Cardiovascular diseases are the most common cause of deaths globally, taking an estimated 17.9 million lives each year, which accounts for 31% of all deaths worldwide. Heart failure is a common event caused by Cardiovascular diseases. It is characterized by the heart’s inability to pump an adequate supply of blood to the body. Without sufficient blood flow, all major body functions are disrupted. Heart failure is a condition or a collection of symptoms that weaken the heart.

IMPORTING LIBRARIES

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from sklearn import preprocessing
from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import train_test_split
import seaborn as sns
from keras.layers import Dense, BatchNormalization, Dropout, LSTM
from keras.models import Sequential
from keras.utils import to_categorical
from keras import callbacks
from sklearn.metrics import precision_score, recall_score, confusion_matrix, classification_report, accuracy_score, f1_score

Using TensorFlow backend.

LOADING DATA

data = pd.read_csv("Desktop\heart_failure_clinical_records_dataset.csv")
data.head()

	age	anaemia	creatinine_phosphokinase	diabetes	ejection_fraction	high_blood_pressure	platelets	serum_creatinine	serum_sodium	sex	smoking	time	DEATH_EVENT
0	75.0	0	582	0	20	1	265000.00	1.9	130	1	0	4	1
1	55.0	0	7861	0	38	0	263358.03	1.1	136	1	0	6	1
2	65.0	0	146	0	20	0	162000.00	1.3	129	1	1	7	1
3	50.0	1	111	0	20	0	210000.00	1.9	137	1	0	7	1
4	65.0	1	160	1	20	0	327000.00	2.7	116	0	0	8	1

data.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 299 entries, 0 to 298
Data columns (total 13 columns):
 #   Column                    Non-Null Count  Dtype  
---  ------                    --------------  -----  
 0   age                       299 non-null    float64
 1   anaemia                   299 non-null    int64  
 2   creatinine_phosphokinase  299 non-null    int64  
 3   diabetes                  299 non-null    int64  
 4   ejection_fraction         299 non-null    int64  
 5   high_blood_pressure       299 non-null    int64  
 6   platelets                 299 non-null    float64
 7   serum_creatinine          299 non-null    float64
 8   serum_sodium              299 non-null    int64  
 9   sex                       299 non-null    int64  
 10  smoking                   299 non-null    int64  
 11  time                      299 non-null    int64  
 12  DEATH_EVENT               299 non-null    int64  
dtypes: float64(3), int64(10)
memory usage: 30.5 KB

About the data:

• age: Age of the patient
• anaemia: If the patient had the haemoglobin below the normal range
• creatinine_phosphokinase: The level of the creatine phosphokinase in the blood in mcg/L
• diabetes: If the patient was diabetic
• ejection_fraction: Ejection fraction is a measurement of how much blood the left ventricle pumps out with each contraction
• high_blood_pressure: If the patient had hypertension
• platelets: Platelet count of blood in kiloplatelets/mL
• serum_creatinine: The level of serum creatinine in the blood in mg/dL
• serum_sodium: The level of serum sodium in the blood in mEq/L
• sex: The sex of the patient
• smoking: If the patient smokes actively or ever did in past
• time: It is the time of the patient's follow-up visit for the disease in months
• DEATH_EVENT: If the patient deceased during the follow-up period

DATA ANALYSIS

#first of all let us evaluate the target and find out if our data is imbalanced or not
cols = ["#000080","#FF0000"]
sns.countplot(x = data["DEATH_EVENT"], palette = cols)

<AxesSubplot:xlabel='DEATH_EVENT', ylabel='count'>

cmap = sns.diverging_palette(250,150, s=50, l=65, n=9)
corrmat = data.corr()
plt.subplots(figsize=(15,15))
sns.heatmap(corrmat, cmap=cmap, annot=True, square=True)

<AxesSubplot:>

• Time of the patient's follow-up visit for the disease is crucial in as initial diagnosis with cardiovascular issue and treatment reduces the chances of any fatality. It holds and inverse relation.
• Ejection fraction is the second most important feature. It is quite expected as it is basically the efficiency of the heart.
• Age of the patient is the third most correlated feature. Clearly as heart's functioning declines with ageing

#Evaluating age distribution 
plt.figure(figsize=(20,12))
Days_of_weeks = sns.countplot(x = data["age"], data=data, hue = "DEATH_EVENT", palette = cols)
Days_of_weeks.set_title("Distribution Of Age")

ERROR! Session/line number was not unique in database. History logging moved to new session 47

Text(0.5, 1.0, 'Distribution Of Age')

# Boxen and swarm plot of some non binary features.
feature = ["age","creatinine_phosphokinase","ejection_fraction","platelets","serum_creatinine","serum_sodium", "time"]
for i in feature:
    plt.figure(figsize = (10,10))
    sns.swarmplot(x = data["DEATH_EVENT"], y = data[i], marker="o", s=4.5, color = "Black", alpha = 0.5)
    sns.boxenplot(x = data["DEATH_EVENT"], y = data[i], palette = cols)
    plt.show()

sns.kdeplot(x = data["time"], y = data["age"], hue = data["DEATH_EVENT"], palette = cols)

<AxesSubplot:xlabel='time', ylabel='age'>

data.describe().T

	count	mean	std	min	25%	50%	75%	max
age	299.0	60.833893	11.894809	40.0	51.0	60.0	70.0	95.0
anaemia	299.0	0.431438	0.496107	0.0	0.0	0.0	1.0	1.0
creatinine_phosphokinase	299.0	581.839465	970.287881	23.0	116.5	250.0	582.0	7861.0
diabetes	299.0	0.418060	0.494067	0.0	0.0	0.0	1.0	1.0
ejection_fraction	299.0	38.083612	11.834841	14.0	30.0	38.0	45.0	80.0
high_blood_pressure	299.0	0.351171	0.478136	0.0	0.0	0.0	1.0	1.0
platelets	299.0	263358.029264	97804.236869	25100.0	212500.0	262000.0	303500.0	850000.0
serum_creatinine	299.0	1.393880	1.034510	0.5	0.9	1.1	1.4	9.4
serum_sodium	299.0	136.625418	4.412477	113.0	134.0	137.0	140.0	148.0
sex	299.0	0.648829	0.478136	0.0	0.0	1.0	1.0	1.0
smoking	299.0	0.321070	0.467670	0.0	0.0	0.0	1.0	1.0
time	299.0	130.260870	77.614208	4.0	73.0	115.0	203.0	285.0
DEATH_EVENT	299.0	0.321070	0.467670	0.0	0.0	0.0	1.0	1.0

DATA PREPROCESSING

• Dropping the outliers based on data analysis
• Assigning values to features as X and target as y
• Perform the scaling of the features
• Split test and training sets

#assigning values to features as X and target as y
X=data.drop(["DEATH_EVENT"],axis=1)
y=data["DEATH_EVENT"]

ERROR! Session/line number was not unique in database. History logging moved to new session 56

#Set up a standard scaler for the features
col_names = list(X.columns)
s_scaler = preprocessing.StandardScaler()
X_df = s_scaler.fit_transform(X)
X_df = pd.DataFrame(X_df, columns=col_names)   
X_df.describe().T

	count	mean	std	min	25%	50%	75%	max
age	299.0	5.703353e-16	1.001676	-1.754448	-0.828124	-0.070223	0.771889	2.877170
anaemia	299.0	1.009969e-16	1.001676	-0.871105	-0.871105	-0.871105	1.147968	1.147968
creatinine_phosphokinase	299.0	0.000000e+00	1.001676	-0.576918	-0.480393	-0.342574	0.000166	7.514640
diabetes	299.0	9.060014e-17	1.001676	-0.847579	-0.847579	-0.847579	1.179830	1.179830
ejection_fraction	299.0	-3.267546e-17	1.001676	-2.038387	-0.684180	-0.007077	0.585389	3.547716
high_blood_pressure	299.0	0.000000e+00	1.001676	-0.735688	-0.735688	-0.735688	1.359272	1.359272
platelets	299.0	7.723291e-17	1.001676	-2.440155	-0.520870	-0.013908	0.411120	6.008180
serum_creatinine	299.0	1.425838e-16	1.001676	-0.865509	-0.478205	-0.284552	0.005926	7.752020
serum_sodium	299.0	-8.673849e-16	1.001676	-5.363206	-0.595996	0.085034	0.766064	2.582144
sex	299.0	-8.911489e-18	1.001676	-1.359272	-1.359272	0.735688	0.735688	0.735688
smoking	299.0	-1.188199e-17	1.001676	-0.687682	-0.687682	-0.687682	1.454161	1.454161
time	299.0	-1.901118e-16	1.001676	-1.629502	-0.739000	-0.196954	0.938759	1.997038

#looking at the scaled features
colours =["#FF0000","#FFFF00","#00FF00" ,"#00FFFF", "#FF00FF"]
plt.figure(figsize=(20,10))
sns.boxenplot(data = X_df, palette = colours)
plt.xticks(rotation = 90)
plt.show()

#spliting test and training sets
X_train, X_test, y_train,y_test = train_test_split(X_df,y,test_size=0.25,random_state=7)

MODEL BUILDING

• In this project, we build an artificial neural network.
  • Initialising the ANN
  • Defining by adding layers
  • Compiling the ANN
  • Train the ANN

early_stopping = callbacks.EarlyStopping(
    min_delta = 0.001,
    patience = 20,
    restore_best_weights = True)

# Initialising the ANN
model = Sequential()

# Layers
model.add(Dense(units = 16, kernel_initializer = "uniform", activation = "relu", input_dim = 12))
model.add(Dense(units = 8, kernel_initializer = 'uniform', activation = 'relu'))
model.add(Dropout(0.25))
model.add(Dense(units = 4, kernel_initializer = 'uniform', activation = 'relu'))
model.add(Dropout(0.5))
model.add(Dense(units = 1, kernel_initializer = 'uniform', activation = 'sigmoid'))
from keras.optimizers import SGD

# Compiling the ANN
model.compile(optimizer = 'adam', loss = 'binary_crossentropy', metrics = ['accuracy'])

# Train the ANN
history = model.fit(X_train, y_train, batch_size = 32, epochs = 500,callbacks=[early_stopping], validation_split=0.2)

Train on 179 samples, validate on 45 samples
Epoch 1/500
179/179 [==============================] - 0s 1ms/step - loss: 0.6928 - accuracy: 0.6425 - val_loss: 0.6922 - val_accuracy: 0.6667
Epoch 2/500
179/179 [==============================] - 0s 50us/step - loss: 0.6920 - accuracy: 0.6480 - val_loss: 0.6912 - val_accuracy: 0.6667
Epoch 3/500
179/179 [==============================] - 0s 50us/step - loss: 0.6911 - accuracy: 0.6480 - val_loss: 0.6902 - val_accuracy: 0.6667
Epoch 4/500
179/179 [==============================] - 0s 45us/step - loss: 0.6901 - accuracy: 0.6480 - val_loss: 0.6892 - val_accuracy: 0.6667
Epoch 5/500
179/179 [==============================] - 0s 45us/step - loss: 0.6893 - accuracy: 0.6480 - val_loss: 0.6880 - val_accuracy: 0.6667
Epoch 6/500
179/179 [==============================] - 0s 45us/step - loss: 0.6883 - accuracy: 0.6480 - val_loss: 0.6869 - val_accuracy: 0.6667
Epoch 7/500
179/179 [==============================] - 0s 45us/step - loss: 0.6875 - accuracy: 0.6480 - val_loss: 0.6856 - val_accuracy: 0.6667
Epoch 8/500
179/179 [==============================] - 0s 45us/step - loss: 0.6858 - accuracy: 0.6480 - val_loss: 0.6843 - val_accuracy: 0.6667
Epoch 9/500
179/179 [==============================] - 0s 45us/step - loss: 0.6846 - accuracy: 0.6480 - val_loss: 0.6826 - val_accuracy: 0.6667
Epoch 10/500
179/179 [==============================] - 0s 50us/step - loss: 0.6834 - accuracy: 0.6480 - val_loss: 0.6806 - val_accuracy: 0.6667
Epoch 11/500
179/179 [==============================] - 0s 45us/step - loss: 0.6816 - accuracy: 0.6480 - val_loss: 0.6786 - val_accuracy: 0.6667
Epoch 12/500
179/179 [==============================] - 0s 50us/step - loss: 0.6792 - accuracy: 0.6480 - val_loss: 0.6758 - val_accuracy: 0.6667
Epoch 13/500
179/179 [==============================] - 0s 45us/step - loss: 0.6760 - accuracy: 0.6480 - val_loss: 0.6722 - val_accuracy: 0.6667
Epoch 14/500
179/179 [==============================] - 0s 45us/step - loss: 0.6719 - accuracy: 0.6480 - val_loss: 0.6678 - val_accuracy: 0.6667
Epoch 15/500
179/179 [==============================] - 0s 50us/step - loss: 0.6687 - accuracy: 0.6480 - val_loss: 0.6622 - val_accuracy: 0.6667
Epoch 16/500
179/179 [==============================] - 0s 45us/step - loss: 0.6641 - accuracy: 0.6480 - val_loss: 0.6555 - val_accuracy: 0.6667
Epoch 17/500
179/179 [==============================] - 0s 45us/step - loss: 0.6527 - accuracy: 0.6480 - val_loss: 0.6474 - val_accuracy: 0.6667
Epoch 18/500
179/179 [==============================] - 0s 50us/step - loss: 0.6406 - accuracy: 0.6480 - val_loss: 0.6375 - val_accuracy: 0.6667
Epoch 19/500
179/179 [==============================] - 0s 50us/step - loss: 0.6358 - accuracy: 0.6480 - val_loss: 0.6265 - val_accuracy: 0.6667
Epoch 20/500
179/179 [==============================] - 0s 45us/step - loss: 0.6237 - accuracy: 0.6480 - val_loss: 0.6146 - val_accuracy: 0.6667
Epoch 21/500
179/179 [==============================] - 0s 50us/step - loss: 0.6136 - accuracy: 0.6480 - val_loss: 0.6020 - val_accuracy: 0.6667
Epoch 22/500
179/179 [==============================] - 0s 50us/step - loss: 0.6029 - accuracy: 0.6480 - val_loss: 0.5891 - val_accuracy: 0.6667
Epoch 23/500
179/179 [==============================] - 0s 45us/step - loss: 0.5934 - accuracy: 0.6480 - val_loss: 0.5770 - val_accuracy: 0.6667
Epoch 24/500
179/179 [==============================] - 0s 50us/step - loss: 0.5717 - accuracy: 0.6480 - val_loss: 0.5653 - val_accuracy: 0.6667
Epoch 25/500
179/179 [==============================] - 0s 67us/step - loss: 0.5750 - accuracy: 0.6480 - val_loss: 0.5558 - val_accuracy: 0.6667
Epoch 26/500
179/179 [==============================] - 0s 61us/step - loss: 0.5369 - accuracy: 0.6480 - val_loss: 0.5472 - val_accuracy: 0.6667
Epoch 27/500
179/179 [==============================] - 0s 50us/step - loss: 0.5503 - accuracy: 0.6480 - val_loss: 0.5403 - val_accuracy: 0.6667
Epoch 28/500
179/179 [==============================] - 0s 50us/step - loss: 0.5446 - accuracy: 0.6480 - val_loss: 0.5366 - val_accuracy: 0.6667
Epoch 29/500
179/179 [==============================] - 0s 50us/step - loss: 0.5416 - accuracy: 0.6480 - val_loss: 0.5328 - val_accuracy: 0.6667
Epoch 30/500
179/179 [==============================] - 0s 39us/step - loss: 0.5550 - accuracy: 0.6704 - val_loss: 0.5299 - val_accuracy: 0.6667
Epoch 31/500
179/179 [==============================] - 0s 50us/step - loss: 0.5488 - accuracy: 0.7263 - val_loss: 0.5269 - val_accuracy: 0.6667
Epoch 32/500
179/179 [==============================] - 0s 45us/step - loss: 0.5292 - accuracy: 0.7039 - val_loss: 0.5223 - val_accuracy: 0.7111
Epoch 33/500
179/179 [==============================] - 0s 45us/step - loss: 0.5361 - accuracy: 0.7430 - val_loss: 0.5185 - val_accuracy: 0.7333
Epoch 34/500
179/179 [==============================] - 0s 50us/step - loss: 0.5326 - accuracy: 0.7430 - val_loss: 0.5153 - val_accuracy: 0.7111
Epoch 35/500
179/179 [==============================] - 0s 45us/step - loss: 0.5308 - accuracy: 0.7542 - val_loss: 0.5114 - val_accuracy: 0.7333
Epoch 36/500
179/179 [==============================] - 0s 45us/step - loss: 0.5152 - accuracy: 0.7430 - val_loss: 0.5084 - val_accuracy: 0.7556
Epoch 37/500
179/179 [==============================] - 0s 50us/step - loss: 0.5210 - accuracy: 0.7430 - val_loss: 0.5062 - val_accuracy: 0.8000
Epoch 38/500
179/179 [==============================] - 0s 50us/step - loss: 0.5125 - accuracy: 0.7654 - val_loss: 0.5052 - val_accuracy: 0.8000
Epoch 39/500
179/179 [==============================] - 0s 50us/step - loss: 0.4906 - accuracy: 0.7933 - val_loss: 0.5041 - val_accuracy: 0.8000
Epoch 40/500
179/179 [==============================] - 0s 45us/step - loss: 0.5023 - accuracy: 0.7709 - val_loss: 0.5030 - val_accuracy: 0.8000
Epoch 41/500
179/179 [==============================] - 0s 45us/step - loss: 0.5138 - accuracy: 0.8156 - val_loss: 0.5011 - val_accuracy: 0.8222
Epoch 42/500
179/179 [==============================] - 0s 39us/step - loss: 0.5015 - accuracy: 0.7765 - val_loss: 0.4977 - val_accuracy: 0.8222
Epoch 43/500
179/179 [==============================] - 0s 39us/step - loss: 0.4787 - accuracy: 0.8212 - val_loss: 0.4944 - val_accuracy: 0.8000
Epoch 44/500
179/179 [==============================] - 0s 45us/step - loss: 0.4935 - accuracy: 0.7542 - val_loss: 0.4916 - val_accuracy: 0.8000
Epoch 45/500
179/179 [==============================] - 0s 45us/step - loss: 0.4716 - accuracy: 0.8101 - val_loss: 0.4910 - val_accuracy: 0.8000
Epoch 46/500
179/179 [==============================] - 0s 45us/step - loss: 0.4768 - accuracy: 0.8156 - val_loss: 0.4900 - val_accuracy: 0.8000
Epoch 47/500
179/179 [==============================] - 0s 39us/step - loss: 0.4824 - accuracy: 0.7933 - val_loss: 0.4902 - val_accuracy: 0.8000
Epoch 48/500
179/179 [==============================] - 0s 45us/step - loss: 0.4764 - accuracy: 0.8045 - val_loss: 0.4890 - val_accuracy: 0.8000
Epoch 49/500
179/179 [==============================] - 0s 39us/step - loss: 0.4644 - accuracy: 0.8324 - val_loss: 0.4888 - val_accuracy: 0.8222
Epoch 50/500
179/179 [==============================] - 0s 45us/step - loss: 0.4933 - accuracy: 0.7709 - val_loss: 0.4875 - val_accuracy: 0.8222
Epoch 51/500
179/179 [==============================] - 0s 50us/step - loss: 0.4360 - accuracy: 0.8156 - val_loss: 0.4874 - val_accuracy: 0.8222
Epoch 52/500
179/179 [==============================] - 0s 45us/step - loss: 0.4715 - accuracy: 0.7933 - val_loss: 0.4882 - val_accuracy: 0.8222
Epoch 53/500
179/179 [==============================] - 0s 56us/step - loss: 0.4684 - accuracy: 0.7933 - val_loss: 0.4881 - val_accuracy: 0.8222
Epoch 54/500
179/179 [==============================] - 0s 50us/step - loss: 0.4588 - accuracy: 0.8212 - val_loss: 0.4881 - val_accuracy: 0.8222
Epoch 55/500
179/179 [==============================] - 0s 45us/step - loss: 0.4544 - accuracy: 0.7821 - val_loss: 0.4886 - val_accuracy: 0.8000
Epoch 56/500
179/179 [==============================] - 0s 50us/step - loss: 0.4846 - accuracy: 0.7877 - val_loss: 0.4875 - val_accuracy: 0.8000
Epoch 57/500
179/179 [==============================] - 0s 39us/step - loss: 0.4262 - accuracy: 0.8156 - val_loss: 0.4865 - val_accuracy: 0.8222
Epoch 58/500
179/179 [==============================] - 0s 45us/step - loss: 0.4471 - accuracy: 0.8268 - val_loss: 0.4849 - val_accuracy: 0.8222
Epoch 59/500
179/179 [==============================] - 0s 45us/step - loss: 0.4192 - accuracy: 0.7989 - val_loss: 0.4842 - val_accuracy: 0.8222
Epoch 60/500
179/179 [==============================] - 0s 39us/step - loss: 0.4718 - accuracy: 0.8212 - val_loss: 0.4833 - val_accuracy: 0.8222
Epoch 61/500
179/179 [==============================] - 0s 45us/step - loss: 0.4486 - accuracy: 0.8156 - val_loss: 0.4821 - val_accuracy: 0.8222
Epoch 62/500
179/179 [==============================] - 0s 39us/step - loss: 0.4463 - accuracy: 0.7933 - val_loss: 0.4796 - val_accuracy: 0.8222
Epoch 63/500
179/179 [==============================] - 0s 45us/step - loss: 0.5011 - accuracy: 0.7877 - val_loss: 0.4781 - val_accuracy: 0.8222
Epoch 64/500
179/179 [==============================] - ETA: 0s - loss: 0.4291 - accuracy: 0.81 - 0s 45us/step - loss: 0.4378 - accuracy: 0.8212 - val_loss: 0.4786 - val_accuracy: 0.8222
Epoch 65/500
179/179 [==============================] - 0s 50us/step - loss: 0.4881 - accuracy: 0.8212 - val_loss: 0.4767 - val_accuracy: 0.8222
Epoch 66/500
179/179 [==============================] - 0s 45us/step - loss: 0.4707 - accuracy: 0.8156 - val_loss: 0.4760 - val_accuracy: 0.8222
Epoch 67/500
179/179 [==============================] - 0s 45us/step - loss: 0.4376 - accuracy: 0.8045 - val_loss: 0.4756 - val_accuracy: 0.8222
Epoch 68/500
179/179 [==============================] - 0s 50us/step - loss: 0.4238 - accuracy: 0.8268 - val_loss: 0.4773 - val_accuracy: 0.8222
Epoch 69/500
179/179 [==============================] - 0s 45us/step - loss: 0.4367 - accuracy: 0.8436 - val_loss: 0.4779 - val_accuracy: 0.8222
Epoch 70/500
179/179 [==============================] - 0s 50us/step - loss: 0.4601 - accuracy: 0.7989 - val_loss: 0.4784 - val_accuracy: 0.8222
Epoch 71/500
179/179 [==============================] - 0s 45us/step - loss: 0.4632 - accuracy: 0.7877 - val_loss: 0.4770 - val_accuracy: 0.8222
Epoch 72/500
179/179 [==============================] - 0s 50us/step - loss: 0.4282 - accuracy: 0.8212 - val_loss: 0.4773 - val_accuracy: 0.8222
Epoch 73/500
179/179 [==============================] - 0s 45us/step - loss: 0.4329 - accuracy: 0.8436 - val_loss: 0.4768 - val_accuracy: 0.8222
Epoch 74/500
179/179 [==============================] - 0s 39us/step - loss: 0.4578 - accuracy: 0.8101 - val_loss: 0.4759 - val_accuracy: 0.8222
Epoch 75/500
179/179 [==============================] - 0s 45us/step - loss: 0.4701 - accuracy: 0.7989 - val_loss: 0.4756 - val_accuracy: 0.8222
Epoch 76/500
179/179 [==============================] - 0s 45us/step - loss: 0.4314 - accuracy: 0.8324 - val_loss: 0.4754 - val_accuracy: 0.8222
Epoch 77/500
179/179 [==============================] - 0s 50us/step - loss: 0.4279 - accuracy: 0.8101 - val_loss: 0.4749 - val_accuracy: 0.8222
Epoch 78/500
179/179 [==============================] - 0s 50us/step - loss: 0.4386 - accuracy: 0.8156 - val_loss: 0.4739 - val_accuracy: 0.8222
Epoch 79/500
179/179 [==============================] - 0s 39us/step - loss: 0.4334 - accuracy: 0.7877 - val_loss: 0.4727 - val_accuracy: 0.8222
Epoch 80/500
179/179 [==============================] - 0s 45us/step - loss: 0.4583 - accuracy: 0.8156 - val_loss: 0.4732 - val_accuracy: 0.8222
Epoch 81/500
179/179 [==============================] - 0s 39us/step - loss: 0.4145 - accuracy: 0.7933 - val_loss: 0.4741 - val_accuracy: 0.8222
Epoch 82/500
179/179 [==============================] - 0s 45us/step - loss: 0.4100 - accuracy: 0.8268 - val_loss: 0.4748 - val_accuracy: 0.8222
Epoch 83/500
179/179 [==============================] - 0s 39us/step - loss: 0.4296 - accuracy: 0.8268 - val_loss: 0.4765 - val_accuracy: 0.8222
Epoch 84/500
179/179 [==============================] - 0s 39us/step - loss: 0.4375 - accuracy: 0.8324 - val_loss: 0.4779 - val_accuracy: 0.8222
Epoch 85/500
179/179 [==============================] - 0s 39us/step - loss: 0.4347 - accuracy: 0.8603 - val_loss: 0.4795 - val_accuracy: 0.8222
Epoch 86/500
179/179 [==============================] - 0s 45us/step - loss: 0.4434 - accuracy: 0.8492 - val_loss: 0.4779 - val_accuracy: 0.8222
Epoch 87/500
179/179 [==============================] - 0s 45us/step - loss: 0.4662 - accuracy: 0.7709 - val_loss: 0.4751 - val_accuracy: 0.8222
Epoch 88/500
179/179 [==============================] - 0s 45us/step - loss: 0.3888 - accuracy: 0.8212 - val_loss: 0.4731 - val_accuracy: 0.8222
Epoch 89/500
179/179 [==============================] - 0s 45us/step - loss: 0.4636 - accuracy: 0.7542 - val_loss: 0.4734 - val_accuracy: 0.8222
Epoch 90/500
179/179 [==============================] - 0s 45us/step - loss: 0.4496 - accuracy: 0.8380 - val_loss: 0.4742 - val_accuracy: 0.8000
Epoch 91/500
179/179 [==============================] - 0s 45us/step - loss: 0.4118 - accuracy: 0.8659 - val_loss: 0.4730 - val_accuracy: 0.8222
Epoch 92/500
179/179 [==============================] - 0s 45us/step - loss: 0.4347 - accuracy: 0.8156 - val_loss: 0.4713 - val_accuracy: 0.8222
Epoch 93/500
179/179 [==============================] - 0s 43us/step - loss: 0.4385 - accuracy: 0.8492 - val_loss: 0.4712 - val_accuracy: 0.8222
Epoch 94/500
179/179 [==============================] - 0s 45us/step - loss: 0.3707 - accuracy: 0.8380 - val_loss: 0.4717 - val_accuracy: 0.8222
Epoch 95/500
179/179 [==============================] - 0s 45us/step - loss: 0.4644 - accuracy: 0.7989 - val_loss: 0.4733 - val_accuracy: 0.8222
Epoch 96/500
179/179 [==============================] - 0s 45us/step - loss: 0.4458 - accuracy: 0.8045 - val_loss: 0.4734 - val_accuracy: 0.8222
Epoch 97/500
179/179 [==============================] - 0s 45us/step - loss: 0.4017 - accuracy: 0.8101 - val_loss: 0.4748 - val_accuracy: 0.8222
Epoch 98/500
179/179 [==============================] - 0s 45us/step - loss: 0.4157 - accuracy: 0.8436 - val_loss: 0.4768 - val_accuracy: 0.8222
Epoch 99/500
179/179 [==============================] - 0s 45us/step - loss: 0.4628 - accuracy: 0.8101 - val_loss: 0.4758 - val_accuracy: 0.8222
Epoch 100/500
179/179 [==============================] - 0s 44us/step - loss: 0.4317 - accuracy: 0.8268 - val_loss: 0.4754 - val_accuracy: 0.8222
Epoch 101/500
179/179 [==============================] - 0s 50us/step - loss: 0.4243 - accuracy: 0.8212 - val_loss: 0.4746 - val_accuracy: 0.8222
Epoch 102/500
179/179 [==============================] - 0s 45us/step - loss: 0.3967 - accuracy: 0.8268 - val_loss: 0.4747 - val_accuracy: 0.8222
Epoch 103/500
179/179 [==============================] - 0s 45us/step - loss: 0.4199 - accuracy: 0.8101 - val_loss: 0.4763 - val_accuracy: 0.8222
Epoch 104/500
179/179 [==============================] - 0s 45us/step - loss: 0.4141 - accuracy: 0.8380 - val_loss: 0.4770 - val_accuracy: 0.8222
Epoch 105/500
179/179 [==============================] - 0s 50us/step - loss: 0.4497 - accuracy: 0.7765 - val_loss: 0.4780 - val_accuracy: 0.8222
Epoch 106/500
179/179 [==============================] - 0s 45us/step - loss: 0.3986 - accuracy: 0.8212 - val_loss: 0.4776 - val_accuracy: 0.8222
Epoch 107/500
179/179 [==============================] - 0s 45us/step - loss: 0.4229 - accuracy: 0.8156 - val_loss: 0.4771 - val_accuracy: 0.8222
Epoch 108/500
179/179 [==============================] - 0s 45us/step - loss: 0.3862 - accuracy: 0.8101 - val_loss: 0.4784 - val_accuracy: 0.8222
Epoch 109/500
179/179 [==============================] - 0s 45us/step - loss: 0.4358 - accuracy: 0.7989 - val_loss: 0.4801 - val_accuracy: 0.8222
Epoch 110/500
179/179 [==============================] - 0s 39us/step - loss: 0.3982 - accuracy: 0.8101 - val_loss: 0.4817 - val_accuracy: 0.8222
Epoch 111/500
179/179 [==============================] - 0s 45us/step - loss: 0.4159 - accuracy: 0.8715 - val_loss: 0.4809 - val_accuracy: 0.8222
Epoch 112/500
179/179 [==============================] - 0s 45us/step - loss: 0.4330 - accuracy: 0.8268 - val_loss: 0.4802 - val_accuracy: 0.8222

val_accuracy = np.mean(history.history['val_accuracy'])
print("\n%s: %.2f%%" % ('val_accuracy', val_accuracy*100))

val_accuracy: 77.24%

history_df = pd.DataFrame(history.history)
plt.plot(history_df.loc[:, ['loss']], "#6daa9f", label='Training loss')
plt.plot(history_df.loc[:, ['val_loss']],"#774571", label='Validation loss')
plt.title('Training and Validation loss')
plt.xlabel('Epochs')
plt.ylabel('Loss')
plt.legend(loc="best")

plt.show()

ERROR! Session/line number was not unique in database. History logging moved to new session 59

history_df = pd.DataFrame(history.history)

plt.plot(history_df.loc[:, ['accuracy']], "#6daa9f", label='Training accuracy')
plt.plot(history_df.loc[:, ['val_accuracy']], "#774571", label='Validation accuracy')

plt.title('Training and Validation accuracy')
plt.xlabel('Epochs')
plt.ylabel('Accuracy')
plt.legend()
plt.show()

CONCLUSIONS

• Testing on the test set
• Evaluating the confusion matrix
• Evaluating the classification report

# Predicting the test set results
y_pred = model.predict(X_test)
y_pred = (y_pred > 0.5)
np.set_printoptions()

# confusion matrix
cmap1 = sns.diverging_palette(275,150,  s=40, l=65, n=6)
plt.subplots(figsize=(12,8))
cf_matrix = confusion_matrix(y_test, y_pred)
sns.heatmap(cf_matrix/np.sum(cf_matrix), cmap = cmap1, annot = True, annot_kws = {'size':15})

<AxesSubplot:>

print(classification_report(y_test, y_pred))

              precision    recall  f1-score   support

           0       0.86      0.88      0.87        57
           1       0.59      0.56      0.57        18

    accuracy                           0.80        75
   macro avg       0.73      0.72      0.72        75
weighted avg       0.80      0.80      0.80        75