SimpleRNN_Experiment2_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",
    "# 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",
    "             \"mds-delirium-scale\", \"mds-long-term-memory-ok\", \"mds-short-term-memory-ok\", \"mds-staff-assess-pain\", \\\n",
    "             \"recent-start-other-fall-risk-rx\"] + [\"mds-pressure-ulcer-stage-%d\" % i for i in range(1, 5)]\n",
    "cate_one_hot = [['Yes', 'No', 'Unable to answer']] * 3 + [['Yes', 'No']] * 9\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 += [\"cms-long-stay\", \"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-dehydrated\", \"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\", \"ws-antibiotic-anticoag-coexposure\", \\\n",
    "            \"ws-psychotropic-with-not-ordered-weight-loss\", \\\n",
    "            \"ws-diuretic-adl\", \"ws-tramadol-antidepressant-coexposure\"] + [\"mds-pain-non-verbal\"]\n",
    "dic_label += [{'Yes': 1, 'No': 0}] * 28 + [{'None/Mild':0,'Moderate/Severe':1}]\n",
    "\n",
    "f_one_hot += [\"race\", \"facility\", \"gender\", \"mds-bims-summary-ranking\", \"ws-meclizine-psych-coexposure\"]\n",
    "cate_one_hot += [['White', 'Black', 'Other'], \\\n",
    "                 ['Sugar Creek', 'Heritage Place', 'Canterbury Place', 'Senaca Place', 'Cranberry Place'], \\\n",
    "                 ['Female', 'Male'], ['Intact or Moderately Intact', 'Moderate Impairment'], \\\n",
    "                 ['No', 'meclAndPsychStartSameR', 'meclPrecedesPsychR', 'psychPrecedesMeclR', 'meclAndPsychOlderStart']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Features need imputation\n",
    "f_one_hot += [\"pneumonia\", \"uti\", \"mdro\", \"anemia\", \"septicemia\", \"constipation\", \"wound\", \"hyponatremia\", \\\n",
    "              \"hyperkalemia\", \"embolisms\", \"alzheimers\", \"anxiety\", \"depression\", \"non-alz-dimentia\", \"bipolar\", \\\n",
    "              \"parkinsons\", \"psychosis\", \"schizophrenia\", \"seizure\", \"aphasia\", \"emphysema\", \"arthritis\", \"ashd\", \\\n",
    "              \"bph\", \"cancer\", \"cerebralpalsy\", \"stroke\", \"cirrhosis\", \"comatose\", \"diabetes\", \"dysrhythmias\", \"gerd\", \\\n",
    "              \"heart-failure\", \"hemiplegia/hemiparesis\", \"hepatitis\", \"huntingtons\", \"hyperlipidemia\", \"hyperthyroidism\", \\\n",
    "              \"hypothyroidism\", \"hypertension\", \"hypotension\", \"multiple-sclerosis\", \"neurogenic-bladder\", \\\n",
    "              \"obstructive-uropathy\", \"osteoporosis\", \"paraplegia\", \"ptsd\", \"pvd\", \"quadriplegia\", \"thyroid-disorder\",\\\n",
    "              \"tourettes\", \"transient-ischemic-attack\", \"traumatic-brain-injury\", \"tuberculosis\", \"renal-failure\",\\\n",
    "              \"mds-delirium-scale\", \"mds-pain-medication\", \"mds-received-pain-tx-non-pharm\", \\\n",
    "              \"mds-received-prn-pain-medication\", \"mds-conduct-staff-assessment-mental-status\", \\\n",
    "              \"mds-pressure-ulcer-prsnt\", \"conduct_bims\", \"acute_mental_change\", \"conduct_pain_assmnt\", \\\n",
    "              \"cane_or_crutch_past_7_days\", \"walker_past_7_days\", \"wheelchair_past_7_days\", \\\n",
    "              \"limb_prosthesis_past_7_days\", \"no_listed_mobility_device\"] + \\\n",
    "              [\"balance_while_standing\", \"balance_while_walking\", \"balance_turning_around\", \\\n",
    "               \"balance_toileting\", \"balance_bed_to_chair\"] + \\\n",
    "              [\"marital-status\", \"mds-entered-from\", \"mds-excess-weight-loss\"]\n",
    "cate_one_hot += [['Yes', 'No']] * 69 + \\\n",
    "                [[\"Steady at all times\", \"able to stabilize without assistance\", \\\n",
    "                  \"able to stabilize with assistance\", \"Activity did not occur\"]] * 5 + \\\n",
    "                [[\"Never Married\", \"Married\", \"Widowed\", \"Separated\", \"Divorced\"], \\\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",
    "                 [\"Yes - prescribed\", \"Yes - not prescribed\", \"No or unknown\"]]\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\", \"cam_motor_retardation\"]\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}] * 4\n",
    "\n",
    "f_transform += [(\"mds-urinary-incontinence\", 3, 0), (\"cam_inattention\", 2, 0), (\"cam_disorganized_thought\", 2, 0), \\\n",
    "                (\"cam_altered_conc\", 2, 0), (\"cam_motor_retardation\", 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 28s, sys: 1.52 s, total: 1min 30s\n",
      "Wall time: 29.5 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 25.4 s, sys: 76.2 ms, total: 25.5 s\n",
      "Wall time: 25.5 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, 524)"
      ]
     },
     "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, 524)"
      ]
     },
     "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>cms-long-stay</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>...</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>0</td>\n",
       "      <td>67</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>1</th>\n",
       "      <td>10364</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>0</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>...</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>0</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>...</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>0</td>\n",
       "      <td>90</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>6</th>\n",
       "      <td>5003</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>0</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>...</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 × 524 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   patient-id  current-stay-days  cumulative-days-in-facility  cms-long-stay  \\\n",
       "0       10353                  7                            7              0   \n",
       "1       10364                  7                            7              0   \n",
       "4       10386                  7                            7              0   \n",
       "5        5000                  7                            7              0   \n",
       "6        5003                  7                            7              0   \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  ...  atc_5_N06AB10  atc_5_N06AX05  \\\n",
       "0                          0  ...              0              0   \n",
       "1                          1  ...              0              0   \n",
       "4                          0  ...              0              0   \n",
       "5                          0  ...              0              0   \n",
       "6                          1  ...              0              0   \n",
       "\n",
       "   atc_5_N06AX11  atc_5_N06AX12  atc_5_N06AX16  atc_5_N06AX21  atc_5_N06AX24  \\\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_N05AA01  atc_5_N05BA05  atc_5_N06AX23  \n",
       "0              0              0              0  \n",
       "1              0              0              0  \n",
       "4              0              0              0  \n",
       "5              0              0              0  \n",
       "6              0              0              0  \n",
       "\n",
       "[5 rows x 524 columns]"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_use.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_use.to_csv (r'/tmp/export_train_dataframe.csv', index = False, header=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "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>cms-long-stay</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>...</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>0</td>\n",
       "      <td>84</td>\n",
       "      <td>0</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>3</th>\n",
       "      <td>10385</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>0</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>...</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>0</td>\n",
       "      <td>85</td>\n",
       "      <td>1</td>\n",
       "      <td>1</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>15</th>\n",
       "      <td>5032</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>0</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>...</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>0</td>\n",
       "      <td>86</td>\n",
       "      <td>1</td>\n",
       "      <td>1</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",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 524 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    patient-id  current-stay-days  cumulative-days-in-facility  cms-long-stay  \\\n",
       "2        10379                  7                            7              0   \n",
       "3        10385                  7                            7              0   \n",
       "14        5032                  7                            7              0   \n",
       "15        5032                  8                            8              0   \n",
       "16        5032                  7                            7              0   \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  ...  atc_5_N06AB10  atc_5_N06AX05  \\\n",
       "2                           0  ...              0              0   \n",
       "3                           1  ...              0              0   \n",
       "14                          1  ...              0              0   \n",
       "15                          1  ...              0              0   \n",
       "16                          1  ...              0              0   \n",
       "\n",
       "    atc_5_N06AX11  atc_5_N06AX12  atc_5_N06AX16  atc_5_N06AX21  atc_5_N06AX24  \\\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_N05AA01  atc_5_N05BA05  atc_5_N06AX23  \n",
       "2               0              0              0  \n",
       "3               0              0              0  \n",
       "14              0              0              0  \n",
       "15              0              0              0  \n",
       "16              0              0              0  \n",
       "\n",
       "[5 rows x 524 columns]"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test_use.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "# The main difference in this experiment from experiment 1 \n",
    "# is that values are imputed so there are no missing values to drop. \n",
    "# This adds a bunch of features\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": 26,
   "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": 27,
   "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": 28,
   "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": 29,
   "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": 30,
   "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": 31,
   "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": 32,
   "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": 33,
   "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": 34,
   "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/exp2_train_transformed_df.csv', index = False, header=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(5301, 526)"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_transformed.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2283, 526)"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test_transformed.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "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": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "525\n"
     ]
    }
   ],
   "source": [
    "n_features = X_train_list[0][0].shape[0]\n",
    "print(n_features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "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": 40,
   "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": 41,
   "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": 42,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING: Logging before flag parsing goes to stderr.\n",
      "W0424 10:03:29.885380 140366486800192 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 10:03:29.887990 140366486800192 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 10:03:29.903437 140366486800192 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 10:03:29.947661 140366486800192 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 10:03:29.957343 140366486800192 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 10:03:30.079508 140366486800192 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 10:03:30.134725 140366486800192 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 10:03:30.154075 140366486800192 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 [==============================] - 2s 690us/step - loss: 0.4843 - binary_accuracy: 0.8148\n",
      "Epoch 2/175\n",
      "2193/2193 [==============================] - 1s 391us/step - loss: 0.4357 - binary_accuracy: 0.8248\n",
      "Epoch 3/175\n",
      "2193/2193 [==============================] - 1s 385us/step - loss: 0.4409 - binary_accuracy: 0.8232\n",
      "Epoch 4/175\n",
      "2193/2193 [==============================] - 1s 402us/step - loss: 0.4391 - binary_accuracy: 0.8235\n",
      "Epoch 5/175\n",
      "2193/2193 [==============================] - 1s 383us/step - loss: 0.4272 - binary_accuracy: 0.8238\n",
      "Epoch 6/175\n",
      "2193/2193 [==============================] - 1s 385us/step - loss: 0.4315 - binary_accuracy: 0.8264\n",
      "Epoch 7/175\n",
      "2193/2193 [==============================] - 1s 408us/step - loss: 0.4314 - binary_accuracy: 0.8220\n",
      "Epoch 8/175\n",
      "2193/2193 [==============================] - 1s 404us/step - loss: 0.4269 - binary_accuracy: 0.8226\n",
      "Epoch 9/175\n",
      "2193/2193 [==============================] - 1s 391us/step - loss: 0.4258 - binary_accuracy: 0.8218\n",
      "Epoch 10/175\n",
      "2193/2193 [==============================] - 1s 429us/step - loss: 0.4199 - binary_accuracy: 0.8227\n",
      "Epoch 11/175\n",
      "2193/2193 [==============================] - 1s 399us/step - loss: 0.4199 - binary_accuracy: 0.8215\n",
      "Epoch 12/175\n",
      "2193/2193 [==============================] - 1s 397us/step - loss: 0.4165 - binary_accuracy: 0.8264\n",
      "Epoch 13/175\n",
      "2193/2193 [==============================] - 1s 415us/step - loss: 0.4225 - binary_accuracy: 0.8217\n",
      "Epoch 14/175\n",
      "2193/2193 [==============================] - 1s 380us/step - loss: 0.4255 - binary_accuracy: 0.8214\n",
      "Epoch 15/175\n",
      "2193/2193 [==============================] - 1s 386us/step - loss: 0.4200 - binary_accuracy: 0.8253\n",
      "Epoch 16/175\n",
      "2193/2193 [==============================] - 1s 384us/step - loss: 0.4214 - binary_accuracy: 0.8253\n",
      "Epoch 17/175\n",
      "2193/2193 [==============================] - 1s 423us/step - loss: 0.4293 - binary_accuracy: 0.8208\n",
      "Epoch 18/175\n",
      "2193/2193 [==============================] - 1s 384us/step - loss: 0.4255 - binary_accuracy: 0.8224\n",
      "Epoch 19/175\n",
      "2193/2193 [==============================] - 1s 377us/step - loss: 0.4142 - binary_accuracy: 0.8254\n",
      "Epoch 20/175\n",
      "2193/2193 [==============================] - 1s 389us/step - loss: 0.4241 - binary_accuracy: 0.8223\n",
      "Epoch 21/175\n",
      "2193/2193 [==============================] - 1s 381us/step - loss: 0.4221 - binary_accuracy: 0.8274\n",
      "Epoch 22/175\n",
      "2193/2193 [==============================] - 1s 381us/step - loss: 0.4218 - binary_accuracy: 0.8258\n",
      "Epoch 23/175\n",
      "2193/2193 [==============================] - 1s 380us/step - loss: 0.4235 - binary_accuracy: 0.8255\n",
      "Epoch 24/175\n",
      "2193/2193 [==============================] - 1s 379us/step - loss: 0.4137 - binary_accuracy: 0.8284\n",
      "Epoch 25/175\n",
      "2193/2193 [==============================] - 1s 378us/step - loss: 0.4219 - binary_accuracy: 0.8227\n",
      "Epoch 26/175\n",
      "2193/2193 [==============================] - 1s 379us/step - loss: 0.4095 - binary_accuracy: 0.8298\n",
      "Epoch 27/175\n",
      "2193/2193 [==============================] - 1s 379us/step - loss: 0.4145 - binary_accuracy: 0.8280\n",
      "Epoch 28/175\n",
      "2193/2193 [==============================] - 1s 382us/step - loss: 0.4176 - binary_accuracy: 0.8234\n",
      "Epoch 29/175\n",
      "2193/2193 [==============================] - 1s 383us/step - loss: 0.4163 - binary_accuracy: 0.8233\n",
      "Epoch 30/175\n",
      "2193/2193 [==============================] - 1s 383us/step - loss: 0.4177 - binary_accuracy: 0.8242\n",
      "Epoch 31/175\n",
      "2193/2193 [==============================] - 1s 379us/step - loss: 0.4180 - binary_accuracy: 0.8229\n",
      "Epoch 32/175\n",
      "2193/2193 [==============================] - 1s 377us/step - loss: 0.4140 - binary_accuracy: 0.8297\n",
      "Epoch 33/175\n",
      "2193/2193 [==============================] - 1s 379us/step - loss: 0.4149 - binary_accuracy: 0.8276\n",
      "Epoch 34/175\n",
      "2193/2193 [==============================] - 1s 379us/step - loss: 0.4117 - binary_accuracy: 0.8297\n",
      "Epoch 35/175\n",
      "2193/2193 [==============================] - 1s 377us/step - loss: 0.4157 - binary_accuracy: 0.8251\n",
      "Epoch 36/175\n",
      "2193/2193 [==============================] - 1s 380us/step - loss: 0.4117 - binary_accuracy: 0.8272\n",
      "Epoch 37/175\n",
      "2193/2193 [==============================] - 1s 381us/step - loss: 0.4147 - binary_accuracy: 0.8262\n",
      "Epoch 38/175\n",
      "2193/2193 [==============================] - 1s 381us/step - loss: 0.4169 - binary_accuracy: 0.8219\n",
      "Epoch 39/175\n",
      "2193/2193 [==============================] - 1s 382us/step - loss: 0.4149 - binary_accuracy: 0.8283\n",
      "Epoch 40/175\n",
      "2193/2193 [==============================] - 1s 381us/step - loss: 0.4131 - binary_accuracy: 0.8271\n",
      "Epoch 41/175\n",
      "2193/2193 [==============================] - 1s 383us/step - loss: 0.4143 - binary_accuracy: 0.8246\n",
      "Epoch 42/175\n",
      "2193/2193 [==============================] - 1s 382us/step - loss: 0.4204 - binary_accuracy: 0.8247\n",
      "Epoch 43/175\n",
      "2193/2193 [==============================] - 1s 383us/step - loss: 0.4105 - binary_accuracy: 0.8286\n",
      "Epoch 44/175\n",
      "2193/2193 [==============================] - 1s 378us/step - loss: 0.4083 - binary_accuracy: 0.8270\n",
      "Epoch 45/175\n",
      "2193/2193 [==============================] - 1s 379us/step - loss: 0.4162 - binary_accuracy: 0.8291\n",
      "Epoch 46/175\n",
      "2193/2193 [==============================] - 1s 384us/step - loss: 0.4108 - binary_accuracy: 0.8285\n",
      "Epoch 47/175\n",
      "2193/2193 [==============================] - 1s 413us/step - loss: 0.4153 - binary_accuracy: 0.8269\n",
      "Epoch 48/175\n",
      "2193/2193 [==============================] - 1s 380us/step - loss: 0.4161 - binary_accuracy: 0.8228\n",
      "Epoch 49/175\n",
      "2193/2193 [==============================] - 1s 383us/step - loss: 0.4118 - binary_accuracy: 0.8275\n",
      "Epoch 50/175\n",
      "2193/2193 [==============================] - 1s 382us/step - loss: 0.4143 - binary_accuracy: 0.8281\n",
      "Epoch 51/175\n",
      "2193/2193 [==============================] - 1s 381us/step - loss: 0.4171 - binary_accuracy: 0.8254\n",
      "Epoch 52/175\n",
      "2193/2193 [==============================] - 1s 380us/step - loss: 0.4118 - binary_accuracy: 0.8256\n",
      "Epoch 53/175\n",
      "2193/2193 [==============================] - 1s 380us/step - loss: 0.4039 - binary_accuracy: 0.8290\n",
      "Epoch 54/175\n",
      "2193/2193 [==============================] - 1s 387us/step - loss: 0.4066 - binary_accuracy: 0.8284\n",
      "Epoch 55/175\n",
      "2193/2193 [==============================] - 1s 381us/step - loss: 0.4096 - binary_accuracy: 0.8282\n",
      "Epoch 56/175\n",
      "2193/2193 [==============================] - 1s 379us/step - loss: 0.4194 - binary_accuracy: 0.8256\n",
      "Epoch 57/175\n",
      "2193/2193 [==============================] - 1s 375us/step - loss: 0.4152 - binary_accuracy: 0.8260\n",
      "Epoch 58/175\n",
      "2193/2193 [==============================] - 1s 379us/step - loss: 0.4122 - binary_accuracy: 0.8306\n",
      "Epoch 59/175\n",
      "2193/2193 [==============================] - 1s 389us/step - loss: 0.4135 - binary_accuracy: 0.8281\n",
      "Epoch 60/175\n",
      "2193/2193 [==============================] - 1s 378us/step - loss: 0.4117 - binary_accuracy: 0.8278\n",
      "Epoch 61/175\n",
      "2193/2193 [==============================] - 1s 379us/step - loss: 0.4114 - binary_accuracy: 0.8278\n",
      "Epoch 62/175\n",
      "2193/2193 [==============================] - 1s 379us/step - loss: 0.4170 - binary_accuracy: 0.8242\n",
      "Epoch 63/175\n",
      "2193/2193 [==============================] - 1s 380us/step - loss: 0.4099 - binary_accuracy: 0.8288\n",
      "Epoch 64/175\n",
      "2193/2193 [==============================] - 1s 376us/step - loss: 0.4110 - binary_accuracy: 0.8326\n",
      "Epoch 65/175\n",
      "2193/2193 [==============================] - 1s 377us/step - loss: 0.4141 - binary_accuracy: 0.8266\n",
      "Epoch 66/175\n",
      "2193/2193 [==============================] - 1s 382us/step - loss: 0.4127 - binary_accuracy: 0.8290\n",
      "Epoch 67/175\n",
      "2193/2193 [==============================] - 1s 375us/step - loss: 0.4131 - binary_accuracy: 0.8232\n",
      "Epoch 68/175\n",
      "2193/2193 [==============================] - 1s 378us/step - loss: 0.4091 - binary_accuracy: 0.8267\n",
      "Epoch 69/175\n",
      "2193/2193 [==============================] - 1s 379us/step - loss: 0.4175 - binary_accuracy: 0.8245\n",
      "Epoch 70/175\n",
      "2193/2193 [==============================] - 1s 377us/step - loss: 0.4090 - binary_accuracy: 0.8277\n",
      "Epoch 71/175\n",
      "2193/2193 [==============================] - 1s 381us/step - loss: 0.4142 - binary_accuracy: 0.8263\n",
      "Epoch 72/175\n",
      "2193/2193 [==============================] - 1s 380us/step - loss: 0.4126 - binary_accuracy: 0.8270\n",
      "Epoch 73/175\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2193/2193 [==============================] - 1s 380us/step - loss: 0.4056 - binary_accuracy: 0.8302\n",
      "Epoch 74/175\n",
      "2193/2193 [==============================] - 1s 377us/step - loss: 0.4081 - binary_accuracy: 0.8292\n",
      "Epoch 75/175\n",
      "2193/2193 [==============================] - 1s 395us/step - loss: 0.4062 - binary_accuracy: 0.8296\n",
      "Epoch 76/175\n",
      "2193/2193 [==============================] - 1s 381us/step - loss: 0.4151 - binary_accuracy: 0.8238\n",
      "Epoch 77/175\n",
      "2193/2193 [==============================] - 1s 381us/step - loss: 0.4048 - binary_accuracy: 0.8293\n",
      "Epoch 78/175\n",
      "2193/2193 [==============================] - 1s 381us/step - loss: 0.4083 - binary_accuracy: 0.8296\n",
      "Epoch 79/175\n",
      "2193/2193 [==============================] - 1s 385us/step - loss: 0.4141 - binary_accuracy: 0.8267\n",
      "Epoch 80/175\n",
      "2193/2193 [==============================] - 1s 381us/step - loss: 0.4058 - binary_accuracy: 0.8282\n",
      "Epoch 81/175\n",
      "2193/2193 [==============================] - 1s 398us/step - loss: 0.4129 - binary_accuracy: 0.8268\n",
      "Epoch 82/175\n",
      "2193/2193 [==============================] - 1s 385us/step - loss: 0.3990 - binary_accuracy: 0.8365\n",
      "Epoch 83/175\n",
      "2193/2193 [==============================] - 1s 380us/step - loss: 0.4117 - binary_accuracy: 0.8298\n",
      "Epoch 84/175\n",
      "2193/2193 [==============================] - 1s 382us/step - loss: 0.4030 - binary_accuracy: 0.8330\n",
      "Epoch 85/175\n",
      "2193/2193 [==============================] - 1s 409us/step - loss: 0.4113 - binary_accuracy: 0.8296\n",
      "Epoch 86/175\n",
      "2193/2193 [==============================] - 1s 386us/step - loss: 0.4078 - binary_accuracy: 0.8277\n",
      "Epoch 87/175\n",
      "2193/2193 [==============================] - 1s 377us/step - loss: 0.4110 - binary_accuracy: 0.8250\n",
      "Epoch 88/175\n",
      "2193/2193 [==============================] - 1s 376us/step - loss: 0.4129 - binary_accuracy: 0.8251\n",
      "Epoch 89/175\n",
      "2193/2193 [==============================] - 1s 377us/step - loss: 0.4144 - binary_accuracy: 0.8259\n",
      "Epoch 90/175\n",
      "2193/2193 [==============================] - 1s 384us/step - loss: 0.4052 - binary_accuracy: 0.8294\n",
      "Epoch 91/175\n",
      "2193/2193 [==============================] - 1s 382us/step - loss: 0.4087 - binary_accuracy: 0.8261\n",
      "Epoch 92/175\n",
      "2193/2193 [==============================] - 1s 387us/step - loss: 0.4067 - binary_accuracy: 0.8309\n",
      "Epoch 93/175\n",
      "2193/2193 [==============================] - 1s 390us/step - loss: 0.4060 - binary_accuracy: 0.8277\n",
      "Epoch 94/175\n",
      "2193/2193 [==============================] - 1s 379us/step - loss: 0.4093 - binary_accuracy: 0.8293\n",
      "Epoch 95/175\n",
      "2193/2193 [==============================] - 1s 386us/step - loss: 0.4098 - binary_accuracy: 0.8281\n",
      "Epoch 96/175\n",
      "2193/2193 [==============================] - 1s 382us/step - loss: 0.4091 - binary_accuracy: 0.8319\n",
      "Epoch 97/175\n",
      "2193/2193 [==============================] - 1s 387us/step - loss: 0.4086 - binary_accuracy: 0.8301\n",
      "Epoch 98/175\n",
      "2193/2193 [==============================] - 1s 384us/step - loss: 0.4054 - binary_accuracy: 0.8267\n",
      "Epoch 99/175\n",
      "2193/2193 [==============================] - 1s 383us/step - loss: 0.4177 - binary_accuracy: 0.8238\n",
      "Epoch 100/175\n",
      "2193/2193 [==============================] - 1s 383us/step - loss: 0.4083 - binary_accuracy: 0.8283\n",
      "Epoch 101/175\n",
      "2193/2193 [==============================] - 1s 387us/step - loss: 0.4126 - binary_accuracy: 0.8254\n",
      "Epoch 102/175\n",
      "2193/2193 [==============================] - 1s 378us/step - loss: 0.4075 - binary_accuracy: 0.8296\n",
      "Epoch 103/175\n",
      "2193/2193 [==============================] - 1s 381us/step - loss: 0.4023 - binary_accuracy: 0.8333\n",
      "Epoch 104/175\n",
      "2193/2193 [==============================] - 1s 381us/step - loss: 0.4015 - binary_accuracy: 0.8329\n",
      "Epoch 105/175\n",
      "2193/2193 [==============================] - 1s 380us/step - loss: 0.4006 - binary_accuracy: 0.8346\n",
      "Epoch 106/175\n",
      "2193/2193 [==============================] - 1s 378us/step - loss: 0.4072 - binary_accuracy: 0.8256\n",
      "Epoch 107/175\n",
      "2193/2193 [==============================] - 1s 383us/step - loss: 0.4112 - binary_accuracy: 0.8296\n",
      "Epoch 108/175\n",
      "2193/2193 [==============================] - 1s 382us/step - loss: 0.4031 - binary_accuracy: 0.8311\n",
      "Epoch 109/175\n",
      "2193/2193 [==============================] - 1s 383us/step - loss: 0.4047 - binary_accuracy: 0.8300\n",
      "Epoch 110/175\n",
      "2193/2193 [==============================] - 1s 380us/step - loss: 0.4066 - binary_accuracy: 0.8285\n",
      "Epoch 111/175\n",
      "2193/2193 [==============================] - 1s 380us/step - loss: 0.4064 - binary_accuracy: 0.8303\n",
      "Epoch 112/175\n",
      "2193/2193 [==============================] - 1s 381us/step - loss: 0.4105 - binary_accuracy: 0.8253\n",
      "Epoch 113/175\n",
      "2193/2193 [==============================] - 1s 384us/step - loss: 0.4054 - binary_accuracy: 0.8328\n",
      "Epoch 114/175\n",
      "2193/2193 [==============================] - 1s 387us/step - loss: 0.4044 - binary_accuracy: 0.8308\n",
      "Epoch 115/175\n",
      "2193/2193 [==============================] - 1s 382us/step - loss: 0.4141 - binary_accuracy: 0.8244\n",
      "Epoch 116/175\n",
      "2193/2193 [==============================] - 1s 381us/step - loss: 0.4072 - binary_accuracy: 0.8262\n",
      "Epoch 117/175\n",
      "2193/2193 [==============================] - 1s 389us/step - loss: 0.4023 - binary_accuracy: 0.8294\n",
      "Epoch 118/175\n",
      "2193/2193 [==============================] - 1s 382us/step - loss: 0.4063 - binary_accuracy: 0.8282\n",
      "Epoch 119/175\n",
      "2193/2193 [==============================] - 1s 382us/step - loss: 0.4080 - binary_accuracy: 0.8293\n",
      "Epoch 120/175\n",
      "2193/2193 [==============================] - 1s 381us/step - loss: 0.4049 - binary_accuracy: 0.8286\n",
      "Epoch 121/175\n",
      "2193/2193 [==============================] - 1s 379us/step - loss: 0.4063 - binary_accuracy: 0.8326\n",
      "Epoch 122/175\n",
      "2193/2193 [==============================] - 1s 382us/step - loss: 0.4030 - binary_accuracy: 0.8299\n",
      "Epoch 123/175\n",
      "2193/2193 [==============================] - 1s 379us/step - loss: 0.4061 - binary_accuracy: 0.8320\n",
      "Epoch 124/175\n",
      "2193/2193 [==============================] - 1s 374us/step - loss: 0.4051 - binary_accuracy: 0.8252\n",
      "Epoch 125/175\n",
      "2193/2193 [==============================] - 1s 373us/step - loss: 0.4023 - binary_accuracy: 0.8324\n",
      "Epoch 126/175\n",
      "2193/2193 [==============================] - 1s 378us/step - loss: 0.4077 - binary_accuracy: 0.8292\n",
      "Epoch 127/175\n",
      "2193/2193 [==============================] - 1s 371us/step - loss: 0.4096 - binary_accuracy: 0.8250\n",
      "Epoch 128/175\n",
      "2193/2193 [==============================] - 1s 377us/step - loss: 0.4027 - binary_accuracy: 0.8318\n",
      "Epoch 129/175\n",
      "2193/2193 [==============================] - 1s 378us/step - loss: 0.4032 - binary_accuracy: 0.8327\n",
      "Epoch 130/175\n",
      "2193/2193 [==============================] - 1s 373us/step - loss: 0.4066 - binary_accuracy: 0.8257\n",
      "Epoch 131/175\n",
      "2193/2193 [==============================] - 1s 373us/step - loss: 0.4064 - binary_accuracy: 0.8322\n",
      "Epoch 132/175\n",
      "2193/2193 [==============================] - 1s 370us/step - loss: 0.4123 - binary_accuracy: 0.8282\n",
      "Epoch 133/175\n",
      "2193/2193 [==============================] - 1s 369us/step - loss: 0.4034 - binary_accuracy: 0.8324\n",
      "Epoch 134/175\n",
      "2193/2193 [==============================] - 1s 375us/step - loss: 0.4000 - binary_accuracy: 0.8315\n",
      "Epoch 135/175\n",
      "2193/2193 [==============================] - 1s 370us/step - loss: 0.3985 - binary_accuracy: 0.8320\n",
      "Epoch 136/175\n",
      "2193/2193 [==============================] - 1s 387us/step - loss: 0.4081 - binary_accuracy: 0.8284\n",
      "Epoch 137/175\n",
      "2193/2193 [==============================] - 1s 387us/step - loss: 0.4069 - binary_accuracy: 0.8312\n",
      "Epoch 138/175\n",
      "2193/2193 [==============================] - 1s 372us/step - loss: 0.4101 - binary_accuracy: 0.8292\n",
      "Epoch 139/175\n",
      "2193/2193 [==============================] - 1s 371us/step - loss: 0.4015 - binary_accuracy: 0.8288\n",
      "Epoch 140/175\n",
      "2193/2193 [==============================] - 1s 372us/step - loss: 0.4098 - binary_accuracy: 0.8285\n",
      "Epoch 141/175\n",
      "2193/2193 [==============================] - 1s 372us/step - loss: 0.4076 - binary_accuracy: 0.8281\n",
      "Epoch 142/175\n",
      "2193/2193 [==============================] - 1s 382us/step - loss: 0.4051 - binary_accuracy: 0.8301\n",
      "Epoch 143/175\n",
      "2193/2193 [==============================] - 1s 369us/step - loss: 0.4038 - binary_accuracy: 0.8354\n",
      "Epoch 144/175\n",
      "2193/2193 [==============================] - 1s 373us/step - loss: 0.4010 - binary_accuracy: 0.8304\n",
      "Epoch 145/175\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2193/2193 [==============================] - 1s 372us/step - loss: 0.4031 - binary_accuracy: 0.8325\n",
      "Epoch 146/175\n",
      "2193/2193 [==============================] - 1s 373us/step - loss: 0.4047 - binary_accuracy: 0.8265\n",
      "Epoch 147/175\n",
      "2193/2193 [==============================] - 1s 379us/step - loss: 0.4042 - binary_accuracy: 0.8266\n",
      "Epoch 148/175\n",
      "2193/2193 [==============================] - 1s 386us/step - loss: 0.4023 - binary_accuracy: 0.8306\n",
      "Epoch 149/175\n",
      "2193/2193 [==============================] - 1s 373us/step - loss: 0.4074 - binary_accuracy: 0.8328\n",
      "Epoch 150/175\n",
      "2193/2193 [==============================] - 1s 374us/step - loss: 0.4107 - binary_accuracy: 0.8306\n",
      "Epoch 151/175\n",
      "2193/2193 [==============================] - 1s 374us/step - loss: 0.3983 - binary_accuracy: 0.8339\n",
      "Epoch 152/175\n",
      "2193/2193 [==============================] - 1s 373us/step - loss: 0.4045 - binary_accuracy: 0.8291\n",
      "Epoch 153/175\n",
      "2193/2193 [==============================] - 1s 376us/step - loss: 0.4099 - binary_accuracy: 0.8272\n",
      "Epoch 154/175\n",
      "2193/2193 [==============================] - 1s 377us/step - loss: 0.4056 - binary_accuracy: 0.8274\n",
      "Epoch 155/175\n",
      "2193/2193 [==============================] - 1s 378us/step - loss: 0.4086 - binary_accuracy: 0.8231\n",
      "Epoch 156/175\n",
      "2193/2193 [==============================] - 1s 376us/step - loss: 0.4040 - binary_accuracy: 0.8284\n",
      "Epoch 157/175\n",
      "2193/2193 [==============================] - 1s 371us/step - loss: 0.4034 - binary_accuracy: 0.8280\n",
      "Epoch 158/175\n",
      "2193/2193 [==============================] - 1s 370us/step - loss: 0.4033 - binary_accuracy: 0.8327\n",
      "Epoch 159/175\n",
      "2193/2193 [==============================] - 1s 372us/step - loss: 0.3997 - binary_accuracy: 0.8329\n",
      "Epoch 160/175\n",
      "2193/2193 [==============================] - 1s 382us/step - loss: 0.3995 - binary_accuracy: 0.8335\n",
      "Epoch 161/175\n",
      "2193/2193 [==============================] - 1s 388us/step - loss: 0.4051 - binary_accuracy: 0.8268\n",
      "Epoch 162/175\n",
      "2193/2193 [==============================] - 1s 378us/step - loss: 0.4025 - binary_accuracy: 0.8307\n",
      "Epoch 163/175\n",
      "2193/2193 [==============================] - 1s 372us/step - loss: 0.4048 - binary_accuracy: 0.8296\n",
      "Epoch 164/175\n",
      "2193/2193 [==============================] - 1s 371us/step - loss: 0.4094 - binary_accuracy: 0.8250\n",
      "Epoch 165/175\n",
      "2193/2193 [==============================] - 1s 370us/step - loss: 0.4061 - binary_accuracy: 0.8299\n",
      "Epoch 166/175\n",
      "2193/2193 [==============================] - 1s 378us/step - loss: 0.4015 - binary_accuracy: 0.8345\n",
      "Epoch 167/175\n",
      "2193/2193 [==============================] - 1s 371us/step - loss: 0.4041 - binary_accuracy: 0.8338\n",
      "Epoch 168/175\n",
      "2193/2193 [==============================] - 1s 371us/step - loss: 0.4001 - binary_accuracy: 0.8322\n",
      "Epoch 169/175\n",
      "2193/2193 [==============================] - 1s 369us/step - loss: 0.4089 - binary_accuracy: 0.8293\n",
      "Epoch 170/175\n",
      "2193/2193 [==============================] - 1s 372us/step - loss: 0.4038 - binary_accuracy: 0.8343\n",
      "Epoch 171/175\n",
      "2193/2193 [==============================] - 1s 371us/step - loss: 0.4005 - binary_accuracy: 0.8299\n",
      "Epoch 172/175\n",
      "2193/2193 [==============================] - 1s 373us/step - loss: 0.4058 - binary_accuracy: 0.8311\n",
      "Epoch 173/175\n",
      "2193/2193 [==============================] - 1s 375us/step - loss: 0.4070 - binary_accuracy: 0.8285\n",
      "Epoch 174/175\n",
      "2193/2193 [==============================] - 1s 377us/step - loss: 0.4077 - binary_accuracy: 0.8303\n",
      "Epoch 175/175\n",
      "2193/2193 [==============================] - 1s 379us/step - loss: 0.4045 - binary_accuracy: 0.8306\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:  0.7164783721365708\n",
      "f1:  0.4327301337529504\n",
      "threshold:  0.15927842\n",
      "precision:  0.32816229116945106\n",
      "recall:  0.6351039260969977\n",
      "accuracy:  0.6841874726237407\n",
      "specificity:  0.6956756756756757\n",
      "CPU times: user 7min 58s, sys: 1min 7s, total: 9min 5s\n",
      "Wall time: 2min 29s\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": 43,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/225\n",
      "1755/1755 [==============================] - 1s 749us/step - loss: 0.4789 - binary_accuracy: 0.8141\n",
      "Epoch 2/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.4507 - binary_accuracy: 0.8200\n",
      "Epoch 3/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.4379 - binary_accuracy: 0.8220\n",
      "Epoch 4/225\n",
      "1755/1755 [==============================] - 1s 401us/step - loss: 0.4345 - binary_accuracy: 0.8157\n",
      "Epoch 5/225\n",
      "1755/1755 [==============================] - 1s 404us/step - loss: 0.4314 - binary_accuracy: 0.8214\n",
      "Epoch 6/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.4274 - binary_accuracy: 0.8204\n",
      "Epoch 7/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.4290 - binary_accuracy: 0.8239\n",
      "Epoch 8/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.4310 - binary_accuracy: 0.8206\n",
      "Epoch 9/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.4264 - binary_accuracy: 0.8205\n",
      "Epoch 10/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.4227 - binary_accuracy: 0.8262\n",
      "Epoch 11/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.4161 - binary_accuracy: 0.8286\n",
      "Epoch 12/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.4186 - binary_accuracy: 0.8248\n",
      "Epoch 13/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.4233 - binary_accuracy: 0.8232\n",
      "Epoch 14/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.4210 - binary_accuracy: 0.8222\n",
      "Epoch 15/225\n",
      "1755/1755 [==============================] - 1s 408us/step - loss: 0.4185 - binary_accuracy: 0.8240\n",
      "Epoch 16/225\n",
      "1755/1755 [==============================] - 1s 404us/step - loss: 0.4211 - binary_accuracy: 0.8221\n",
      "Epoch 17/225\n",
      "1755/1755 [==============================] - 1s 403us/step - loss: 0.4095 - binary_accuracy: 0.8297\n",
      "Epoch 18/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.4171 - binary_accuracy: 0.8274\n",
      "Epoch 19/225\n",
      "1755/1755 [==============================] - 1s 405us/step - loss: 0.4153 - binary_accuracy: 0.8295\n",
      "Epoch 20/225\n",
      "1755/1755 [==============================] - 1s 403us/step - loss: 0.4115 - binary_accuracy: 0.8281\n",
      "Epoch 21/225\n",
      "1755/1755 [==============================] - 1s 404us/step - loss: 0.4166 - binary_accuracy: 0.8253\n",
      "Epoch 22/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.4201 - binary_accuracy: 0.8207\n",
      "Epoch 23/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.4111 - binary_accuracy: 0.8287\n",
      "Epoch 24/225\n",
      "1755/1755 [==============================] - 1s 401us/step - loss: 0.4152 - binary_accuracy: 0.8251\n",
      "Epoch 25/225\n",
      "1755/1755 [==============================] - 1s 413us/step - loss: 0.4083 - binary_accuracy: 0.8267\n",
      "Epoch 26/225\n",
      "1755/1755 [==============================] - 1s 406us/step - loss: 0.4187 - binary_accuracy: 0.8281\n",
      "Epoch 27/225\n",
      "1755/1755 [==============================] - 1s 403us/step - loss: 0.4136 - binary_accuracy: 0.8248\n",
      "Epoch 28/225\n",
      "1755/1755 [==============================] - 1s 404us/step - loss: 0.4115 - binary_accuracy: 0.8309\n",
      "Epoch 29/225\n",
      "1755/1755 [==============================] - 1s 414us/step - loss: 0.4161 - binary_accuracy: 0.8261\n",
      "Epoch 30/225\n",
      "1755/1755 [==============================] - 1s 407us/step - loss: 0.4168 - binary_accuracy: 0.8271\n",
      "Epoch 31/225\n",
      "1755/1755 [==============================] - 1s 411us/step - loss: 0.4174 - binary_accuracy: 0.8253\n",
      "Epoch 32/225\n",
      "1755/1755 [==============================] - 1s 408us/step - loss: 0.4175 - binary_accuracy: 0.8238\n",
      "Epoch 33/225\n",
      "1755/1755 [==============================] - 1s 407us/step - loss: 0.4169 - binary_accuracy: 0.8246\n",
      "Epoch 34/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.4093 - binary_accuracy: 0.8271\n",
      "Epoch 35/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.4158 - binary_accuracy: 0.8253\n",
      "Epoch 36/225\n",
      "1755/1755 [==============================] - 1s 411us/step - loss: 0.4075 - binary_accuracy: 0.8294\n",
      "Epoch 37/225\n",
      "1755/1755 [==============================] - 1s 403us/step - loss: 0.4103 - binary_accuracy: 0.8288\n",
      "Epoch 38/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.4142 - binary_accuracy: 0.8261\n",
      "Epoch 39/225\n",
      "1755/1755 [==============================] - 1s 406us/step - loss: 0.4158 - binary_accuracy: 0.8243\n",
      "Epoch 40/225\n",
      "1755/1755 [==============================] - 1s 401us/step - loss: 0.4154 - binary_accuracy: 0.8278\n",
      "Epoch 41/225\n",
      "1755/1755 [==============================] - 1s 417us/step - loss: 0.4127 - binary_accuracy: 0.8272\n",
      "Epoch 42/225\n",
      "1755/1755 [==============================] - 1s 416us/step - loss: 0.4087 - binary_accuracy: 0.8275\n",
      "Epoch 43/225\n",
      "1755/1755 [==============================] - 1s 415us/step - loss: 0.4133 - binary_accuracy: 0.8230\n",
      "Epoch 44/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.4127 - binary_accuracy: 0.8290\n",
      "Epoch 45/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.4028 - binary_accuracy: 0.8331\n",
      "Epoch 46/225\n",
      "1755/1755 [==============================] - 1s 401us/step - loss: 0.4092 - binary_accuracy: 0.8271\n",
      "Epoch 47/225\n",
      "1755/1755 [==============================] - 1s 418us/step - loss: 0.4090 - binary_accuracy: 0.8302\n",
      "Epoch 48/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.4184 - binary_accuracy: 0.8240\n",
      "Epoch 49/225\n",
      "1755/1755 [==============================] - 1s 405us/step - loss: 0.4043 - binary_accuracy: 0.8298\n",
      "Epoch 50/225\n",
      "1755/1755 [==============================] - 1s 415us/step - loss: 0.4128 - binary_accuracy: 0.8249\n",
      "Epoch 51/225\n",
      "1755/1755 [==============================] - 1s 403us/step - loss: 0.4116 - binary_accuracy: 0.8278\n",
      "Epoch 52/225\n",
      "1755/1755 [==============================] - 1s 401us/step - loss: 0.4086 - binary_accuracy: 0.8268\n",
      "Epoch 53/225\n",
      "1755/1755 [==============================] - 1s 403us/step - loss: 0.4071 - binary_accuracy: 0.8302\n",
      "Epoch 54/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.4098 - binary_accuracy: 0.8276\n",
      "Epoch 55/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.4113 - binary_accuracy: 0.8241\n",
      "Epoch 56/225\n",
      "1755/1755 [==============================] - 1s 406us/step - loss: 0.4117 - binary_accuracy: 0.8260\n",
      "Epoch 57/225\n",
      "1755/1755 [==============================] - 1s 401us/step - loss: 0.4151 - binary_accuracy: 0.8255\n",
      "Epoch 58/225\n",
      "1755/1755 [==============================] - 1s 405us/step - loss: 0.4000 - binary_accuracy: 0.8344\n",
      "Epoch 59/225\n",
      "1755/1755 [==============================] - 1s 407us/step - loss: 0.4120 - binary_accuracy: 0.8283\n",
      "Epoch 60/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.4075 - binary_accuracy: 0.8295\n",
      "Epoch 61/225\n",
      "1755/1755 [==============================] - 1s 436us/step - loss: 0.4036 - binary_accuracy: 0.8307\n",
      "Epoch 62/225\n",
      "1755/1755 [==============================] - 1s 427us/step - loss: 0.4141 - binary_accuracy: 0.8284\n",
      "Epoch 63/225\n",
      "1755/1755 [==============================] - 1s 425us/step - loss: 0.4100 - binary_accuracy: 0.8282\n",
      "Epoch 64/225\n",
      "1755/1755 [==============================] - 1s 432us/step - loss: 0.4049 - binary_accuracy: 0.8295\n",
      "Epoch 65/225\n",
      "1755/1755 [==============================] - 1s 426us/step - loss: 0.4061 - binary_accuracy: 0.8262\n",
      "Epoch 66/225\n",
      "1755/1755 [==============================] - 1s 424us/step - loss: 0.4024 - binary_accuracy: 0.8325\n",
      "Epoch 67/225\n",
      "1755/1755 [==============================] - 1s 423us/step - loss: 0.4069 - binary_accuracy: 0.8257\n",
      "Epoch 68/225\n",
      "1755/1755 [==============================] - 1s 419us/step - loss: 0.4071 - binary_accuracy: 0.8316\n",
      "Epoch 69/225\n",
      "1755/1755 [==============================] - 1s 407us/step - loss: 0.4089 - binary_accuracy: 0.8294\n",
      "Epoch 70/225\n",
      "1755/1755 [==============================] - 1s 406us/step - loss: 0.4090 - binary_accuracy: 0.8291\n",
      "Epoch 71/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.4037 - binary_accuracy: 0.8377\n",
      "Epoch 72/225\n",
      "1755/1755 [==============================] - 1s 404us/step - loss: 0.4059 - binary_accuracy: 0.8311\n",
      "Epoch 73/225\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.4086 - binary_accuracy: 0.8297\n",
      "Epoch 74/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.4107 - binary_accuracy: 0.8293\n",
      "Epoch 75/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.4126 - binary_accuracy: 0.8254\n",
      "Epoch 76/225\n",
      "1755/1755 [==============================] - 1s 391us/step - loss: 0.4059 - binary_accuracy: 0.8291\n",
      "Epoch 77/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.4166 - binary_accuracy: 0.8290\n",
      "Epoch 78/225\n",
      "1755/1755 [==============================] - 1s 407us/step - loss: 0.4047 - binary_accuracy: 0.8313\n",
      "Epoch 79/225\n",
      "1755/1755 [==============================] - 1s 403us/step - loss: 0.4072 - binary_accuracy: 0.8287\n",
      "Epoch 80/225\n",
      "1755/1755 [==============================] - 1s 403us/step - loss: 0.4007 - binary_accuracy: 0.8345\n",
      "Epoch 81/225\n",
      "1755/1755 [==============================] - 1s 404us/step - loss: 0.4026 - binary_accuracy: 0.8320\n",
      "Epoch 82/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.4046 - binary_accuracy: 0.8295\n",
      "Epoch 83/225\n",
      "1755/1755 [==============================] - 1s 401us/step - loss: 0.4100 - binary_accuracy: 0.8289\n",
      "Epoch 84/225\n",
      "1755/1755 [==============================] - 1s 404us/step - loss: 0.4074 - binary_accuracy: 0.8301\n",
      "Epoch 85/225\n",
      "1755/1755 [==============================] - 1s 403us/step - loss: 0.4019 - binary_accuracy: 0.8314\n",
      "Epoch 86/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.4064 - binary_accuracy: 0.8303\n",
      "Epoch 87/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.4042 - binary_accuracy: 0.8254\n",
      "Epoch 88/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.3997 - binary_accuracy: 0.8305\n",
      "Epoch 89/225\n",
      "1755/1755 [==============================] - 1s 392us/step - loss: 0.4061 - binary_accuracy: 0.8278\n",
      "Epoch 90/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.4081 - binary_accuracy: 0.8280\n",
      "Epoch 91/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.4018 - binary_accuracy: 0.8334\n",
      "Epoch 92/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.4029 - binary_accuracy: 0.8299\n",
      "Epoch 93/225\n",
      "1755/1755 [==============================] - 1s 393us/step - loss: 0.4062 - binary_accuracy: 0.8271\n",
      "Epoch 94/225\n",
      "1755/1755 [==============================] - 1s 404us/step - loss: 0.4031 - binary_accuracy: 0.8278\n",
      "Epoch 95/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.4112 - binary_accuracy: 0.8294\n",
      "Epoch 96/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.4004 - binary_accuracy: 0.8282\n",
      "Epoch 97/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.4041 - binary_accuracy: 0.8321\n",
      "Epoch 98/225\n",
      "1755/1755 [==============================] - 1s 411us/step - loss: 0.4058 - binary_accuracy: 0.8254\n",
      "Epoch 99/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.4070 - binary_accuracy: 0.8280\n",
      "Epoch 100/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.4069 - binary_accuracy: 0.8299\n",
      "Epoch 101/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.4053 - binary_accuracy: 0.8280\n",
      "Epoch 102/225\n",
      "1755/1755 [==============================] - 1s 406us/step - loss: 0.4067 - binary_accuracy: 0.8271\n",
      "Epoch 103/225\n",
      "1755/1755 [==============================] - 1s 406us/step - loss: 0.4117 - binary_accuracy: 0.8252\n",
      "Epoch 104/225\n",
      "1755/1755 [==============================] - 1s 408us/step - loss: 0.4075 - binary_accuracy: 0.8296\n",
      "Epoch 105/225\n",
      "1755/1755 [==============================] - 1s 403us/step - loss: 0.4063 - binary_accuracy: 0.8292\n",
      "Epoch 106/225\n",
      "1755/1755 [==============================] - 1s 403us/step - loss: 0.4105 - binary_accuracy: 0.8271\n",
      "Epoch 107/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.4006 - binary_accuracy: 0.8312\n",
      "Epoch 108/225\n",
      "1755/1755 [==============================] - 1s 407us/step - loss: 0.4039 - binary_accuracy: 0.8311\n",
      "Epoch 109/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.4108 - binary_accuracy: 0.8268\n",
      "Epoch 110/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.4100 - binary_accuracy: 0.8243\n",
      "Epoch 111/225\n",
      "1755/1755 [==============================] - 1s 408us/step - loss: 0.3979 - binary_accuracy: 0.8345\n",
      "Epoch 112/225\n",
      "1755/1755 [==============================] - 1s 403us/step - loss: 0.4010 - binary_accuracy: 0.8303\n",
      "Epoch 113/225\n",
      "1755/1755 [==============================] - 1s 441us/step - loss: 0.4092 - binary_accuracy: 0.8298\n",
      "Epoch 114/225\n",
      "1755/1755 [==============================] - 1s 417us/step - loss: 0.4000 - binary_accuracy: 0.8349\n",
      "Epoch 115/225\n",
      "1755/1755 [==============================] - 1s 409us/step - loss: 0.4018 - binary_accuracy: 0.8283\n",
      "Epoch 116/225\n",
      "1755/1755 [==============================] - 1s 405us/step - loss: 0.4050 - binary_accuracy: 0.8292\n",
      "Epoch 117/225\n",
      "1755/1755 [==============================] - 1s 418us/step - loss: 0.4033 - binary_accuracy: 0.8314\n",
      "Epoch 118/225\n",
      "1755/1755 [==============================] - 1s 418us/step - loss: 0.4060 - binary_accuracy: 0.8275\n",
      "Epoch 119/225\n",
      "1755/1755 [==============================] - 1s 417us/step - loss: 0.4049 - binary_accuracy: 0.8253\n",
      "Epoch 120/225\n",
      "1755/1755 [==============================] - 1s 413us/step - loss: 0.3991 - binary_accuracy: 0.8368\n",
      "Epoch 121/225\n",
      "1755/1755 [==============================] - 1s 417us/step - loss: 0.4050 - binary_accuracy: 0.8274\n",
      "Epoch 122/225\n",
      "1755/1755 [==============================] - 1s 419us/step - loss: 0.4072 - binary_accuracy: 0.8268\n",
      "Epoch 123/225\n",
      "1755/1755 [==============================] - 1s 403us/step - loss: 0.4029 - binary_accuracy: 0.8327\n",
      "Epoch 124/225\n",
      "1755/1755 [==============================] - 1s 404us/step - loss: 0.4030 - binary_accuracy: 0.8330\n",
      "Epoch 125/225\n",
      "1755/1755 [==============================] - 1s 403us/step - loss: 0.4001 - binary_accuracy: 0.8301\n",
      "Epoch 126/225\n",
      "1755/1755 [==============================] - 1s 404us/step - loss: 0.4032 - binary_accuracy: 0.8256\n",
      "Epoch 127/225\n",
      "1755/1755 [==============================] - 1s 406us/step - loss: 0.4034 - binary_accuracy: 0.8301\n",
      "Epoch 128/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.4051 - binary_accuracy: 0.8288\n",
      "Epoch 129/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.4019 - binary_accuracy: 0.8302\n",
      "Epoch 130/225\n",
      "1755/1755 [==============================] - 1s 406us/step - loss: 0.4026 - binary_accuracy: 0.8350\n",
      "Epoch 131/225\n",
      "1755/1755 [==============================] - 1s 393us/step - loss: 0.4048 - binary_accuracy: 0.8261\n",
      "Epoch 132/225\n",
      "1755/1755 [==============================] - 1s 395us/step - loss: 0.4002 - binary_accuracy: 0.8288\n",
      "Epoch 133/225\n",
      "1755/1755 [==============================] - 1s 395us/step - loss: 0.4045 - binary_accuracy: 0.8319\n",
      "Epoch 134/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.4007 - binary_accuracy: 0.8376\n",
      "Epoch 135/225\n",
      "1755/1755 [==============================] - 1s 391us/step - loss: 0.3989 - binary_accuracy: 0.8322\n",
      "Epoch 136/225\n",
      "1755/1755 [==============================] - 1s 391us/step - loss: 0.4026 - binary_accuracy: 0.8299\n",
      "Epoch 137/225\n",
      "1755/1755 [==============================] - 1s 410us/step - loss: 0.3948 - binary_accuracy: 0.8356\n",
      "Epoch 138/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.4072 - binary_accuracy: 0.8262\n",
      "Epoch 139/225\n",
      "1755/1755 [==============================] - 1s 395us/step - loss: 0.4018 - binary_accuracy: 0.8311\n",
      "Epoch 140/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.4029 - binary_accuracy: 0.8337\n",
      "Epoch 141/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.3958 - binary_accuracy: 0.8292\n",
      "Epoch 142/225\n",
      "1755/1755 [==============================] - 1s 394us/step - loss: 0.3981 - binary_accuracy: 0.8311\n",
      "Epoch 143/225\n",
      "1755/1755 [==============================] - 1s 394us/step - loss: 0.3985 - binary_accuracy: 0.8278\n",
      "Epoch 144/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.3970 - binary_accuracy: 0.8319\n",
      "Epoch 145/225\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1755/1755 [==============================] - 1s 394us/step - loss: 0.4020 - binary_accuracy: 0.8305\n",
      "Epoch 146/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.4027 - binary_accuracy: 0.8309\n",
      "Epoch 147/225\n",
      "1755/1755 [==============================] - 1s 416us/step - loss: 0.3949 - binary_accuracy: 0.8335\n",
      "Epoch 148/225\n",
      "1755/1755 [==============================] - 1s 426us/step - loss: 0.4043 - binary_accuracy: 0.8276\n",
      "Epoch 149/225\n",
      "1755/1755 [==============================] - 1s 420us/step - loss: 0.3977 - binary_accuracy: 0.8341\n",
      "Epoch 150/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.3984 - binary_accuracy: 0.8371\n",
      "Epoch 151/225\n",
      "1755/1755 [==============================] - 1s 412us/step - loss: 0.4017 - binary_accuracy: 0.8314\n",
      "Epoch 152/225\n",
      "1755/1755 [==============================] - 1s 401us/step - loss: 0.3996 - binary_accuracy: 0.8344\n",
      "Epoch 153/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.3959 - binary_accuracy: 0.8355\n",
      "Epoch 154/225\n",
      "1755/1755 [==============================] - 1s 405us/step - loss: 0.3992 - binary_accuracy: 0.8332\n",
      "Epoch 155/225\n",
      "1755/1755 [==============================] - 1s 426us/step - loss: 0.4026 - binary_accuracy: 0.8340\n",
      "Epoch 156/225\n",
      "1755/1755 [==============================] - 1s 440us/step - loss: 0.3986 - binary_accuracy: 0.8357\n",
      "Epoch 157/225\n",
      "1755/1755 [==============================] - 1s 405us/step - loss: 0.3991 - binary_accuracy: 0.8341\n",
      "Epoch 158/225\n",
      "1755/1755 [==============================] - 1s 414us/step - loss: 0.3942 - binary_accuracy: 0.8305\n",
      "Epoch 159/225\n",
      "1755/1755 [==============================] - 1s 408us/step - loss: 0.3985 - binary_accuracy: 0.8320\n",
      "Epoch 160/225\n",
      "1755/1755 [==============================] - 1s 406us/step - loss: 0.4063 - binary_accuracy: 0.8274\n",
      "Epoch 161/225\n",
      "1755/1755 [==============================] - 1s 418us/step - loss: 0.3998 - binary_accuracy: 0.8310\n",
      "Epoch 162/225\n",
      "1755/1755 [==============================] - 1s 395us/step - loss: 0.4017 - binary_accuracy: 0.8343\n",
      "Epoch 163/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.3983 - binary_accuracy: 0.8331\n",
      "Epoch 164/225\n",
      "1755/1755 [==============================] - 1s 394us/step - loss: 0.3986 - binary_accuracy: 0.8356\n",
      "Epoch 165/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.3950 - binary_accuracy: 0.8303\n",
      "Epoch 166/225\n",
      "1755/1755 [==============================] - 1s 392us/step - loss: 0.4053 - binary_accuracy: 0.8259\n",
      "Epoch 167/225\n",
      "1755/1755 [==============================] - 1s 407us/step - loss: 0.4041 - binary_accuracy: 0.8309\n",
      "Epoch 168/225\n",
      "1755/1755 [==============================] - 1s 390us/step - loss: 0.4000 - binary_accuracy: 0.8341\n",
      "Epoch 169/225\n",
      "1755/1755 [==============================] - 1s 392us/step - loss: 0.3999 - binary_accuracy: 0.8324\n",
      "Epoch 170/225\n",
      "1755/1755 [==============================] - 1s 393us/step - loss: 0.3964 - binary_accuracy: 0.8316\n",
      "Epoch 171/225\n",
      "1755/1755 [==============================] - 1s 392us/step - loss: 0.3925 - binary_accuracy: 0.8351\n",
      "Epoch 172/225\n",
      "1755/1755 [==============================] - 1s 395us/step - loss: 0.3997 - binary_accuracy: 0.8339\n",
      "Epoch 173/225\n",
      "1755/1755 [==============================] - 1s 393us/step - loss: 0.3936 - binary_accuracy: 0.8356\n",
      "Epoch 174/225\n",
      "1755/1755 [==============================] - 1s 393us/step - loss: 0.3953 - binary_accuracy: 0.8295\n",
      "Epoch 175/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.4112 - binary_accuracy: 0.8300\n",
      "Epoch 176/225\n",
      "1755/1755 [==============================] - 1s 391us/step - loss: 0.4026 - binary_accuracy: 0.8289\n",
      "Epoch 177/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.3936 - binary_accuracy: 0.8344\n",
      "Epoch 178/225\n",
      "1755/1755 [==============================] - 1s 409us/step - loss: 0.3937 - binary_accuracy: 0.8324\n",
      "Epoch 179/225\n",
      "1755/1755 [==============================] - 1s 406us/step - loss: 0.4024 - binary_accuracy: 0.8314\n",
      "Epoch 180/225\n",
      "1755/1755 [==============================] - 1s 441us/step - loss: 0.3937 - binary_accuracy: 0.8335\n",
      "Epoch 181/225\n",
      "1755/1755 [==============================] - 1s 404us/step - loss: 0.3908 - binary_accuracy: 0.8368\n",
      "Epoch 182/225\n",
      "1755/1755 [==============================] - 1s 401us/step - loss: 0.3955 - binary_accuracy: 0.8372\n",
      "Epoch 183/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.4024 - binary_accuracy: 0.8313\n",
      "Epoch 184/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.3918 - binary_accuracy: 0.8321\n",
      "Epoch 185/225\n",
      "1755/1755 [==============================] - 1s 401us/step - loss: 0.4003 - binary_accuracy: 0.8295\n",
      "Epoch 186/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.3998 - binary_accuracy: 0.8304\n",
      "Epoch 187/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.3970 - binary_accuracy: 0.8320\n",
      "Epoch 188/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.4009 - binary_accuracy: 0.8343\n",
      "Epoch 189/225\n",
      "1755/1755 [==============================] - 1s 394us/step - loss: 0.3925 - binary_accuracy: 0.8356\n",
      "Epoch 190/225\n",
      "1755/1755 [==============================] - 1s 401us/step - loss: 0.3909 - binary_accuracy: 0.8356\n",
      "Epoch 191/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.4086 - binary_accuracy: 0.8283\n",
      "Epoch 192/225\n",
      "1755/1755 [==============================] - 1s 408us/step - loss: 0.3986 - binary_accuracy: 0.8292\n",
      "Epoch 193/225\n",
      "1755/1755 [==============================] - 1s 405us/step - loss: 0.4009 - binary_accuracy: 0.8313\n",
      "Epoch 194/225\n",
      "1755/1755 [==============================] - 1s 403us/step - loss: 0.3920 - binary_accuracy: 0.8315\n",
      "Epoch 195/225\n",
      "1755/1755 [==============================] - 1s 404us/step - loss: 0.3937 - binary_accuracy: 0.8380\n",
      "Epoch 196/225\n",
      "1755/1755 [==============================] - 1s 406us/step - loss: 0.3963 - binary_accuracy: 0.8301\n",
      "Epoch 197/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.4044 - binary_accuracy: 0.8325\n",
      "Epoch 198/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.3900 - binary_accuracy: 0.8433\n",
      "Epoch 199/225\n",
      "1755/1755 [==============================] - 1s 421us/step - loss: 0.3984 - binary_accuracy: 0.8366\n",
      "Epoch 200/225\n",
      "1755/1755 [==============================] - 1s 408us/step - loss: 0.4019 - binary_accuracy: 0.8324\n",
      "Epoch 201/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.3978 - binary_accuracy: 0.8315\n",
      "Epoch 202/225\n",
      "1755/1755 [==============================] - 1s 393us/step - loss: 0.3924 - binary_accuracy: 0.8363\n",
      "Epoch 203/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.3997 - binary_accuracy: 0.8330\n",
      "Epoch 204/225\n",
      "1755/1755 [==============================] - 1s 403us/step - loss: 0.4029 - binary_accuracy: 0.8323\n",
      "Epoch 205/225\n",
      "1755/1755 [==============================] - 1s 394us/step - loss: 0.3962 - binary_accuracy: 0.8359\n",
      "Epoch 206/225\n",
      "1755/1755 [==============================] - 1s 403us/step - loss: 0.3907 - binary_accuracy: 0.8339\n",
      "Epoch 207/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.3951 - binary_accuracy: 0.8336\n",
      "Epoch 208/225\n",
      "1755/1755 [==============================] - 1s 423us/step - loss: 0.3952 - binary_accuracy: 0.8314\n",
      "Epoch 209/225\n",
      "1755/1755 [==============================] - 1s 420us/step - loss: 0.4045 - binary_accuracy: 0.8296\n",
      "Epoch 210/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.3982 - binary_accuracy: 0.8322\n",
      "Epoch 211/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.4001 - binary_accuracy: 0.8348\n",
      "Epoch 212/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.3994 - binary_accuracy: 0.8344\n",
      "Epoch 213/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.4035 - binary_accuracy: 0.8296\n",
      "Epoch 214/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.3944 - binary_accuracy: 0.8368\n",
      "Epoch 215/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.3985 - binary_accuracy: 0.8310\n",
      "Epoch 216/225\n",
      "1755/1755 [==============================] - 1s 405us/step - loss: 0.3917 - binary_accuracy: 0.8370\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 217/225\n",
      "1755/1755 [==============================] - 1s 393us/step - loss: 0.3965 - binary_accuracy: 0.8362\n",
      "Epoch 218/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.3917 - binary_accuracy: 0.8400\n",
      "Epoch 219/225\n",
      "1755/1755 [==============================] - 1s 393us/step - loss: 0.3972 - binary_accuracy: 0.8334\n",
      "Epoch 220/225\n",
      "1755/1755 [==============================] - 1s 401us/step - loss: 0.3908 - binary_accuracy: 0.8348\n",
      "Epoch 221/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.3969 - binary_accuracy: 0.8328\n",
      "Epoch 222/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.3938 - binary_accuracy: 0.8356\n",
      "Epoch 223/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.3943 - binary_accuracy: 0.8345\n",
      "Epoch 224/225\n",
      "1755/1755 [==============================] - 1s 408us/step - loss: 0.4040 - binary_accuracy: 0.8302\n",
      "Epoch 225/225\n",
      "1755/1755 [==============================] - 1s 411us/step - loss: 0.3945 - binary_accuracy: 0.8374\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 785us/step - loss: 0.4800 - binary_accuracy: 0.8135\n",
      "Epoch 2/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.4498 - binary_accuracy: 0.8210\n",
      "Epoch 3/225\n",
      "1755/1755 [==============================] - 1s 394us/step - loss: 0.4319 - binary_accuracy: 0.8196\n",
      "Epoch 4/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.4342 - binary_accuracy: 0.8227\n",
      "Epoch 5/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.4244 - binary_accuracy: 0.8264\n",
      "Epoch 6/225\n",
      "1755/1755 [==============================] - 1s 394us/step - loss: 0.4272 - binary_accuracy: 0.8187\n",
      "Epoch 7/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.4305 - binary_accuracy: 0.8215\n",
      "Epoch 8/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.4308 - binary_accuracy: 0.8224\n",
      "Epoch 9/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.4247 - binary_accuracy: 0.8278\n",
      "Epoch 10/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.4292 - binary_accuracy: 0.8225\n",
      "Epoch 11/225\n",
      "1755/1755 [==============================] - 1s 393us/step - loss: 0.4194 - binary_accuracy: 0.8283\n",
      "Epoch 12/225\n",
      "1755/1755 [==============================] - 1s 391us/step - loss: 0.4213 - binary_accuracy: 0.8204\n",
      "Epoch 13/225\n",
      "1755/1755 [==============================] - 1s 388us/step - loss: 0.4190 - binary_accuracy: 0.8263\n",
      "Epoch 14/225\n",
      "1755/1755 [==============================] - 1s 411us/step - loss: 0.4275 - binary_accuracy: 0.8207\n",
      "Epoch 15/225\n",
      "1755/1755 [==============================] - 1s 429us/step - loss: 0.4185 - binary_accuracy: 0.8272\n",
      "Epoch 16/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.4211 - binary_accuracy: 0.8306\n",
      "Epoch 17/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.4153 - binary_accuracy: 0.8321\n",
      "Epoch 18/225\n",
      "1755/1755 [==============================] - 1s 394us/step - loss: 0.4203 - binary_accuracy: 0.8200\n",
      "Epoch 19/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.4245 - binary_accuracy: 0.8277\n",
      "Epoch 20/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.4171 - binary_accuracy: 0.8277\n",
      "Epoch 21/225\n",
      "1755/1755 [==============================] - 1s 395us/step - loss: 0.4221 - binary_accuracy: 0.8269\n",
      "Epoch 22/225\n",
      "1755/1755 [==============================] - 1s 393us/step - loss: 0.4206 - binary_accuracy: 0.8234\n",
      "Epoch 23/225\n",
      "1755/1755 [==============================] - 1s 404us/step - loss: 0.4159 - binary_accuracy: 0.8308\n",
      "Epoch 24/225\n",
      "1755/1755 [==============================] - 1s 404us/step - loss: 0.4192 - binary_accuracy: 0.8300\n",
      "Epoch 25/225\n",
      "1755/1755 [==============================] - 1s 394us/step - loss: 0.4199 - binary_accuracy: 0.8267\n",
      "Epoch 26/225\n",
      "1755/1755 [==============================] - 1s 394us/step - loss: 0.4154 - binary_accuracy: 0.8312\n",
      "Epoch 27/225\n",
      "1755/1755 [==============================] - 1s 407us/step - loss: 0.4152 - binary_accuracy: 0.8230\n",
      "Epoch 28/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.4168 - binary_accuracy: 0.8254\n",
      "Epoch 29/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.4143 - binary_accuracy: 0.8294\n",
      "Epoch 30/225\n",
      "1755/1755 [==============================] - 1s 409us/step - loss: 0.4116 - binary_accuracy: 0.8331\n",
      "Epoch 31/225\n",
      "1755/1755 [==============================] - 1s 404us/step - loss: 0.4120 - binary_accuracy: 0.8307\n",
      "Epoch 32/225\n",
      "1755/1755 [==============================] - 1s 415us/step - loss: 0.4122 - binary_accuracy: 0.8299\n",
      "Epoch 33/225\n",
      "1755/1755 [==============================] - 1s 405us/step - loss: 0.4080 - binary_accuracy: 0.8306\n",
      "Epoch 34/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.4116 - binary_accuracy: 0.8318\n",
      "Epoch 35/225\n",
      "1755/1755 [==============================] - 1s 401us/step - loss: 0.4093 - binary_accuracy: 0.8288\n",
      "Epoch 36/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.4120 - binary_accuracy: 0.8298\n",
      "Epoch 37/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.4099 - binary_accuracy: 0.8343\n",
      "Epoch 38/225\n",
      "1755/1755 [==============================] - 1s 406us/step - loss: 0.4166 - binary_accuracy: 0.8278\n",
      "Epoch 39/225\n",
      "1755/1755 [==============================] - 1s 403us/step - loss: 0.4113 - binary_accuracy: 0.8302\n",
      "Epoch 40/225\n",
      "1755/1755 [==============================] - 1s 401us/step - loss: 0.4114 - binary_accuracy: 0.8317\n",
      "Epoch 41/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.4149 - binary_accuracy: 0.8272\n",
      "Epoch 42/225\n",
      "1755/1755 [==============================] - 1s 405us/step - loss: 0.4128 - binary_accuracy: 0.8299\n",
      "Epoch 43/225\n",
      "1755/1755 [==============================] - 1s 405us/step - loss: 0.4098 - binary_accuracy: 0.8313\n",
      "Epoch 44/225\n",
      "1755/1755 [==============================] - 1s 403us/step - loss: 0.4133 - binary_accuracy: 0.8314\n",
      "Epoch 45/225\n",
      "1755/1755 [==============================] - 1s 410us/step - loss: 0.4031 - binary_accuracy: 0.8350\n",
      "Epoch 46/225\n",
      "1755/1755 [==============================] - 1s 409us/step - loss: 0.4106 - binary_accuracy: 0.8295\n",
      "Epoch 47/225\n",
      "1755/1755 [==============================] - 1s 445us/step - loss: 0.4102 - binary_accuracy: 0.8271\n",
      "Epoch 48/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.4036 - binary_accuracy: 0.8328\n",
      "Epoch 49/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.4094 - binary_accuracy: 0.8307\n",
      "Epoch 50/225\n",
      "1755/1755 [==============================] - 1s 405us/step - loss: 0.4091 - binary_accuracy: 0.8285\n",
      "Epoch 51/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.4183 - binary_accuracy: 0.8281\n",
      "Epoch 52/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.4066 - binary_accuracy: 0.8267\n",
      "Epoch 53/225\n",
      "1755/1755 [==============================] - 1s 404us/step - loss: 0.4039 - binary_accuracy: 0.8372\n",
      "Epoch 54/225\n",
      "1755/1755 [==============================] - 1s 403us/step - loss: 0.4127 - binary_accuracy: 0.8272\n",
      "Epoch 55/225\n",
      "1755/1755 [==============================] - 1s 404us/step - loss: 0.4112 - binary_accuracy: 0.8298\n",
      "Epoch 56/225\n",
      "1755/1755 [==============================] - 1s 418us/step - loss: 0.4083 - binary_accuracy: 0.8306\n",
      "Epoch 57/225\n",
      "1755/1755 [==============================] - 1s 405us/step - loss: 0.4037 - binary_accuracy: 0.8304\n",
      "Epoch 58/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.4099 - binary_accuracy: 0.8285\n",
      "Epoch 59/225\n",
      "1755/1755 [==============================] - 1s 404us/step - loss: 0.4053 - binary_accuracy: 0.8352\n",
      "Epoch 60/225\n",
      "1755/1755 [==============================] - 1s 405us/step - loss: 0.4080 - binary_accuracy: 0.8334\n",
      "Epoch 61/225\n",
      "1755/1755 [==============================] - 1s 405us/step - loss: 0.4048 - binary_accuracy: 0.8351\n",
      "Epoch 62/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.4092 - binary_accuracy: 0.8341\n",
      "Epoch 63/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.4073 - binary_accuracy: 0.8334\n",
      "Epoch 64/225\n",
      "1755/1755 [==============================] - 1s 395us/step - loss: 0.4072 - binary_accuracy: 0.8262\n",
      "Epoch 65/225\n",
      "1755/1755 [==============================] - 1s 405us/step - loss: 0.4045 - binary_accuracy: 0.8291\n",
      "Epoch 66/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.4030 - binary_accuracy: 0.8326\n",
      "Epoch 67/225\n",
      "1755/1755 [==============================] - 1s 392us/step - loss: 0.4057 - binary_accuracy: 0.8306\n",
      "Epoch 68/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.4096 - binary_accuracy: 0.8293\n",
      "Epoch 69/225\n",
      "1755/1755 [==============================] - 1s 395us/step - loss: 0.4051 - binary_accuracy: 0.8280\n",
      "Epoch 70/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.4055 - binary_accuracy: 0.8296\n",
      "Epoch 71/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.4045 - binary_accuracy: 0.8306\n",
      "Epoch 72/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.4120 - binary_accuracy: 0.8285\n",
      "Epoch 73/225\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.4085 - binary_accuracy: 0.8289\n",
      "Epoch 74/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.4032 - binary_accuracy: 0.8350\n",
      "Epoch 75/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.4063 - binary_accuracy: 0.8312\n",
      "Epoch 76/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.4094 - binary_accuracy: 0.8311\n",
      "Epoch 77/225\n",
      "1755/1755 [==============================] - 1s 394us/step - loss: 0.4003 - binary_accuracy: 0.8300\n",
      "Epoch 78/225\n",
      "1755/1755 [==============================] - 1s 395us/step - loss: 0.4110 - binary_accuracy: 0.8300\n",
      "Epoch 79/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.4034 - binary_accuracy: 0.8338\n",
      "Epoch 80/225\n",
      "1755/1755 [==============================] - 1s 440us/step - loss: 0.4033 - binary_accuracy: 0.8361\n",
      "Epoch 81/225\n",
      "1755/1755 [==============================] - 1s 408us/step - loss: 0.4046 - binary_accuracy: 0.8331\n",
      "Epoch 82/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.4084 - binary_accuracy: 0.8317\n",
      "Epoch 83/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.4029 - binary_accuracy: 0.8303\n",
      "Epoch 84/225\n",
      "1755/1755 [==============================] - 1s 394us/step - loss: 0.4059 - binary_accuracy: 0.8309\n",
      "Epoch 85/225\n",
      "1755/1755 [==============================] - 1s 390us/step - loss: 0.4081 - binary_accuracy: 0.8321\n",
      "Epoch 86/225\n",
      "1755/1755 [==============================] - 1s 391us/step - loss: 0.3996 - binary_accuracy: 0.8340\n",
      "Epoch 87/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.4004 - binary_accuracy: 0.8318\n",
      "Epoch 88/225\n",
      "1755/1755 [==============================] - 1s 403us/step - loss: 0.4021 - binary_accuracy: 0.8375\n",
      "Epoch 89/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.4074 - binary_accuracy: 0.8266\n",
      "Epoch 90/225\n",
      "1755/1755 [==============================] - 1s 395us/step - loss: 0.4081 - binary_accuracy: 0.8313\n",
      "Epoch 91/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.3986 - binary_accuracy: 0.8380\n",
      "Epoch 92/225\n",
      "1755/1755 [==============================] - 1s 401us/step - loss: 0.3962 - binary_accuracy: 0.8364\n",
      "Epoch 93/225\n",
      "1755/1755 [==============================] - 1s 393us/step - loss: 0.4091 - binary_accuracy: 0.8311\n",
      "Epoch 94/225\n",
      "1755/1755 [==============================] - 1s 393us/step - loss: 0.3996 - binary_accuracy: 0.8312\n",
      "Epoch 95/225\n",
      "1755/1755 [==============================] - 1s 393us/step - loss: 0.4020 - binary_accuracy: 0.8366\n",
      "Epoch 96/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.4110 - binary_accuracy: 0.8261\n",
      "Epoch 97/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.4053 - binary_accuracy: 0.8310\n",
      "Epoch 98/225\n",
      "1755/1755 [==============================] - 1s 394us/step - loss: 0.4007 - binary_accuracy: 0.8335\n",
      "Epoch 99/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.4119 - binary_accuracy: 0.8314\n",
      "Epoch 100/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.4015 - binary_accuracy: 0.8328\n",
      "Epoch 101/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.4046 - binary_accuracy: 0.8345\n",
      "Epoch 102/225\n",
      "1755/1755 [==============================] - 1s 403us/step - loss: 0.4061 - binary_accuracy: 0.8293\n",
      "Epoch 103/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.4004 - binary_accuracy: 0.8348\n",
      "Epoch 104/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.4016 - binary_accuracy: 0.8339\n",
      "Epoch 105/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.4050 - binary_accuracy: 0.8277\n",
      "Epoch 106/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.4025 - binary_accuracy: 0.8296\n",
      "Epoch 107/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.4003 - binary_accuracy: 0.8368\n",
      "Epoch 108/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.4059 - binary_accuracy: 0.8316\n",
      "Epoch 109/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.3971 - binary_accuracy: 0.8359\n",
      "Epoch 110/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.4000 - binary_accuracy: 0.8367\n",
      "Epoch 111/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.4067 - binary_accuracy: 0.8312\n",
      "Epoch 112/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.4036 - binary_accuracy: 0.8341\n",
      "Epoch 113/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.3953 - binary_accuracy: 0.8372\n",
      "Epoch 114/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.4078 - binary_accuracy: 0.8301\n",
      "Epoch 115/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.3992 - binary_accuracy: 0.8376\n",
      "Epoch 116/225\n",
      "1755/1755 [==============================] - 1s 394us/step - loss: 0.3909 - binary_accuracy: 0.8403\n",
      "Epoch 117/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.3928 - binary_accuracy: 0.8346\n",
      "Epoch 118/225\n",
      "1755/1755 [==============================] - 1s 446us/step - loss: 0.4049 - binary_accuracy: 0.8326\n",
      "Epoch 119/225\n",
      "1755/1755 [==============================] - 1s 410us/step - loss: 0.4004 - binary_accuracy: 0.8274\n",
      "Epoch 120/225\n",
      "1755/1755 [==============================] - 1s 395us/step - loss: 0.4012 - binary_accuracy: 0.8347\n",
      "Epoch 121/225\n",
      "1755/1755 [==============================] - 1s 394us/step - loss: 0.4070 - binary_accuracy: 0.8301\n",
      "Epoch 122/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.3919 - binary_accuracy: 0.8394\n",
      "Epoch 123/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.4018 - binary_accuracy: 0.8347\n",
      "Epoch 124/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.3966 - binary_accuracy: 0.8376\n",
      "Epoch 125/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.4019 - binary_accuracy: 0.8302\n",
      "Epoch 126/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.4018 - binary_accuracy: 0.8339\n",
      "Epoch 127/225\n",
      "1755/1755 [==============================] - 1s 403us/step - loss: 0.3950 - binary_accuracy: 0.8341\n",
      "Epoch 128/225\n",
      "1755/1755 [==============================] - 1s 409us/step - loss: 0.4003 - binary_accuracy: 0.8345\n",
      "Epoch 129/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.3980 - binary_accuracy: 0.8353\n",
      "Epoch 130/225\n",
      "1755/1755 [==============================] - 1s 401us/step - loss: 0.4136 - binary_accuracy: 0.8297\n",
      "Epoch 131/225\n",
      "1755/1755 [==============================] - 1s 394us/step - loss: 0.4000 - binary_accuracy: 0.8334\n",
      "Epoch 132/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.4018 - binary_accuracy: 0.8336\n",
      "Epoch 133/225\n",
      "1755/1755 [==============================] - 1s 391us/step - loss: 0.3964 - binary_accuracy: 0.8363\n",
      "Epoch 134/225\n",
      "1755/1755 [==============================] - 1s 392us/step - loss: 0.4025 - binary_accuracy: 0.8291\n",
      "Epoch 135/225\n",
      "1755/1755 [==============================] - 1s 409us/step - loss: 0.4055 - binary_accuracy: 0.8322\n",
      "Epoch 136/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.3953 - binary_accuracy: 0.8330\n",
      "Epoch 137/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.3931 - binary_accuracy: 0.8362\n",
      "Epoch 138/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.4001 - binary_accuracy: 0.8302\n",
      "Epoch 139/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.3935 - binary_accuracy: 0.8362\n",
      "Epoch 140/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.3968 - binary_accuracy: 0.8353\n",
      "Epoch 141/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.4014 - binary_accuracy: 0.8373\n",
      "Epoch 142/225\n",
      "1755/1755 [==============================] - 1s 412us/step - loss: 0.4008 - binary_accuracy: 0.8351\n",
      "Epoch 143/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.4022 - binary_accuracy: 0.8277\n",
      "Epoch 144/225\n",
      "1755/1755 [==============================] - 1s 408us/step - loss: 0.4040 - binary_accuracy: 0.8285\n",
      "Epoch 145/225\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1755/1755 [==============================] - 1s 401us/step - loss: 0.3981 - binary_accuracy: 0.8331\n",
      "Epoch 146/225\n",
      "1755/1755 [==============================] - 1s 406us/step - loss: 0.3941 - binary_accuracy: 0.8395\n",
      "Epoch 147/225\n",
      "1755/1755 [==============================] - 1s 403us/step - loss: 0.4013 - binary_accuracy: 0.8311\n",
      "Epoch 148/225\n",
      "1755/1755 [==============================] - 1s 405us/step - loss: 0.3984 - binary_accuracy: 0.8401\n",
      "Epoch 149/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.4004 - binary_accuracy: 0.8365\n",
      "Epoch 150/225\n",
      "1755/1755 [==============================] - 1s 401us/step - loss: 0.3947 - binary_accuracy: 0.8360\n",
      "Epoch 151/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.3951 - binary_accuracy: 0.8331\n",
      "Epoch 152/225\n",
      "1755/1755 [==============================] - 1s 395us/step - loss: 0.3994 - binary_accuracy: 0.8342\n",
      "Epoch 153/225\n",
      "1755/1755 [==============================] - 1s 405us/step - loss: 0.4000 - binary_accuracy: 0.8330\n",
      "Epoch 154/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.3960 - binary_accuracy: 0.8407\n",
      "Epoch 155/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.4036 - binary_accuracy: 0.8302\n",
      "Epoch 156/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.4034 - binary_accuracy: 0.8352\n",
      "Epoch 157/225\n",
      "1755/1755 [==============================] - 1s 393us/step - loss: 0.4007 - binary_accuracy: 0.8307\n",
      "Epoch 158/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.3939 - binary_accuracy: 0.8347\n",
      "Epoch 159/225\n",
      "1755/1755 [==============================] - 1s 403us/step - loss: 0.4006 - binary_accuracy: 0.8343\n",
      "Epoch 160/225\n",
      "1755/1755 [==============================] - 1s 403us/step - loss: 0.4006 - binary_accuracy: 0.8267\n",
      "Epoch 161/225\n",
      "1755/1755 [==============================] - 1s 401us/step - loss: 0.3901 - binary_accuracy: 0.8346\n",
      "Epoch 162/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.3960 - binary_accuracy: 0.8398\n",
      "Epoch 163/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.3896 - binary_accuracy: 0.8400\n",
      "Epoch 164/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.3965 - binary_accuracy: 0.8341\n",
      "Epoch 165/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.3985 - binary_accuracy: 0.8341\n",
      "Epoch 166/225\n",
      "1755/1755 [==============================] - 1s 412us/step - loss: 0.3927 - binary_accuracy: 0.8387\n",
      "Epoch 167/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.3945 - binary_accuracy: 0.8361\n",
      "Epoch 168/225\n",
      "1755/1755 [==============================] - 1s 426us/step - loss: 0.3993 - binary_accuracy: 0.8327\n",
      "Epoch 169/225\n",
      "1755/1755 [==============================] - 1s 421us/step - loss: 0.3932 - binary_accuracy: 0.8366\n",
      "Epoch 170/225\n",
      "1755/1755 [==============================] - 1s 403us/step - loss: 0.3929 - binary_accuracy: 0.8385\n",
      "Epoch 171/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.3910 - binary_accuracy: 0.8344\n",
      "Epoch 172/225\n",
      "1755/1755 [==============================] - 1s 395us/step - loss: 0.3959 - binary_accuracy: 0.8376\n",
      "Epoch 173/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.3919 - binary_accuracy: 0.8377\n",
      "Epoch 174/225\n",
      "1755/1755 [==============================] - 1s 392us/step - loss: 0.3933 - binary_accuracy: 0.8394\n",
      "Epoch 175/225\n",
      "1755/1755 [==============================] - 1s 392us/step - loss: 0.4035 - binary_accuracy: 0.8303\n",
      "Epoch 176/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.3968 - binary_accuracy: 0.8353\n",
      "Epoch 177/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.3974 - binary_accuracy: 0.8390\n",
      "Epoch 178/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.3919 - binary_accuracy: 0.8397\n",
      "Epoch 179/225\n",
      "1755/1755 [==============================] - 1s 391us/step - loss: 0.3910 - binary_accuracy: 0.8377\n",
      "Epoch 180/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.3895 - binary_accuracy: 0.8356\n",
      "Epoch 181/225\n",
      "1755/1755 [==============================] - 1s 395us/step - loss: 0.3916 - binary_accuracy: 0.8353\n",
      "Epoch 182/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.4063 - binary_accuracy: 0.8293\n",
      "Epoch 183/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.3953 - binary_accuracy: 0.8354\n",
      "Epoch 184/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.3933 - binary_accuracy: 0.8360\n",
      "Epoch 185/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.3983 - binary_accuracy: 0.8353\n",
      "Epoch 186/225\n",
      "1755/1755 [==============================] - 1s 407us/step - loss: 0.3990 - binary_accuracy: 0.8342\n",
      "Epoch 187/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.3963 - binary_accuracy: 0.8348\n",
      "Epoch 188/225\n",
      "1755/1755 [==============================] - 1s 393us/step - loss: 0.3973 - binary_accuracy: 0.8396\n",
      "Epoch 189/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.3927 - binary_accuracy: 0.8338\n",
      "Epoch 190/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.3987 - binary_accuracy: 0.8305\n",
      "Epoch 191/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.3938 - binary_accuracy: 0.8372\n",
      "Epoch 192/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.3894 - binary_accuracy: 0.8348\n",
      "Epoch 193/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.4024 - binary_accuracy: 0.8304\n",
      "Epoch 194/225\n",
      "1755/1755 [==============================] - 1s 404us/step - loss: 0.4018 - binary_accuracy: 0.8336\n",
      "Epoch 195/225\n",
      "1755/1755 [==============================] - 1s 394us/step - loss: 0.3957 - binary_accuracy: 0.8365\n",
      "Epoch 196/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.3982 - binary_accuracy: 0.8371\n",
      "Epoch 197/225\n",
      "1755/1755 [==============================] - 1s 394us/step - loss: 0.3947 - binary_accuracy: 0.8360\n",
      "Epoch 198/225\n",
      "1755/1755 [==============================] - 1s 410us/step - loss: 0.3932 - binary_accuracy: 0.8367\n",
      "Epoch 199/225\n",
      "1755/1755 [==============================] - 1s 406us/step - loss: 0.3991 - binary_accuracy: 0.8341\n",
      "Epoch 200/225\n",
      "1755/1755 [==============================] - 1s 398us/step - loss: 0.3841 - binary_accuracy: 0.8433\n",
      "Epoch 201/225\n",
      "1755/1755 [==============================] - 1s 395us/step - loss: 0.3866 - binary_accuracy: 0.8402\n",
      "Epoch 202/225\n",
      "1755/1755 [==============================] - 1s 406us/step - loss: 0.3992 - binary_accuracy: 0.8341\n",
      "Epoch 203/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.4034 - binary_accuracy: 0.8309\n",
      "Epoch 204/225\n",
      "1755/1755 [==============================] - 1s 422us/step - loss: 0.3963 - binary_accuracy: 0.8340\n",
      "Epoch 205/225\n",
      "1755/1755 [==============================] - 1s 429us/step - loss: 0.3922 - binary_accuracy: 0.8390\n",
      "Epoch 206/225\n",
      "1755/1755 [==============================] - 1s 395us/step - loss: 0.3918 - binary_accuracy: 0.8374\n",
      "Epoch 207/225\n",
      "1755/1755 [==============================] - 1s 402us/step - loss: 0.3971 - binary_accuracy: 0.8373\n",
      "Epoch 208/225\n",
      "1755/1755 [==============================] - 1s 405us/step - loss: 0.3910 - binary_accuracy: 0.8375\n",
      "Epoch 209/225\n",
      "1755/1755 [==============================] - 1s 410us/step - loss: 0.3932 - binary_accuracy: 0.8325\n",
      "Epoch 210/225\n",
      "1755/1755 [==============================] - 1s 407us/step - loss: 0.3997 - binary_accuracy: 0.8267\n",
      "Epoch 211/225\n",
      "1755/1755 [==============================] - 1s 405us/step - loss: 0.3911 - binary_accuracy: 0.8402\n",
      "Epoch 212/225\n",
      "1755/1755 [==============================] - 1s 405us/step - loss: 0.3892 - binary_accuracy: 0.8380\n",
      "Epoch 213/225\n",
      "1755/1755 [==============================] - 1s 405us/step - loss: 0.3893 - binary_accuracy: 0.8398\n",
      "Epoch 214/225\n",
      "1755/1755 [==============================] - 1s 408us/step - loss: 0.3882 - binary_accuracy: 0.8388\n",
      "Epoch 215/225\n",
      "1755/1755 [==============================] - 1s 416us/step - loss: 0.3918 - binary_accuracy: 0.8366\n",
      "Epoch 216/225\n",
      "1755/1755 [==============================] - 1s 405us/step - loss: 0.3908 - binary_accuracy: 0.8377\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 217/225\n",
      "1755/1755 [==============================] - 1s 407us/step - loss: 0.3940 - binary_accuracy: 0.8337\n",
      "Epoch 218/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.3969 - binary_accuracy: 0.8373\n",
      "Epoch 219/225\n",
      "1755/1755 [==============================] - 1s 395us/step - loss: 0.3966 - binary_accuracy: 0.8373\n",
      "Epoch 220/225\n",
      "1755/1755 [==============================] - 1s 409us/step - loss: 0.3935 - binary_accuracy: 0.8348\n",
      "Epoch 221/225\n",
      "1755/1755 [==============================] - 1s 399us/step - loss: 0.3911 - binary_accuracy: 0.8377\n",
      "Epoch 222/225\n",
      "1755/1755 [==============================] - 1s 401us/step - loss: 0.3931 - binary_accuracy: 0.8372\n",
      "Epoch 223/225\n",
      "1755/1755 [==============================] - 1s 397us/step - loss: 0.3951 - binary_accuracy: 0.8363\n",
      "Epoch 224/225\n",
      "1755/1755 [==============================] - 1s 396us/step - loss: 0.3914 - binary_accuracy: 0.8372\n",
      "Epoch 225/225\n",
      "1755/1755 [==============================] - 1s 400us/step - loss: 0.3874 - binary_accuracy: 0.8407\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 842us/step - loss: 0.4740 - binary_accuracy: 0.8179\n",
      "Epoch 2/225\n",
      "1754/1754 [==============================] - 1s 417us/step - loss: 0.4382 - binary_accuracy: 0.8274\n",
      "Epoch 3/225\n",
      "1754/1754 [==============================] - 1s 437us/step - loss: 0.4355 - binary_accuracy: 0.8261\n",
      "Epoch 4/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.4349 - binary_accuracy: 0.8261\n",
      "Epoch 5/225\n",
      "1754/1754 [==============================] - 1s 403us/step - loss: 0.4297 - binary_accuracy: 0.8206\n",
      "Epoch 6/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.4258 - binary_accuracy: 0.8258\n",
      "Epoch 7/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.4193 - binary_accuracy: 0.8256\n",
      "Epoch 8/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4208 - binary_accuracy: 0.8264\n",
      "Epoch 9/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.4211 - binary_accuracy: 0.8229\n",
      "Epoch 10/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.4204 - binary_accuracy: 0.8250\n",
      "Epoch 11/225\n",
      "1754/1754 [==============================] - 1s 405us/step - loss: 0.4109 - binary_accuracy: 0.8227\n",
      "Epoch 12/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4171 - binary_accuracy: 0.8297\n",
      "Epoch 13/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4063 - binary_accuracy: 0.8305\n",
      "Epoch 14/225\n",
      "1754/1754 [==============================] - 1s 406us/step - loss: 0.4173 - binary_accuracy: 0.8244\n",
      "Epoch 15/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.4098 - binary_accuracy: 0.8333\n",
      "Epoch 16/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4152 - binary_accuracy: 0.8297\n",
      "Epoch 17/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4073 - binary_accuracy: 0.8268\n",
      "Epoch 18/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4105 - binary_accuracy: 0.8341\n",
      "Epoch 19/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4123 - binary_accuracy: 0.8285\n",
      "Epoch 20/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4099 - binary_accuracy: 0.8283\n",
      "Epoch 21/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4120 - binary_accuracy: 0.8321\n",
      "Epoch 22/225\n",
      "1754/1754 [==============================] - 1s 409us/step - loss: 0.4045 - binary_accuracy: 0.8253\n",
      "Epoch 23/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.4165 - binary_accuracy: 0.8267\n",
      "Epoch 24/225\n",
      "1754/1754 [==============================] - 1s 403us/step - loss: 0.4158 - binary_accuracy: 0.8227\n",
      "Epoch 25/225\n",
      "1754/1754 [==============================] - 1s 405us/step - loss: 0.3986 - binary_accuracy: 0.8342\n",
      "Epoch 26/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4036 - binary_accuracy: 0.8337\n",
      "Epoch 27/225\n",
      "1754/1754 [==============================] - 1s 403us/step - loss: 0.4088 - binary_accuracy: 0.8316\n",
      "Epoch 28/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4071 - binary_accuracy: 0.8285\n",
      "Epoch 29/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4073 - binary_accuracy: 0.8253\n",
      "Epoch 30/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4067 - binary_accuracy: 0.8347\n",
      "Epoch 31/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.4092 - binary_accuracy: 0.8318\n",
      "Epoch 32/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4085 - binary_accuracy: 0.8294\n",
      "Epoch 33/225\n",
      "1754/1754 [==============================] - 1s 406us/step - loss: 0.4138 - binary_accuracy: 0.8240\n",
      "Epoch 34/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4094 - binary_accuracy: 0.8266\n",
      "Epoch 35/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4052 - binary_accuracy: 0.8323\n",
      "Epoch 36/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4066 - binary_accuracy: 0.8321\n",
      "Epoch 37/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4036 - binary_accuracy: 0.8294\n",
      "Epoch 38/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4053 - binary_accuracy: 0.8332\n",
      "Epoch 39/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4061 - binary_accuracy: 0.8323\n",
      "Epoch 40/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4087 - binary_accuracy: 0.8295\n",
      "Epoch 41/225\n",
      "1754/1754 [==============================] - 1s 438us/step - loss: 0.4070 - binary_accuracy: 0.8321\n",
      "Epoch 42/225\n",
      "1754/1754 [==============================] - 1s 415us/step - loss: 0.4028 - binary_accuracy: 0.8394\n",
      "Epoch 43/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4077 - binary_accuracy: 0.8329\n",
      "Epoch 44/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4030 - binary_accuracy: 0.8295\n",
      "Epoch 45/225\n",
      "1754/1754 [==============================] - 1s 404us/step - loss: 0.4041 - binary_accuracy: 0.8311\n",
      "Epoch 46/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.4103 - binary_accuracy: 0.8289\n",
      "Epoch 47/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.3959 - binary_accuracy: 0.8350\n",
      "Epoch 48/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4074 - binary_accuracy: 0.8299\n",
      "Epoch 49/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4072 - binary_accuracy: 0.8282\n",
      "Epoch 50/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4001 - binary_accuracy: 0.8293\n",
      "Epoch 51/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4033 - binary_accuracy: 0.8315\n",
      "Epoch 52/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.3980 - binary_accuracy: 0.8342\n",
      "Epoch 53/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.3961 - binary_accuracy: 0.8331\n",
      "Epoch 54/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4011 - binary_accuracy: 0.8303\n",
      "Epoch 55/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.3981 - binary_accuracy: 0.8337\n",
      "Epoch 56/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4007 - binary_accuracy: 0.8344\n",
      "Epoch 57/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4104 - binary_accuracy: 0.8278\n",
      "Epoch 58/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4009 - binary_accuracy: 0.8314\n",
      "Epoch 59/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4077 - binary_accuracy: 0.8280\n",
      "Epoch 60/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.4005 - binary_accuracy: 0.8353\n",
      "Epoch 61/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.3980 - binary_accuracy: 0.8344\n",
      "Epoch 62/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4002 - binary_accuracy: 0.8288\n",
      "Epoch 63/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.4030 - binary_accuracy: 0.8276\n",
      "Epoch 64/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.3989 - binary_accuracy: 0.8353\n",
      "Epoch 65/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4067 - binary_accuracy: 0.8335\n",
      "Epoch 66/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4030 - binary_accuracy: 0.8346\n",
      "Epoch 67/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.4000 - binary_accuracy: 0.8342\n",
      "Epoch 68/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.3978 - binary_accuracy: 0.8283\n",
      "Epoch 69/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4069 - binary_accuracy: 0.8326\n",
      "Epoch 70/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4074 - binary_accuracy: 0.8282\n",
      "Epoch 71/225\n",
      "1754/1754 [==============================] - 1s 405us/step - loss: 0.4054 - binary_accuracy: 0.8336\n",
      "Epoch 72/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.3972 - binary_accuracy: 0.8339\n",
      "Epoch 73/225\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4014 - binary_accuracy: 0.8330\n",
      "Epoch 74/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4008 - binary_accuracy: 0.8339\n",
      "Epoch 75/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4027 - binary_accuracy: 0.8327\n",
      "Epoch 76/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.3960 - binary_accuracy: 0.8414\n",
      "Epoch 77/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.3980 - binary_accuracy: 0.8380\n",
      "Epoch 78/225\n",
      "1754/1754 [==============================] - 1s 405us/step - loss: 0.4072 - binary_accuracy: 0.8257\n",
      "Epoch 79/225\n",
      "1754/1754 [==============================] - 1s 438us/step - loss: 0.3957 - binary_accuracy: 0.8337\n",
      "Epoch 80/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4008 - binary_accuracy: 0.8297\n",
      "Epoch 81/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.3961 - binary_accuracy: 0.8316\n",
      "Epoch 82/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4099 - binary_accuracy: 0.8268\n",
      "Epoch 83/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.4020 - binary_accuracy: 0.8334\n",
      "Epoch 84/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4002 - binary_accuracy: 0.8313\n",
      "Epoch 85/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.3965 - binary_accuracy: 0.8297\n",
      "Epoch 86/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4118 - binary_accuracy: 0.8285\n",
      "Epoch 87/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.3938 - binary_accuracy: 0.8396\n",
      "Epoch 88/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4003 - binary_accuracy: 0.8320\n",
      "Epoch 89/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.4009 - binary_accuracy: 0.8319\n",
      "Epoch 90/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.3984 - binary_accuracy: 0.8325\n",
      "Epoch 91/225\n",
      "1754/1754 [==============================] - 1s 405us/step - loss: 0.3953 - binary_accuracy: 0.8312\n",
      "Epoch 92/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.3939 - binary_accuracy: 0.8378\n",
      "Epoch 93/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.4000 - binary_accuracy: 0.8353\n",
      "Epoch 94/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4010 - binary_accuracy: 0.8303\n",
      "Epoch 95/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.3997 - binary_accuracy: 0.8341\n",
      "Epoch 96/225\n",
      "1754/1754 [==============================] - 1s 405us/step - loss: 0.4029 - binary_accuracy: 0.8357\n",
      "Epoch 97/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.3985 - binary_accuracy: 0.8327\n",
      "Epoch 98/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.3936 - binary_accuracy: 0.8354\n",
      "Epoch 99/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.3941 - binary_accuracy: 0.8320\n",
      "Epoch 100/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4019 - binary_accuracy: 0.8323\n",
      "Epoch 101/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4026 - binary_accuracy: 0.8343\n",
      "Epoch 102/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.3956 - binary_accuracy: 0.8362\n",
      "Epoch 103/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.3948 - binary_accuracy: 0.8312\n",
      "Epoch 104/225\n",
      "1754/1754 [==============================] - 1s 410us/step - loss: 0.3973 - binary_accuracy: 0.8383\n",
      "Epoch 105/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.3978 - binary_accuracy: 0.8377\n",
      "Epoch 106/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4022 - binary_accuracy: 0.8314\n",
      "Epoch 107/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.3966 - binary_accuracy: 0.8306\n",
      "Epoch 108/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.3955 - binary_accuracy: 0.8371\n",
      "Epoch 109/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4013 - binary_accuracy: 0.8347\n",
      "Epoch 110/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.3923 - binary_accuracy: 0.8372\n",
      "Epoch 111/225\n",
      "1754/1754 [==============================] - 1s 403us/step - loss: 0.3976 - binary_accuracy: 0.8364\n",
      "Epoch 112/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.3954 - binary_accuracy: 0.8330\n",
      "Epoch 113/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.3990 - binary_accuracy: 0.8312\n",
      "Epoch 114/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.3994 - binary_accuracy: 0.8364\n",
      "Epoch 115/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.4027 - binary_accuracy: 0.8336\n",
      "Epoch 116/225\n",
      "1754/1754 [==============================] - 1s 406us/step - loss: 0.3986 - binary_accuracy: 0.8334\n",
      "Epoch 117/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.3974 - binary_accuracy: 0.8351\n",
      "Epoch 118/225\n",
      "1754/1754 [==============================] - 1s 405us/step - loss: 0.3919 - binary_accuracy: 0.8390\n",
      "Epoch 119/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4014 - binary_accuracy: 0.8306\n",
      "Epoch 120/225\n",
      "1754/1754 [==============================] - 1s 443us/step - loss: 0.4043 - binary_accuracy: 0.8280\n",
      "Epoch 121/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.3961 - binary_accuracy: 0.8365\n",
      "Epoch 122/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.3974 - binary_accuracy: 0.8357\n",
      "Epoch 123/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.3889 - binary_accuracy: 0.8367\n",
      "Epoch 124/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.3950 - binary_accuracy: 0.8346\n",
      "Epoch 125/225\n",
      "1754/1754 [==============================] - 1s 404us/step - loss: 0.4036 - binary_accuracy: 0.8336\n",
      "Epoch 126/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.3893 - binary_accuracy: 0.8383\n",
      "Epoch 127/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.3902 - binary_accuracy: 0.8332\n",
      "Epoch 128/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4007 - binary_accuracy: 0.8322\n",
      "Epoch 129/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.4011 - binary_accuracy: 0.8251\n",
      "Epoch 130/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.3960 - binary_accuracy: 0.8341\n",
      "Epoch 131/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.3993 - binary_accuracy: 0.8320\n",
      "Epoch 132/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.3865 - binary_accuracy: 0.8327\n",
      "Epoch 133/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.3986 - binary_accuracy: 0.8321\n",
      "Epoch 134/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.3940 - binary_accuracy: 0.8375\n",
      "Epoch 135/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.3924 - binary_accuracy: 0.8327\n",
      "Epoch 136/225\n",
      "1754/1754 [==============================] - 1s 406us/step - loss: 0.3995 - binary_accuracy: 0.8317\n",
      "Epoch 137/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.3950 - binary_accuracy: 0.8335\n",
      "Epoch 138/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.3915 - binary_accuracy: 0.8314\n",
      "Epoch 139/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.3972 - binary_accuracy: 0.8307\n",
      "Epoch 140/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.3962 - binary_accuracy: 0.8337\n",
      "Epoch 141/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.3935 - binary_accuracy: 0.8349\n",
      "Epoch 142/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.3932 - binary_accuracy: 0.8357\n",
      "Epoch 143/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.3927 - binary_accuracy: 0.8346\n",
      "Epoch 144/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.3895 - binary_accuracy: 0.8333\n",
      "Epoch 145/225\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.3884 - binary_accuracy: 0.8344\n",
      "Epoch 146/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.3976 - binary_accuracy: 0.8346\n",
      "Epoch 147/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.4022 - binary_accuracy: 0.8378\n",
      "Epoch 148/225\n",
      "1754/1754 [==============================] - 1s 403us/step - loss: 0.3975 - binary_accuracy: 0.8341\n",
      "Epoch 149/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.3956 - binary_accuracy: 0.8318\n",
      "Epoch 150/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.3990 - binary_accuracy: 0.8327\n",
      "Epoch 151/225\n",
      "1754/1754 [==============================] - 1s 406us/step - loss: 0.3954 - binary_accuracy: 0.8313\n",
      "Epoch 152/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.3927 - binary_accuracy: 0.8310\n",
      "Epoch 153/225\n",
      "1754/1754 [==============================] - 1s 404us/step - loss: 0.3909 - binary_accuracy: 0.8362\n",
      "Epoch 154/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.3898 - binary_accuracy: 0.8370\n",
      "Epoch 155/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.3957 - binary_accuracy: 0.8383\n",
      "Epoch 156/225\n",
      "1754/1754 [==============================] - 1s 403us/step - loss: 0.3922 - binary_accuracy: 0.8331\n",
      "Epoch 157/225\n",
      "1754/1754 [==============================] - 1s 406us/step - loss: 0.3982 - binary_accuracy: 0.8332\n",
      "Epoch 158/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.3980 - binary_accuracy: 0.8358\n",
      "Epoch 159/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4002 - binary_accuracy: 0.8345\n",
      "Epoch 160/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4004 - binary_accuracy: 0.8300\n",
      "Epoch 161/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.3963 - binary_accuracy: 0.8365\n",
      "Epoch 162/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.3994 - binary_accuracy: 0.8301\n",
      "Epoch 163/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.3918 - binary_accuracy: 0.8356\n",
      "Epoch 164/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.3952 - binary_accuracy: 0.8340\n",
      "Epoch 165/225\n",
      "1754/1754 [==============================] - 1s 407us/step - loss: 0.3916 - binary_accuracy: 0.8342\n",
      "Epoch 166/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.3966 - binary_accuracy: 0.8371\n",
      "Epoch 167/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.3953 - binary_accuracy: 0.8400\n",
      "Epoch 168/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.3914 - binary_accuracy: 0.8376\n",
      "Epoch 169/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.3875 - binary_accuracy: 0.8334\n",
      "Epoch 170/225\n",
      "1754/1754 [==============================] - 1s 408us/step - loss: 0.3907 - binary_accuracy: 0.8337\n",
      "Epoch 171/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.3919 - binary_accuracy: 0.8375\n",
      "Epoch 172/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.3880 - binary_accuracy: 0.8424\n",
      "Epoch 173/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.3967 - binary_accuracy: 0.8320\n",
      "Epoch 174/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.3854 - binary_accuracy: 0.8385\n",
      "Epoch 175/225\n",
      "1754/1754 [==============================] - 1s 405us/step - loss: 0.3911 - binary_accuracy: 0.8361\n",
      "Epoch 176/225\n",
      "1754/1754 [==============================] - 1s 407us/step - loss: 0.3924 - binary_accuracy: 0.8365\n",
      "Epoch 177/225\n",
      "1754/1754 [==============================] - 1s 445us/step - loss: 0.3914 - binary_accuracy: 0.8349\n",
      "Epoch 178/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.3888 - binary_accuracy: 0.8309\n",
      "Epoch 179/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4011 - binary_accuracy: 0.8299\n",
      "Epoch 180/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.3950 - binary_accuracy: 0.8360\n",
      "Epoch 181/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.3921 - binary_accuracy: 0.8368\n",
      "Epoch 182/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.3945 - binary_accuracy: 0.8282\n",
      "Epoch 183/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.3875 - binary_accuracy: 0.8382\n",
      "Epoch 184/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.3897 - binary_accuracy: 0.8369\n",
      "Epoch 185/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.3821 - binary_accuracy: 0.8411\n",
      "Epoch 186/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.3874 - binary_accuracy: 0.8349\n",
      "Epoch 187/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.3955 - binary_accuracy: 0.8377\n",
      "Epoch 188/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.3821 - binary_accuracy: 0.8408\n",
      "Epoch 189/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.3969 - binary_accuracy: 0.8349\n",
      "Epoch 190/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.3869 - binary_accuracy: 0.8414\n",
      "Epoch 191/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.3983 - binary_accuracy: 0.8298\n",
      "Epoch 192/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.3947 - binary_accuracy: 0.8351\n",
      "Epoch 193/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.4008 - binary_accuracy: 0.8312\n",
      "Epoch 194/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.3929 - binary_accuracy: 0.8359\n",
      "Epoch 195/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.3913 - binary_accuracy: 0.8362\n",
      "Epoch 196/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.3928 - binary_accuracy: 0.8347\n",
      "Epoch 197/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4010 - binary_accuracy: 0.8325\n",
      "Epoch 198/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.3908 - binary_accuracy: 0.8364\n",
      "Epoch 199/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.3903 - binary_accuracy: 0.8303\n",
      "Epoch 200/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.3882 - binary_accuracy: 0.8396\n",
      "Epoch 201/225\n",
      "1754/1754 [==============================] - 1s 430us/step - loss: 0.3873 - binary_accuracy: 0.8386\n",
      "Epoch 202/225\n",
      "1754/1754 [==============================] - 1s 406us/step - loss: 0.3908 - binary_accuracy: 0.8371\n",
      "Epoch 203/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.3932 - binary_accuracy: 0.8291\n",
      "Epoch 204/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.3926 - binary_accuracy: 0.8363\n",
      "Epoch 205/225\n",
      "1754/1754 [==============================] - 1s 404us/step - loss: 0.3869 - binary_accuracy: 0.8395\n",
      "Epoch 206/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.3916 - binary_accuracy: 0.8345\n",
      "Epoch 207/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.3863 - binary_accuracy: 0.8323\n",
      "Epoch 208/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.3835 - binary_accuracy: 0.8372\n",
      "Epoch 209/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.3855 - binary_accuracy: 0.8434\n",
      "Epoch 210/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.3859 - binary_accuracy: 0.8405\n",
      "Epoch 211/225\n",
      "1754/1754 [==============================] - 1s 405us/step - loss: 0.3918 - binary_accuracy: 0.8371\n",
      "Epoch 212/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.3915 - binary_accuracy: 0.8384\n",
      "Epoch 213/225\n",
      "1754/1754 [==============================] - 1s 444us/step - loss: 0.3885 - binary_accuracy: 0.8338\n",
      "Epoch 214/225\n",
      "1754/1754 [==============================] - 1s 404us/step - loss: 0.3911 - binary_accuracy: 0.8378\n",
      "Epoch 215/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.3875 - binary_accuracy: 0.8341\n",
      "Epoch 216/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.3907 - binary_accuracy: 0.8378\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 217/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.3835 - binary_accuracy: 0.8394\n",
      "Epoch 218/225\n",
      "1754/1754 [==============================] - 1s 403us/step - loss: 0.3894 - binary_accuracy: 0.8336\n",
      "Epoch 219/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.3858 - binary_accuracy: 0.8408\n",
      "Epoch 220/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.3882 - binary_accuracy: 0.8393\n",
      "Epoch 221/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.3808 - binary_accuracy: 0.8423\n",
      "Epoch 222/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.3870 - binary_accuracy: 0.8390\n",
      "Epoch 223/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.3932 - binary_accuracy: 0.8345\n",
      "Epoch 224/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.3938 - binary_accuracy: 0.8316\n",
      "Epoch 225/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.3934 - binary_accuracy: 0.8326\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 893us/step - loss: 0.4855 - binary_accuracy: 0.8146\n",
      "Epoch 2/225\n",
      "1754/1754 [==============================] - 1s 405us/step - loss: 0.4503 - binary_accuracy: 0.8192\n",
      "Epoch 3/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.4438 - binary_accuracy: 0.8159\n",
      "Epoch 4/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4364 - binary_accuracy: 0.8231\n",
      "Epoch 5/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.4419 - binary_accuracy: 0.8207\n",
      "Epoch 6/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4325 - binary_accuracy: 0.8229\n",
      "Epoch 7/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4376 - binary_accuracy: 0.8212\n",
      "Epoch 8/225\n",
      "1754/1754 [==============================] - 1s 404us/step - loss: 0.4217 - binary_accuracy: 0.8260\n",
      "Epoch 9/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4283 - binary_accuracy: 0.8222\n",
      "Epoch 10/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4254 - binary_accuracy: 0.8241\n",
      "Epoch 11/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4296 - binary_accuracy: 0.8216\n",
      "Epoch 12/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4198 - binary_accuracy: 0.8205\n",
      "Epoch 13/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4311 - binary_accuracy: 0.8183\n",
      "Epoch 14/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4193 - binary_accuracy: 0.8225\n",
      "Epoch 15/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4247 - binary_accuracy: 0.8238\n",
      "Epoch 16/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4304 - binary_accuracy: 0.8194\n",
      "Epoch 17/225\n",
      "1754/1754 [==============================] - 1s 404us/step - loss: 0.4083 - binary_accuracy: 0.8270\n",
      "Epoch 18/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4277 - binary_accuracy: 0.8211\n",
      "Epoch 19/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4278 - binary_accuracy: 0.8225\n",
      "Epoch 20/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4233 - binary_accuracy: 0.8251\n",
      "Epoch 21/225\n",
      "1754/1754 [==============================] - 1s 407us/step - loss: 0.4148 - binary_accuracy: 0.8278\n",
      "Epoch 22/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.4238 - binary_accuracy: 0.8234\n",
      "Epoch 23/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4265 - binary_accuracy: 0.8222\n",
      "Epoch 24/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4168 - binary_accuracy: 0.8269\n",
      "Epoch 25/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4131 - binary_accuracy: 0.8282\n",
      "Epoch 26/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4217 - binary_accuracy: 0.8240\n",
      "Epoch 27/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4189 - binary_accuracy: 0.8257\n",
      "Epoch 28/225\n",
      "1754/1754 [==============================] - 1s 419us/step - loss: 0.4183 - binary_accuracy: 0.8254\n",
      "Epoch 29/225\n",
      "1754/1754 [==============================] - 1s 429us/step - loss: 0.4197 - binary_accuracy: 0.8215\n",
      "Epoch 30/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4206 - binary_accuracy: 0.8249\n",
      "Epoch 31/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4244 - binary_accuracy: 0.8226\n",
      "Epoch 32/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4171 - binary_accuracy: 0.8273\n",
      "Epoch 33/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4189 - binary_accuracy: 0.8185\n",
      "Epoch 34/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4157 - binary_accuracy: 0.8241\n",
      "Epoch 35/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4158 - binary_accuracy: 0.8269\n",
      "Epoch 36/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4256 - binary_accuracy: 0.8232\n",
      "Epoch 37/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4154 - binary_accuracy: 0.8287\n",
      "Epoch 38/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4202 - binary_accuracy: 0.8242\n",
      "Epoch 39/225\n",
      "1754/1754 [==============================] - 1s 391us/step - loss: 0.4172 - binary_accuracy: 0.8206\n",
      "Epoch 40/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4147 - binary_accuracy: 0.8262\n",
      "Epoch 41/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4110 - binary_accuracy: 0.8270\n",
      "Epoch 42/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4123 - binary_accuracy: 0.8278\n",
      "Epoch 43/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4194 - binary_accuracy: 0.8290\n",
      "Epoch 44/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4217 - binary_accuracy: 0.8204\n",
      "Epoch 45/225\n",
      "1754/1754 [==============================] - 1s 404us/step - loss: 0.4071 - binary_accuracy: 0.8273\n",
      "Epoch 46/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4189 - binary_accuracy: 0.8270\n",
      "Epoch 47/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4176 - binary_accuracy: 0.8258\n",
      "Epoch 48/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4146 - binary_accuracy: 0.8222\n",
      "Epoch 49/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4105 - binary_accuracy: 0.8323\n",
      "Epoch 50/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4118 - binary_accuracy: 0.8277\n",
      "Epoch 51/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4157 - binary_accuracy: 0.8272\n",
      "Epoch 52/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4113 - binary_accuracy: 0.8270\n",
      "Epoch 53/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4071 - binary_accuracy: 0.8250\n",
      "Epoch 54/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4178 - binary_accuracy: 0.8316\n",
      "Epoch 55/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4113 - binary_accuracy: 0.8270\n",
      "Epoch 56/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4137 - binary_accuracy: 0.8283\n",
      "Epoch 57/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4117 - binary_accuracy: 0.8248\n",
      "Epoch 58/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.4175 - binary_accuracy: 0.8228\n",
      "Epoch 59/225\n",
      "1754/1754 [==============================] - 1s 404us/step - loss: 0.4136 - binary_accuracy: 0.8302\n",
      "Epoch 60/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4129 - binary_accuracy: 0.8263\n",
      "Epoch 61/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4086 - binary_accuracy: 0.8274\n",
      "Epoch 62/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4090 - binary_accuracy: 0.8245\n",
      "Epoch 63/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4187 - binary_accuracy: 0.8246\n",
      "Epoch 64/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4122 - binary_accuracy: 0.8253\n",
      "Epoch 65/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4139 - binary_accuracy: 0.8280\n",
      "Epoch 66/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.4119 - binary_accuracy: 0.8292\n",
      "Epoch 67/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.4134 - binary_accuracy: 0.8220\n",
      "Epoch 68/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4113 - binary_accuracy: 0.8244\n",
      "Epoch 69/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4147 - binary_accuracy: 0.8260\n",
      "Epoch 70/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4151 - binary_accuracy: 0.8250\n",
      "Epoch 71/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.4154 - binary_accuracy: 0.8271\n",
      "Epoch 72/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4119 - binary_accuracy: 0.8264\n",
      "Epoch 73/225\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4169 - binary_accuracy: 0.8264\n",
      "Epoch 74/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.4098 - binary_accuracy: 0.8264\n",
      "Epoch 75/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.4038 - binary_accuracy: 0.8289\n",
      "Epoch 76/225\n",
      "1754/1754 [==============================] - 1s 408us/step - loss: 0.4128 - binary_accuracy: 0.8276\n",
      "Epoch 77/225\n",
      "1754/1754 [==============================] - 1s 389us/step - loss: 0.4118 - binary_accuracy: 0.8298\n",
      "Epoch 78/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4035 - binary_accuracy: 0.8278\n",
      "Epoch 79/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4108 - binary_accuracy: 0.8251\n",
      "Epoch 80/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4088 - binary_accuracy: 0.8294\n",
      "Epoch 81/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4102 - binary_accuracy: 0.8305\n",
      "Epoch 82/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4099 - binary_accuracy: 0.8279\n",
      "Epoch 83/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4120 - binary_accuracy: 0.8234\n",
      "Epoch 84/225\n",
      "1754/1754 [==============================] - 1s 445us/step - loss: 0.4072 - binary_accuracy: 0.8254\n",
      "Epoch 85/225\n",
      "1754/1754 [==============================] - 1s 406us/step - loss: 0.4081 - binary_accuracy: 0.8303\n",
      "Epoch 86/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4135 - binary_accuracy: 0.8285\n",
      "Epoch 87/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4103 - binary_accuracy: 0.8253\n",
      "Epoch 88/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4109 - binary_accuracy: 0.8278\n",
      "Epoch 89/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4091 - binary_accuracy: 0.8243\n",
      "Epoch 90/225\n",
      "1754/1754 [==============================] - 1s 389us/step - loss: 0.4056 - binary_accuracy: 0.8313\n",
      "Epoch 91/225\n",
      "1754/1754 [==============================] - 1s 390us/step - loss: 0.4094 - binary_accuracy: 0.8283\n",
      "Epoch 92/225\n",
      "1754/1754 [==============================] - 1s 391us/step - loss: 0.4027 - binary_accuracy: 0.8320\n",
      "Epoch 93/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4010 - binary_accuracy: 0.8337\n",
      "Epoch 94/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4082 - binary_accuracy: 0.8292\n",
      "Epoch 95/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4042 - binary_accuracy: 0.8347\n",
      "Epoch 96/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4088 - binary_accuracy: 0.8290\n",
      "Epoch 97/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4042 - binary_accuracy: 0.8320\n",
      "Epoch 98/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4124 - binary_accuracy: 0.8246\n",
      "Epoch 99/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4102 - binary_accuracy: 0.8232\n",
      "Epoch 100/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4055 - binary_accuracy: 0.8297\n",
      "Epoch 101/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4087 - binary_accuracy: 0.8288\n",
      "Epoch 102/225\n",
      "1754/1754 [==============================] - 1s 389us/step - loss: 0.4119 - binary_accuracy: 0.8244\n",
      "Epoch 103/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4088 - binary_accuracy: 0.8266\n",
      "Epoch 104/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4100 - binary_accuracy: 0.8331\n",
      "Epoch 105/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4029 - binary_accuracy: 0.8324\n",
      "Epoch 106/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.4060 - binary_accuracy: 0.8289\n",
      "Epoch 107/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4062 - binary_accuracy: 0.8307\n",
      "Epoch 108/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4038 - binary_accuracy: 0.8331\n",
      "Epoch 109/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.3984 - binary_accuracy: 0.8342\n",
      "Epoch 110/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.4036 - binary_accuracy: 0.8333\n",
      "Epoch 111/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4150 - binary_accuracy: 0.8260\n",
      "Epoch 112/225\n",
      "1754/1754 [==============================] - 1s 389us/step - loss: 0.4045 - binary_accuracy: 0.8332\n",
      "Epoch 113/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4039 - binary_accuracy: 0.8365\n",
      "Epoch 114/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4179 - binary_accuracy: 0.8228\n",
      "Epoch 115/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4026 - binary_accuracy: 0.8314\n",
      "Epoch 116/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4023 - binary_accuracy: 0.8338\n",
      "Epoch 117/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4008 - binary_accuracy: 0.8252\n",
      "Epoch 118/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4038 - binary_accuracy: 0.8352\n",
      "Epoch 119/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4134 - binary_accuracy: 0.8287\n",
      "Epoch 120/225\n",
      "1754/1754 [==============================] - 1s 389us/step - loss: 0.4059 - binary_accuracy: 0.8329\n",
      "Epoch 121/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4003 - binary_accuracy: 0.8328\n",
      "Epoch 122/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4065 - binary_accuracy: 0.8294\n",
      "Epoch 123/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4043 - binary_accuracy: 0.8276\n",
      "Epoch 124/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4041 - binary_accuracy: 0.8320\n",
      "Epoch 125/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4058 - binary_accuracy: 0.8344\n",
      "Epoch 126/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4110 - binary_accuracy: 0.8275\n",
      "Epoch 127/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4022 - binary_accuracy: 0.8327\n",
      "Epoch 128/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4049 - binary_accuracy: 0.8303\n",
      "Epoch 129/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4054 - binary_accuracy: 0.8263\n",
      "Epoch 130/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4020 - binary_accuracy: 0.8290\n",
      "Epoch 131/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4003 - binary_accuracy: 0.8314\n",
      "Epoch 132/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4048 - binary_accuracy: 0.8290\n",
      "Epoch 133/225\n",
      "1754/1754 [==============================] - 1s 406us/step - loss: 0.4001 - binary_accuracy: 0.8313\n",
      "Epoch 134/225\n",
      "1754/1754 [==============================] - 1s 432us/step - loss: 0.3986 - binary_accuracy: 0.8358\n",
      "Epoch 135/225\n",
      "1754/1754 [==============================] - 1s 390us/step - loss: 0.4050 - binary_accuracy: 0.8268\n",
      "Epoch 136/225\n",
      "1754/1754 [==============================] - 1s 390us/step - loss: 0.4040 - binary_accuracy: 0.8298\n",
      "Epoch 137/225\n",
      "1754/1754 [==============================] - 1s 386us/step - loss: 0.4060 - binary_accuracy: 0.8284\n",
      "Epoch 138/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4081 - binary_accuracy: 0.8315\n",
      "Epoch 139/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4060 - binary_accuracy: 0.8270\n",
      "Epoch 140/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4012 - binary_accuracy: 0.8298\n",
      "Epoch 141/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4037 - binary_accuracy: 0.8330\n",
      "Epoch 142/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4054 - binary_accuracy: 0.8272\n",
      "Epoch 143/225\n",
      "1754/1754 [==============================] - 1s 391us/step - loss: 0.4042 - binary_accuracy: 0.8325\n",
      "Epoch 144/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.3971 - binary_accuracy: 0.8323\n",
      "Epoch 145/225\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.3973 - binary_accuracy: 0.8302\n",
      "Epoch 146/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.3978 - binary_accuracy: 0.8387\n",
      "Epoch 147/225\n",
      "1754/1754 [==============================] - 1s 391us/step - loss: 0.3995 - binary_accuracy: 0.8327\n",
      "Epoch 148/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.4008 - binary_accuracy: 0.8305\n",
      "Epoch 149/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4004 - binary_accuracy: 0.8336\n",
      "Epoch 150/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.3947 - binary_accuracy: 0.8362\n",
      "Epoch 151/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.4038 - binary_accuracy: 0.8277\n",
      "Epoch 152/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4054 - binary_accuracy: 0.8306\n",
      "Epoch 153/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.3988 - binary_accuracy: 0.8295\n",
      "Epoch 154/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4057 - binary_accuracy: 0.8360\n",
      "Epoch 155/225\n",
      "1754/1754 [==============================] - 1s 407us/step - loss: 0.4025 - binary_accuracy: 0.8269\n",
      "Epoch 156/225\n",
      "1754/1754 [==============================] - 1s 389us/step - loss: 0.4094 - binary_accuracy: 0.8265\n",
      "Epoch 157/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.3958 - binary_accuracy: 0.8392\n",
      "Epoch 158/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4068 - binary_accuracy: 0.8277\n",
      "Epoch 159/225\n",
      "1754/1754 [==============================] - 1s 390us/step - loss: 0.4001 - binary_accuracy: 0.8326\n",
      "Epoch 160/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4023 - binary_accuracy: 0.8321\n",
      "Epoch 161/225\n",
      "1754/1754 [==============================] - 1s 389us/step - loss: 0.3963 - binary_accuracy: 0.8380\n",
      "Epoch 162/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4020 - binary_accuracy: 0.8333\n",
      "Epoch 163/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4035 - binary_accuracy: 0.8309\n",
      "Epoch 164/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4044 - binary_accuracy: 0.8297\n",
      "Epoch 165/225\n",
      "1754/1754 [==============================] - 1s 403us/step - loss: 0.3981 - binary_accuracy: 0.8342\n",
      "Epoch 166/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4014 - binary_accuracy: 0.8262\n",
      "Epoch 167/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.3999 - binary_accuracy: 0.8318\n",
      "Epoch 168/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4000 - binary_accuracy: 0.8362\n",
      "Epoch 169/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4083 - binary_accuracy: 0.8279\n",
      "Epoch 170/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.3981 - binary_accuracy: 0.8382\n",
      "Epoch 171/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.4101 - binary_accuracy: 0.8249\n",
      "Epoch 172/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.3984 - binary_accuracy: 0.8315\n",
      "Epoch 173/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.4034 - binary_accuracy: 0.8253\n",
      "Epoch 174/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4004 - binary_accuracy: 0.8338\n",
      "Epoch 175/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4060 - binary_accuracy: 0.8304\n",
      "Epoch 176/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4086 - binary_accuracy: 0.8314\n",
      "Epoch 177/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4005 - binary_accuracy: 0.8295\n",
      "Epoch 178/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4011 - binary_accuracy: 0.8367\n",
      "Epoch 179/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.3953 - binary_accuracy: 0.8340\n",
      "Epoch 180/225\n",
      "1754/1754 [==============================] - 1s 391us/step - loss: 0.3933 - binary_accuracy: 0.8377\n",
      "Epoch 181/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.3971 - binary_accuracy: 0.8310\n",
      "Epoch 182/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4004 - binary_accuracy: 0.8301\n",
      "Epoch 183/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4064 - binary_accuracy: 0.8246\n",
      "Epoch 184/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4019 - binary_accuracy: 0.8323\n",
      "Epoch 185/225\n",
      "1754/1754 [==============================] - 1s 390us/step - loss: 0.3973 - binary_accuracy: 0.8350\n",
      "Epoch 186/225\n",
      "1754/1754 [==============================] - 1s 388us/step - loss: 0.4000 - binary_accuracy: 0.8368\n",
      "Epoch 187/225\n",
      "1754/1754 [==============================] - 1s 389us/step - loss: 0.3964 - binary_accuracy: 0.8297\n",
      "Epoch 188/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.3981 - binary_accuracy: 0.8273\n",
      "Epoch 189/225\n",
      "1754/1754 [==============================] - 1s 390us/step - loss: 0.3942 - binary_accuracy: 0.8358\n",
      "Epoch 190/225\n",
      "1754/1754 [==============================] - 1s 390us/step - loss: 0.3985 - binary_accuracy: 0.8310\n",
      "Epoch 191/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.3957 - binary_accuracy: 0.8367\n",
      "Epoch 192/225\n",
      "1754/1754 [==============================] - 1s 389us/step - loss: 0.4001 - binary_accuracy: 0.8308\n",
      "Epoch 193/225\n",
      "1754/1754 [==============================] - 1s 391us/step - loss: 0.3988 - binary_accuracy: 0.8330\n",
      "Epoch 194/225\n",
      "1754/1754 [==============================] - 1s 389us/step - loss: 0.3978 - binary_accuracy: 0.8306\n",
      "Epoch 195/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4070 - binary_accuracy: 0.8245\n",
      "Epoch 196/225\n",
      "1754/1754 [==============================] - 1s 389us/step - loss: 0.3997 - binary_accuracy: 0.8326\n",
      "Epoch 197/225\n",
      "1754/1754 [==============================] - 1s 391us/step - loss: 0.4012 - binary_accuracy: 0.8278\n",
      "Epoch 198/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.3962 - binary_accuracy: 0.8344\n",
      "Epoch 199/225\n",
      "1754/1754 [==============================] - 1s 390us/step - loss: 0.4012 - binary_accuracy: 0.8305\n",
      "Epoch 200/225\n",
      "1754/1754 [==============================] - 1s 440us/step - loss: 0.3972 - binary_accuracy: 0.8288\n",
      "Epoch 201/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.3972 - binary_accuracy: 0.8349\n",
      "Epoch 202/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.3980 - binary_accuracy: 0.8335\n",
      "Epoch 203/225\n",
      "1754/1754 [==============================] - 1s 388us/step - loss: 0.3964 - binary_accuracy: 0.8323\n",
      "Epoch 204/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4004 - binary_accuracy: 0.8356\n",
      "Epoch 205/225\n",
      "1754/1754 [==============================] - 1s 387us/step - loss: 0.3939 - binary_accuracy: 0.8312\n",
      "Epoch 206/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.3988 - binary_accuracy: 0.8309\n",
      "Epoch 207/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.3977 - binary_accuracy: 0.8283\n",
      "Epoch 208/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4031 - binary_accuracy: 0.8303\n",
      "Epoch 209/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.3952 - binary_accuracy: 0.8340\n",
      "Epoch 210/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.3949 - binary_accuracy: 0.8421\n",
      "Epoch 211/225\n",
      "1754/1754 [==============================] - 1s 390us/step - loss: 0.4028 - binary_accuracy: 0.8305\n",
      "Epoch 212/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.3970 - binary_accuracy: 0.8333\n",
      "Epoch 213/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.3902 - binary_accuracy: 0.8358\n",
      "Epoch 214/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4003 - binary_accuracy: 0.8307\n",
      "Epoch 215/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4007 - binary_accuracy: 0.8335\n",
      "Epoch 216/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.4001 - binary_accuracy: 0.8294\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 217/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4025 - binary_accuracy: 0.8312\n",
      "Epoch 218/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4017 - binary_accuracy: 0.8325\n",
      "Epoch 219/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4009 - binary_accuracy: 0.8265\n",
      "Epoch 220/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.3956 - binary_accuracy: 0.8356\n",
      "Epoch 221/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.3970 - binary_accuracy: 0.8299\n",
      "Epoch 222/225\n",
      "1754/1754 [==============================] - 1s 390us/step - loss: 0.3966 - binary_accuracy: 0.8317\n",
      "Epoch 223/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.3944 - binary_accuracy: 0.8316\n",
      "Epoch 224/225\n",
      "1754/1754 [==============================] - 1s 389us/step - loss: 0.3997 - binary_accuracy: 0.8294\n",
      "Epoch 225/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4041 - binary_accuracy: 0.8286\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 919us/step - loss: 0.4897 - binary_accuracy: 0.8085\n",
      "Epoch 2/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4486 - binary_accuracy: 0.8184\n",
      "Epoch 3/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4440 - binary_accuracy: 0.8187\n",
      "Epoch 4/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4393 - binary_accuracy: 0.8203\n",
      "Epoch 5/225\n",
      "1754/1754 [==============================] - 1s 390us/step - loss: 0.4235 - binary_accuracy: 0.8249\n",
      "Epoch 6/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4318 - binary_accuracy: 0.8195\n",
      "Epoch 7/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4249 - binary_accuracy: 0.8226\n",
      "Epoch 8/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.4297 - binary_accuracy: 0.8197\n",
      "Epoch 9/225\n",
      "1754/1754 [==============================] - 1s 421us/step - loss: 0.4349 - binary_accuracy: 0.8202\n",
      "Epoch 10/225\n",
      "1754/1754 [==============================] - 1s 424us/step - loss: 0.4223 - binary_accuracy: 0.8282\n",
      "Epoch 11/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.4265 - binary_accuracy: 0.8174\n",
      "Epoch 12/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4119 - binary_accuracy: 0.8291\n",
      "Epoch 13/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4216 - binary_accuracy: 0.8199\n",
      "Epoch 14/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4232 - binary_accuracy: 0.8216\n",
      "Epoch 15/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4208 - binary_accuracy: 0.8272\n",
      "Epoch 16/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4251 - binary_accuracy: 0.8210\n",
      "Epoch 17/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.4190 - binary_accuracy: 0.8265\n",
      "Epoch 18/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4228 - binary_accuracy: 0.8225\n",
      "Epoch 19/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.4239 - binary_accuracy: 0.8236\n",
      "Epoch 20/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.4194 - binary_accuracy: 0.8217\n",
      "Epoch 21/225\n",
      "1754/1754 [==============================] - 1s 390us/step - loss: 0.4188 - binary_accuracy: 0.8284\n",
      "Epoch 22/225\n",
      "1754/1754 [==============================] - 1s 391us/step - loss: 0.4160 - binary_accuracy: 0.8225\n",
      "Epoch 23/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4163 - binary_accuracy: 0.8268\n",
      "Epoch 24/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4135 - binary_accuracy: 0.8279\n",
      "Epoch 25/225\n",
      "1754/1754 [==============================] - 1s 404us/step - loss: 0.4178 - binary_accuracy: 0.8243\n",
      "Epoch 26/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4085 - binary_accuracy: 0.8294\n",
      "Epoch 27/225\n",
      "1754/1754 [==============================] - 1s 403us/step - loss: 0.4185 - binary_accuracy: 0.8240\n",
      "Epoch 28/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4200 - binary_accuracy: 0.8261\n",
      "Epoch 29/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4175 - binary_accuracy: 0.8224\n",
      "Epoch 30/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.4137 - binary_accuracy: 0.8278\n",
      "Epoch 31/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4152 - binary_accuracy: 0.8226\n",
      "Epoch 32/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.4128 - binary_accuracy: 0.8263\n",
      "Epoch 33/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4169 - binary_accuracy: 0.8247\n",
      "Epoch 34/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4134 - binary_accuracy: 0.8282\n",
      "Epoch 35/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4162 - binary_accuracy: 0.8202\n",
      "Epoch 36/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4077 - binary_accuracy: 0.8315\n",
      "Epoch 37/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.4085 - binary_accuracy: 0.8278\n",
      "Epoch 38/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4152 - binary_accuracy: 0.8268\n",
      "Epoch 39/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4097 - binary_accuracy: 0.8217\n",
      "Epoch 40/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.4114 - binary_accuracy: 0.8282\n",
      "Epoch 41/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4163 - binary_accuracy: 0.8250\n",
      "Epoch 42/225\n",
      "1754/1754 [==============================] - 1s 403us/step - loss: 0.4142 - binary_accuracy: 0.8252\n",
      "Epoch 43/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4094 - binary_accuracy: 0.8203\n",
      "Epoch 44/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4117 - binary_accuracy: 0.8265\n",
      "Epoch 45/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4079 - binary_accuracy: 0.8302\n",
      "Epoch 46/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4103 - binary_accuracy: 0.8292\n",
      "Epoch 47/225\n",
      "1754/1754 [==============================] - 1s 405us/step - loss: 0.4164 - binary_accuracy: 0.8230\n",
      "Epoch 48/225\n",
      "1754/1754 [==============================] - 1s 418us/step - loss: 0.4118 - binary_accuracy: 0.8259\n",
      "Epoch 49/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.4054 - binary_accuracy: 0.8320\n",
      "Epoch 50/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4120 - binary_accuracy: 0.8271\n",
      "Epoch 51/225\n",
      "1754/1754 [==============================] - 1s 443us/step - loss: 0.4085 - binary_accuracy: 0.8284\n",
      "Epoch 52/225\n",
      "1754/1754 [==============================] - 1s 406us/step - loss: 0.4069 - binary_accuracy: 0.8275\n",
      "Epoch 53/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.4099 - binary_accuracy: 0.8276\n",
      "Epoch 54/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4139 - binary_accuracy: 0.8286\n",
      "Epoch 55/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4059 - binary_accuracy: 0.8305\n",
      "Epoch 56/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4115 - binary_accuracy: 0.8316\n",
      "Epoch 57/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4149 - binary_accuracy: 0.8236\n",
      "Epoch 58/225\n",
      "1754/1754 [==============================] - 1s 390us/step - loss: 0.4064 - binary_accuracy: 0.8272\n",
      "Epoch 59/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.4090 - binary_accuracy: 0.8296\n",
      "Epoch 60/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.4088 - binary_accuracy: 0.8275\n",
      "Epoch 61/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.4079 - binary_accuracy: 0.8298\n",
      "Epoch 62/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.4055 - binary_accuracy: 0.8298\n",
      "Epoch 63/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4051 - binary_accuracy: 0.8296\n",
      "Epoch 64/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4042 - binary_accuracy: 0.8239\n",
      "Epoch 65/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4041 - binary_accuracy: 0.8316\n",
      "Epoch 66/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.4075 - binary_accuracy: 0.8268\n",
      "Epoch 67/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4102 - binary_accuracy: 0.8244\n",
      "Epoch 68/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4005 - binary_accuracy: 0.8354\n",
      "Epoch 69/225\n",
      "1754/1754 [==============================] - 1s 391us/step - loss: 0.4113 - binary_accuracy: 0.8297\n",
      "Epoch 70/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4093 - binary_accuracy: 0.8295\n",
      "Epoch 71/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4114 - binary_accuracy: 0.8278\n",
      "Epoch 72/225\n",
      "1754/1754 [==============================] - 1s 391us/step - loss: 0.4093 - binary_accuracy: 0.8297\n",
      "Epoch 73/225\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4079 - binary_accuracy: 0.8275\n",
      "Epoch 74/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4133 - binary_accuracy: 0.8258\n",
      "Epoch 75/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4099 - binary_accuracy: 0.8269\n",
      "Epoch 76/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4105 - binary_accuracy: 0.8221\n",
      "Epoch 77/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4084 - binary_accuracy: 0.8265\n",
      "Epoch 78/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4024 - binary_accuracy: 0.8343\n",
      "Epoch 79/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4109 - binary_accuracy: 0.8295\n",
      "Epoch 80/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.4060 - binary_accuracy: 0.8292\n",
      "Epoch 81/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4078 - binary_accuracy: 0.8302\n",
      "Epoch 82/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.4031 - binary_accuracy: 0.8325\n",
      "Epoch 83/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4139 - binary_accuracy: 0.8232\n",
      "Epoch 84/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4072 - binary_accuracy: 0.8280\n",
      "Epoch 85/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4056 - binary_accuracy: 0.8247\n",
      "Epoch 86/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.4121 - binary_accuracy: 0.8215\n",
      "Epoch 87/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4094 - binary_accuracy: 0.8247\n",
      "Epoch 88/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4088 - binary_accuracy: 0.8230\n",
      "Epoch 89/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4124 - binary_accuracy: 0.8219\n",
      "Epoch 90/225\n",
      "1754/1754 [==============================] - 1s 388us/step - loss: 0.4030 - binary_accuracy: 0.8290\n",
      "Epoch 91/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4054 - binary_accuracy: 0.8314\n",
      "Epoch 92/225\n",
      "1754/1754 [==============================] - 1s 437us/step - loss: 0.4037 - binary_accuracy: 0.8330\n",
      "Epoch 93/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4046 - binary_accuracy: 0.8290\n",
      "Epoch 94/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4019 - binary_accuracy: 0.8320\n",
      "Epoch 95/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4044 - binary_accuracy: 0.8308\n",
      "Epoch 96/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4078 - binary_accuracy: 0.8318\n",
      "Epoch 97/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4123 - binary_accuracy: 0.8267\n",
      "Epoch 98/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.3995 - binary_accuracy: 0.8339\n",
      "Epoch 99/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4114 - binary_accuracy: 0.8298\n",
      "Epoch 100/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.4050 - binary_accuracy: 0.8268\n",
      "Epoch 101/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4025 - binary_accuracy: 0.8340\n",
      "Epoch 102/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.4040 - binary_accuracy: 0.8279\n",
      "Epoch 103/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4087 - binary_accuracy: 0.8301\n",
      "Epoch 104/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4014 - binary_accuracy: 0.8241\n",
      "Epoch 105/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4048 - binary_accuracy: 0.8283\n",
      "Epoch 106/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4040 - binary_accuracy: 0.8335\n",
      "Epoch 107/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4066 - binary_accuracy: 0.8289\n",
      "Epoch 108/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.3955 - binary_accuracy: 0.8374\n",
      "Epoch 109/225\n",
      "1754/1754 [==============================] - 1s 404us/step - loss: 0.4037 - binary_accuracy: 0.8346\n",
      "Epoch 110/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4077 - binary_accuracy: 0.8262\n",
      "Epoch 111/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4055 - binary_accuracy: 0.8308\n",
      "Epoch 112/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4063 - binary_accuracy: 0.8306\n",
      "Epoch 113/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.3998 - binary_accuracy: 0.8266\n",
      "Epoch 114/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.3998 - binary_accuracy: 0.8279\n",
      "Epoch 115/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.4016 - binary_accuracy: 0.8302\n",
      "Epoch 116/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4062 - binary_accuracy: 0.8271\n",
      "Epoch 117/225\n",
      "1754/1754 [==============================] - 1s 403us/step - loss: 0.4079 - binary_accuracy: 0.8260\n",
      "Epoch 118/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4077 - binary_accuracy: 0.8304\n",
      "Epoch 119/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4027 - binary_accuracy: 0.8306\n",
      "Epoch 120/225\n",
      "1754/1754 [==============================] - 1s 403us/step - loss: 0.4131 - binary_accuracy: 0.8241\n",
      "Epoch 121/225\n",
      "1754/1754 [==============================] - 1s 404us/step - loss: 0.3979 - binary_accuracy: 0.8324\n",
      "Epoch 122/225\n",
      "1754/1754 [==============================] - 1s 408us/step - loss: 0.4042 - binary_accuracy: 0.8322\n",
      "Epoch 123/225\n",
      "1754/1754 [==============================] - 1s 414us/step - loss: 0.4066 - binary_accuracy: 0.8290\n",
      "Epoch 124/225\n",
      "1754/1754 [==============================] - 1s 392us/step - loss: 0.4004 - binary_accuracy: 0.8293\n",
      "Epoch 125/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.4027 - binary_accuracy: 0.8314\n",
      "Epoch 126/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4096 - binary_accuracy: 0.8236\n",
      "Epoch 127/225\n",
      "1754/1754 [==============================] - 1s 421us/step - loss: 0.4085 - binary_accuracy: 0.8275\n",
      "Epoch 128/225\n",
      "1754/1754 [==============================] - 1s 410us/step - loss: 0.3945 - binary_accuracy: 0.8369\n",
      "Epoch 129/225\n",
      "1754/1754 [==============================] - 1s 416us/step - loss: 0.4002 - binary_accuracy: 0.8335\n",
      "Epoch 130/225\n",
      "1754/1754 [==============================] - 1s 415us/step - loss: 0.3979 - binary_accuracy: 0.8357\n",
      "Epoch 131/225\n",
      "1754/1754 [==============================] - 1s 411us/step - loss: 0.3952 - binary_accuracy: 0.8353\n",
      "Epoch 132/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.3918 - binary_accuracy: 0.8352\n",
      "Epoch 133/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.4018 - binary_accuracy: 0.8257\n",
      "Epoch 134/225\n",
      "1754/1754 [==============================] - 1s 406us/step - loss: 0.3992 - binary_accuracy: 0.8307\n",
      "Epoch 135/225\n",
      "1754/1754 [==============================] - 1s 404us/step - loss: 0.3993 - binary_accuracy: 0.8326\n",
      "Epoch 136/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.3984 - binary_accuracy: 0.8281\n",
      "Epoch 137/225\n",
      "1754/1754 [==============================] - 1s 404us/step - loss: 0.4032 - binary_accuracy: 0.8332\n",
      "Epoch 138/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.3997 - binary_accuracy: 0.8267\n",
      "Epoch 139/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.3982 - binary_accuracy: 0.8315\n",
      "Epoch 140/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.3995 - binary_accuracy: 0.8335\n",
      "Epoch 141/225\n",
      "1754/1754 [==============================] - 1s 441us/step - loss: 0.4011 - binary_accuracy: 0.8304\n",
      "Epoch 142/225\n",
      "1754/1754 [==============================] - 1s 416us/step - loss: 0.3971 - binary_accuracy: 0.8319\n",
      "Epoch 143/225\n",
      "1754/1754 [==============================] - 1s 403us/step - loss: 0.4029 - binary_accuracy: 0.8282\n",
      "Epoch 144/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.3962 - binary_accuracy: 0.8364\n",
      "Epoch 145/225\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.4035 - binary_accuracy: 0.8315\n",
      "Epoch 146/225\n",
      "1754/1754 [==============================] - 1s 412us/step - loss: 0.3980 - binary_accuracy: 0.8330\n",
      "Epoch 147/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.3997 - binary_accuracy: 0.8281\n",
      "Epoch 148/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.4000 - binary_accuracy: 0.8292\n",
      "Epoch 149/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.4055 - binary_accuracy: 0.8295\n",
      "Epoch 150/225\n",
      "1754/1754 [==============================] - 1s 404us/step - loss: 0.4006 - binary_accuracy: 0.8351\n",
      "Epoch 151/225\n",
      "1754/1754 [==============================] - 1s 405us/step - loss: 0.4024 - binary_accuracy: 0.8315\n",
      "Epoch 152/225\n",
      "1754/1754 [==============================] - 1s 403us/step - loss: 0.4044 - binary_accuracy: 0.8290\n",
      "Epoch 153/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4006 - binary_accuracy: 0.8324\n",
      "Epoch 154/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.3989 - binary_accuracy: 0.8271\n",
      "Epoch 155/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.4070 - binary_accuracy: 0.8319\n",
      "Epoch 156/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.3999 - binary_accuracy: 0.8309\n",
      "Epoch 157/225\n",
      "1754/1754 [==============================] - 1s 404us/step - loss: 0.3924 - binary_accuracy: 0.8335\n",
      "Epoch 158/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4098 - binary_accuracy: 0.8272\n",
      "Epoch 159/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.3892 - binary_accuracy: 0.8329\n",
      "Epoch 160/225\n",
      "1754/1754 [==============================] - 1s 405us/step - loss: 0.4023 - binary_accuracy: 0.8298\n",
      "Epoch 161/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.3955 - binary_accuracy: 0.8311\n",
      "Epoch 162/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.4013 - binary_accuracy: 0.8320\n",
      "Epoch 163/225\n",
      "1754/1754 [==============================] - 1s 405us/step - loss: 0.4053 - binary_accuracy: 0.8286\n",
      "Epoch 164/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4035 - binary_accuracy: 0.8304\n",
      "Epoch 165/225\n",
      "1754/1754 [==============================] - 1s 403us/step - loss: 0.4001 - binary_accuracy: 0.8287\n",
      "Epoch 166/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.3970 - binary_accuracy: 0.8311\n",
      "Epoch 167/225\n",
      "1754/1754 [==============================] - 1s 407us/step - loss: 0.3993 - binary_accuracy: 0.8295\n",
      "Epoch 168/225\n",
      "1754/1754 [==============================] - 1s 403us/step - loss: 0.3998 - binary_accuracy: 0.8337\n",
      "Epoch 169/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4069 - binary_accuracy: 0.8266\n",
      "Epoch 170/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.3966 - binary_accuracy: 0.8305\n",
      "Epoch 171/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.3958 - binary_accuracy: 0.8349\n",
      "Epoch 172/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.3966 - binary_accuracy: 0.8350\n",
      "Epoch 173/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.3930 - binary_accuracy: 0.8375\n",
      "Epoch 174/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4033 - binary_accuracy: 0.8296\n",
      "Epoch 175/225\n",
      "1754/1754 [==============================] - 1s 388us/step - loss: 0.4006 - binary_accuracy: 0.8308\n",
      "Epoch 176/225\n",
      "1754/1754 [==============================] - 1s 391us/step - loss: 0.3965 - binary_accuracy: 0.8297\n",
      "Epoch 177/225\n",
      "1754/1754 [==============================] - 1s 390us/step - loss: 0.3992 - binary_accuracy: 0.8304\n",
      "Epoch 178/225\n",
      "1754/1754 [==============================] - 1s 391us/step - loss: 0.4044 - binary_accuracy: 0.8300\n",
      "Epoch 179/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.3942 - binary_accuracy: 0.8295\n",
      "Epoch 180/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.3960 - binary_accuracy: 0.8335\n",
      "Epoch 181/225\n",
      "1754/1754 [==============================] - 1s 390us/step - loss: 0.4033 - binary_accuracy: 0.8295\n",
      "Epoch 182/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.3982 - binary_accuracy: 0.8313\n",
      "Epoch 183/225\n",
      "1754/1754 [==============================] - 1s 405us/step - loss: 0.4051 - binary_accuracy: 0.8273\n",
      "Epoch 184/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.3944 - binary_accuracy: 0.8324\n",
      "Epoch 185/225\n",
      "1754/1754 [==============================] - 1s 396us/step - loss: 0.4008 - binary_accuracy: 0.8342\n",
      "Epoch 186/225\n",
      "1754/1754 [==============================] - 1s 390us/step - loss: 0.4002 - binary_accuracy: 0.8297\n",
      "Epoch 187/225\n",
      "1754/1754 [==============================] - 1s 387us/step - loss: 0.3961 - binary_accuracy: 0.8308\n",
      "Epoch 188/225\n",
      "1754/1754 [==============================] - 1s 388us/step - loss: 0.3934 - binary_accuracy: 0.8332\n",
      "Epoch 189/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.3968 - binary_accuracy: 0.8286\n",
      "Epoch 190/225\n",
      "1754/1754 [==============================] - 1s 441us/step - loss: 0.3966 - binary_accuracy: 0.8331\n",
      "Epoch 191/225\n",
      "1754/1754 [==============================] - 1s 404us/step - loss: 0.4016 - binary_accuracy: 0.8346\n",
      "Epoch 192/225\n",
      "1754/1754 [==============================] - 1s 393us/step - loss: 0.4041 - binary_accuracy: 0.8287\n",
      "Epoch 193/225\n",
      "1754/1754 [==============================] - 1s 408us/step - loss: 0.3962 - binary_accuracy: 0.8329\n",
      "Epoch 194/225\n",
      "1754/1754 [==============================] - 1s 397us/step - loss: 0.4051 - binary_accuracy: 0.8279\n",
      "Epoch 195/225\n",
      "1754/1754 [==============================] - 1s 408us/step - loss: 0.4021 - binary_accuracy: 0.8297\n",
      "Epoch 196/225\n",
      "1754/1754 [==============================] - 1s 400us/step - loss: 0.4007 - binary_accuracy: 0.8296\n",
      "Epoch 197/225\n",
      "1754/1754 [==============================] - 1s 412us/step - loss: 0.4049 - binary_accuracy: 0.8303\n",
      "Epoch 198/225\n",
      "1754/1754 [==============================] - 1s 405us/step - loss: 0.3948 - binary_accuracy: 0.8310\n",
      "Epoch 199/225\n",
      "1754/1754 [==============================] - 1s 408us/step - loss: 0.3905 - binary_accuracy: 0.8340\n",
      "Epoch 200/225\n",
      "1754/1754 [==============================] - 1s 407us/step - loss: 0.3988 - binary_accuracy: 0.8309\n",
      "Epoch 201/225\n",
      "1754/1754 [==============================] - 1s 405us/step - loss: 0.3966 - binary_accuracy: 0.8348\n",
      "Epoch 202/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.3990 - binary_accuracy: 0.8332\n",
      "Epoch 203/225\n",
      "1754/1754 [==============================] - 1s 404us/step - loss: 0.3885 - binary_accuracy: 0.8359\n",
      "Epoch 204/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.3982 - binary_accuracy: 0.8338\n",
      "Epoch 205/225\n",
      "1754/1754 [==============================] - 1s 422us/step - loss: 0.4000 - binary_accuracy: 0.8290\n",
      "Epoch 206/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.3934 - binary_accuracy: 0.8368\n",
      "Epoch 207/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.3917 - binary_accuracy: 0.8351\n",
      "Epoch 208/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.3990 - binary_accuracy: 0.8309\n",
      "Epoch 209/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.3974 - binary_accuracy: 0.8311\n",
      "Epoch 210/225\n",
      "1754/1754 [==============================] - 1s 394us/step - loss: 0.4000 - binary_accuracy: 0.8324\n",
      "Epoch 211/225\n",
      "1754/1754 [==============================] - 1s 388us/step - loss: 0.3920 - binary_accuracy: 0.8338\n",
      "Epoch 212/225\n",
      "1754/1754 [==============================] - 1s 391us/step - loss: 0.4002 - binary_accuracy: 0.8297\n",
      "Epoch 213/225\n",
      "1754/1754 [==============================] - 1s 403us/step - loss: 0.3955 - binary_accuracy: 0.8352\n",
      "Epoch 214/225\n",
      "1754/1754 [==============================] - 1s 403us/step - loss: 0.4074 - binary_accuracy: 0.8313\n",
      "Epoch 215/225\n",
      "1754/1754 [==============================] - 1s 398us/step - loss: 0.3972 - binary_accuracy: 0.8277\n",
      "Epoch 216/225\n",
      "1754/1754 [==============================] - 1s 409us/step - loss: 0.3907 - binary_accuracy: 0.8366\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 217/225\n",
      "1754/1754 [==============================] - 1s 405us/step - loss: 0.3903 - binary_accuracy: 0.8375\n",
      "Epoch 218/225\n",
      "1754/1754 [==============================] - 1s 404us/step - loss: 0.3969 - binary_accuracy: 0.8308\n",
      "Epoch 219/225\n",
      "1754/1754 [==============================] - 1s 402us/step - loss: 0.3909 - binary_accuracy: 0.8382\n",
      "Epoch 220/225\n",
      "1754/1754 [==============================] - 1s 399us/step - loss: 0.3954 - binary_accuracy: 0.8316\n",
      "Epoch 221/225\n",
      "1754/1754 [==============================] - 1s 395us/step - loss: 0.3941 - binary_accuracy: 0.8342\n",
      "Epoch 222/225\n",
      "1754/1754 [==============================] - 1s 389us/step - loss: 0.3953 - binary_accuracy: 0.8300\n",
      "Epoch 223/225\n",
      "1754/1754 [==============================] - 1s 388us/step - loss: 0.3911 - binary_accuracy: 0.8382\n",
      "Epoch 224/225\n",
      "1754/1754 [==============================] - 1s 389us/step - loss: 0.3952 - binary_accuracy: 0.8368\n",
      "Epoch 225/225\n",
      "1754/1754 [==============================] - 1s 401us/step - loss: 0.3930 - binary_accuracy: 0.8374\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": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "average precision:0.3507831646105487\n",
      "average recall:0.5914853241664627\n",
      "average specificity:0.755527543395597\n",
      "average F1:0.43823327416139934\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": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "max F1: 0.45875251509054327\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": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "unit: 32\tepoch: 225\tauroc_mean: 0.723179\tauroc_std: 0.012437\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
}