
      "1754/1754 [==============================] - 1s 388us/step - loss: 0.4126 - binary_accuracy: 0.8276\n",
      "Epoch 111/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4133 - binary_accuracy: 0.8199\n",
      "Epoch 112/225\n",
      "1754/1754 [==============================] - 1s 389us/step - loss: 0.4114 - binary_accuracy: 0.8226\n",
      "Epoch 113/225\n",
      "1754/1754 [==============================] - 1s 387us/step - loss: 0.4144 - binary_accuracy: 0.8223\n",
      "Epoch 114/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4079 - binary_accuracy: 0.8282\n",
      "Epoch 115/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4148 - binary_accuracy: 0.8273\n",
      "Epoch 116/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4228 - binary_accuracy: 0.8178\n",
      "Epoch 117/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4222 - binary_accuracy: 0.8224\n",
      "Epoch 118/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4110 - binary_accuracy: 0.8283\n",
      "Epoch 119/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4074 - binary_accuracy: 0.8300\n",
      "Epoch 120/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4191 - binary_accuracy: 0.8230\n",
      "Epoch 121/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4080 - binary_accuracy: 0.8297\n",
      "Epoch 122/225\n",
      "1754/1754 [==============================] - 1s 387us/step - loss: 0.4061 - binary_accuracy: 0.8319\n",
      "Epoch 123/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4090 - binary_accuracy: 0.8271\n",
      "Epoch 124/225\n",
      "1754/1754 [==============================] - 1s 376us/step - loss: 0.4070 - binary_accuracy: 0.8332\n",
      "Epoch 125/225\n",
      "1754/1754 [==============================] - 1s 376us/step - loss: 0.4142 - binary_accuracy: 0.8257\n",
      "Epoch 126/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4111 - binary_accuracy: 0.8289\n",
      "Epoch 127/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4197 - binary_accuracy: 0.8262\n",
      "Epoch 128/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4099 - binary_accuracy: 0.8291\n",
      "Epoch 129/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4136 - binary_accuracy: 0.8267\n",
      "Epoch 130/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4074 - binary_accuracy: 0.8267\n",
      "Epoch 131/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4097 - binary_accuracy: 0.8284\n",
      "Epoch 132/225\n",
      "1754/1754 [==============================] - 1s 375us/step - loss: 0.4119 - binary_accuracy: 0.8309\n",
      "Epoch 133/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4123 - binary_accuracy: 0.8264\n",
      "Epoch 134/225\n",
      "1754/1754 [==============================] - 1s 375us/step - loss: 0.4135 - binary_accuracy: 0.8262\n",
      "Epoch 135/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4097 - binary_accuracy: 0.8235\n",
      "Epoch 136/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4109 - binary_accuracy: 0.8282\n",
      "Epoch 137/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.4078 - binary_accuracy: 0.8265\n",
      "Epoch 138/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4166 - binary_accuracy: 0.8237\n",
      "Epoch 139/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4086 - binary_accuracy: 0.8303\n",
      "Epoch 140/225\n",
      "1754/1754 [==============================] - 1s 375us/step - loss: 0.4088 - binary_accuracy: 0.8282\n",
      "Epoch 141/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4102 - binary_accuracy: 0.8291\n",
      "Epoch 142/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4068 - binary_accuracy: 0.8288\n",
      "Epoch 143/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4148 - binary_accuracy: 0.8252\n",
      "Epoch 144/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.3990 - binary_accuracy: 0.8375\n",
      "Epoch 145/225\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4134 - binary_accuracy: 0.8252\n",
      "Epoch 146/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4166 - binary_accuracy: 0.8203\n",
      "Epoch 147/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4101 - binary_accuracy: 0.8285\n",
      "Epoch 148/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4149 - binary_accuracy: 0.8292\n",
      "Epoch 149/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4123 - binary_accuracy: 0.8236\n",
      "Epoch 150/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4120 - binary_accuracy: 0.8245\n",
      "Epoch 151/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4147 - binary_accuracy: 0.8305\n",
      "Epoch 152/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.3970 - binary_accuracy: 0.8357\n",
      "Epoch 153/225\n",
      "1754/1754 [==============================] - 1s 387us/step - loss: 0.4117 - binary_accuracy: 0.8315\n",
      "Epoch 154/225\n",
      "1754/1754 [==============================] - 1s 375us/step - loss: 0.4051 - binary_accuracy: 0.8283\n",
      "Epoch 155/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4158 - binary_accuracy: 0.8213\n",
      "Epoch 156/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4008 - binary_accuracy: 0.8333\n",
      "Epoch 157/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4046 - binary_accuracy: 0.8294\n",
      "Epoch 158/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4102 - binary_accuracy: 0.8293\n",
      "Epoch 159/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4071 - binary_accuracy: 0.8316\n",
      "Epoch 160/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.3999 - binary_accuracy: 0.8313\n",
      "Epoch 161/225\n",
      "1754/1754 [==============================] - 1s 425us/step - loss: 0.4063 - binary_accuracy: 0.8276\n",
      "Epoch 162/225\n",
      "1754/1754 [==============================] - 1s 391us/step - loss: 0.4115 - binary_accuracy: 0.8285\n",
      "Epoch 163/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4161 - binary_accuracy: 0.8233\n",
      "Epoch 164/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4080 - binary_accuracy: 0.8243\n",
      "Epoch 165/225\n",
      "1754/1754 [==============================] - 1s 388us/step - loss: 0.4064 - binary_accuracy: 0.8286\n",
      "Epoch 166/225\n",
      "1754/1754 [==============================] - 1s 388us/step - loss: 0.4093 - binary_accuracy: 0.8296\n",
      "Epoch 167/225\n",
      "1754/1754 [==============================] - 1s 385us/step - loss: 0.4067 - binary_accuracy: 0.8287\n",
      "Epoch 168/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4089 - binary_accuracy: 0.8324\n",
      "Epoch 169/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4127 - binary_accuracy: 0.8245\n",
      "Epoch 170/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4112 - binary_accuracy: 0.8236\n",
      "Epoch 171/225\n",
      "1754/1754 [==============================] - 1s 385us/step - loss: 0.4042 - binary_accuracy: 0.8317\n",
      "Epoch 172/225\n",
      "1754/1754 [==============================] - 1s 387us/step - loss: 0.4090 - binary_accuracy: 0.8258\n",
      "Epoch 173/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4076 - binary_accuracy: 0.8261\n",
      "Epoch 174/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4123 - binary_accuracy: 0.8244\n",
      "Epoch 175/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4125 - binary_accuracy: 0.8271\n",
      "Epoch 176/225\n",
      "1754/1754 [==============================] - 1s 388us/step - loss: 0.4053 - binary_accuracy: 0.8329\n",
      "Epoch 177/225\n",
      "1754/1754 [==============================] - 1s 384us/step - loss: 0.4111 - binary_accuracy: 0.8307\n",
      "Epoch 178/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4068 - binary_accuracy: 0.8271\n",
      "Epoch 179/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4086 - binary_accuracy: 0.8264\n",
      "Epoch 180/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4122 - binary_accuracy: 0.8282\n",
      "Epoch 181/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4089 - binary_accuracy: 0.8286\n",
      "Epoch 182/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4120 - binary_accuracy: 0.8299\n",
      "Epoch 183/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4087 - binary_accuracy: 0.8343\n",
      "Epoch 184/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4005 - binary_accuracy: 0.8314\n",
      "Epoch 185/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.4160 - binary_accuracy: 0.8260\n",
      "Epoch 186/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.4051 - binary_accuracy: 0.8261\n",
      "Epoch 187/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4059 - binary_accuracy: 0.8295\n",
      "Epoch 188/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.4139 - binary_accuracy: 0.8233\n",
      "Epoch 189/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.4033 - binary_accuracy: 0.8285\n",
      "Epoch 190/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.4028 - binary_accuracy: 0.8325\n",
      "Epoch 191/225\n",
      "1754/1754 [==============================] - 1s 372us/step - loss: 0.4086 - binary_accuracy: 0.8259\n",
      "Epoch 192/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4074 - binary_accuracy: 0.8289\n",
      "Epoch 193/225\n",
      "1754/1754 [==============================] - 1s 379us/step - loss: 0.4069 - binary_accuracy: 0.8311\n",
      "Epoch 194/225\n",
      "1754/1754 [==============================] - 1s 376us/step - loss: 0.4103 - binary_accuracy: 0.8305\n",
      "Epoch 195/225\n",
      "1754/1754 [==============================] - 1s 387us/step - loss: 0.4033 - binary_accuracy: 0.8304\n",
      "Epoch 196/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4087 - binary_accuracy: 0.8274\n",
      "Epoch 197/225\n",
      "1754/1754 [==============================] - 1s 375us/step - loss: 0.4061 - binary_accuracy: 0.8304\n",
      "Epoch 198/225\n",
      "1754/1754 [==============================] - 1s 375us/step - loss: 0.4055 - binary_accuracy: 0.8317\n",
      "Epoch 199/225\n",
      "1754/1754 [==============================] - 1s 376us/step - loss: 0.4043 - binary_accuracy: 0.8297\n",
      "Epoch 200/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4075 - binary_accuracy: 0.8287\n",
      "Epoch 201/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4009 - binary_accuracy: 0.8327\n",
      "Epoch 202/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4132 - binary_accuracy: 0.8256\n",
      "Epoch 203/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4093 - binary_accuracy: 0.8276\n",
      "Epoch 204/225\n",
      "1754/1754 [==============================] - 1s 377us/step - loss: 0.4178 - binary_accuracy: 0.8243\n",
      "Epoch 205/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4082 - binary_accuracy: 0.8307\n",
      "Epoch 206/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4022 - binary_accuracy: 0.8321\n",
      "Epoch 207/225\n",
      "1754/1754 [==============================] - 1s 387us/step - loss: 0.4101 - binary_accuracy: 0.8292\n",
      "Epoch 208/225\n",
      "1754/1754 [==============================] - 1s 375us/step - loss: 0.4076 - binary_accuracy: 0.8292\n",
      "Epoch 209/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.4020 - binary_accuracy: 0.8292\n",
      "Epoch 210/225\n",
      "1754/1754 [==============================] - 1s 380us/step - loss: 0.4063 - binary_accuracy: 0.8267\n",
      "Epoch 211/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4034 - binary_accuracy: 0.8297\n",
      "Epoch 212/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4064 - binary_accuracy: 0.8263\n",
      "Epoch 213/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4027 - binary_accuracy: 0.8274\n",
      "Epoch 214/225\n",
      "1754/1754 [==============================] - 1s 391us/step - loss: 0.4099 - binary_accuracy: 0.8270\n",
      "Epoch 215/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4091 - binary_accuracy: 0.8255\n",
      "Epoch 216/225\n",
      "1754/1754 [==============================] - 1s 388us/step - loss: 0.4000 - binary_accuracy: 0.8367\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 217/225\n",
      "1754/1754 [==============================] - 1s 382us/step - loss: 0.4061 - binary_accuracy: 0.8277\n",
      "Epoch 218/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4092 - binary_accuracy: 0.8303\n",
      "Epoch 219/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4132 - binary_accuracy: 0.8260\n",
      "Epoch 220/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4080 - binary_accuracy: 0.8271\n",
      "Epoch 221/225\n",
      "1754/1754 [==============================] - 1s 423us/step - loss: 0.4080 - binary_accuracy: 0.8292\n",
      "Epoch 222/225\n",
      "1754/1754 [==============================] - 1s 385us/step - loss: 0.4041 - binary_accuracy: 0.8326\n",
      "Epoch 223/225\n",
      "1754/1754 [==============================] - 1s 383us/step - loss: 0.4065 - binary_accuracy: 0.8280\n",
      "Epoch 224/225\n",
      "1754/1754 [==============================] - 1s 378us/step - loss: 0.3996 - binary_accuracy: 0.8332\n",
      "Epoch 225/225\n",
      "1754/1754 [==============================] - 1s 381us/step - loss: 0.4086 - binary_accuracy: 0.8295\n"
     ]
    },
    {
     "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": [
    "## Run 5-fold cross-validation to get average performance characteristics\n",
    "unit = 32\n",
    "ratio = 0.5\n",
    "epoch = 225\n",
    "(auroc_list, other_metric_list) = kfold_validation(unit, ratio, epoch, X_train_transformed, y_train, 'patient-id', \\\n",
    "                                                   n_features=n_features, verbose=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "average precision:0.37557251911057643\n",
      "average recall:0.5695732149907149\n",
      "average specificity:0.7870743807158543\n",
      "average F1:0.44813238795595717\n"
     ]
    }
   ],
   "source": [
    "## other_metric_list has precision, recall, accuracy, specificity, f1, threshold for metrics\n",
    "ave_prec = np.mean([x[0] for x in other_metric_list])\n",
    "print(\"average precision:\" + str(ave_prec))\n",
    "\n",
    "ave_recall = np.mean([x[1] for x in other_metric_list])\n",
    "print(\"average recall:\" + str(ave_recall))\n",
    "\n",
    "ave_specificity = np.mean([x[3] for x in other_metric_list])\n",
    "print(\"average specificity:\" + str(ave_specificity))\n",
    "\n",
    "ave_f1 = np.mean([x[4] for x in other_metric_list])\n",
    "print(\"average F1:\" + str(ave_f1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "max F1: 0.45569620253164556\n"
     ]
    }
   ],
   "source": [
    "max_f1 = np.max([x[4] for x in other_metric_list])\n",
    "print(\"max F1: \" + str(max_f1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "unit: 32\tepoch: 225\tauroc_mean: 0.722735\tauroc_std: 0.012118\n"
     ]
    }
   ],
   "source": [
    "## Average metrics performance\n",
    "print(\"unit: %d\\tepoch: %d\\tauroc_mean: %f\\tauroc_std: %f\"  % (unit, epoch, np.mean(auroc_list), np.std(auroc_list)))"
   ]
  },
  {
   "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}