Table 3.

Artificial neural network models’ architectures and accuracy metrics on the testing data.

Parameter	Model 1	Model 2	Model 3	Model 4
Input features	18	18	18	18
Hidden layers	4	4	2	2
Activation function	Sigmoid	ReLU	ReLU	Sigmoid
Accuracy metrics
Sensitivity	0.79	0.76	0.80	0.71
Positive predictive value	0.72	0.73	0.76	0.75
F1-Score	0.76	0.75	0.78	0.73
Accuracy (ACC)	0.74	0.74	0.77	0.73