Table 2. Deep learning results with different combination of hidden layers.

Total trainable parameters (DNN Model)	TPR (sensitivity)	TNR (specifity)	FPR	FNR	Precision	Recall	Accuracy	AUC	WFM	Runtime (min:sec)
20152 (50,50)	0.946	0.987	0.129	0.053	0.987	0.992	0.980	0.955	0.982	00:40
195902 (300,300)	0.942	0.987	0.129	0.057	0.987	0.991	0.980	0.955	0.982	01:45
45352 (100,50,100)	0.954	0.981	0.181	0.045	0.982	0.993	0.980	0.940	0.981	00:50
166002 (300,100,300)	0.937	0.986	0.137	0.062	0.986	0.990	0.980	0.952	0.982	01:49
65502 (100,100,100,100)	0.942	0.986	0.137	0.057	0.988	0.991	0.980	0.957	0.982	00:52
376502 (300,300,300,300)	0.956	0.987	0.129	0.043	0.987	0.993	0.980	0.956	0.983	02:05
75602 (100,100,100,100,100)	0.942	0.986	0.013	0.058	0.986	0.991	0.980	0.953	0.982	00:54