Table 2.

Effects of the varied hyper-parameters through a 10-fold cross-validation in terms of AUC based on the validation and test datasets.

Hyper-parameter	Parameter	Validation	Test
Kernel size	2	0.8310	0.8321
	3	0.8121	0.8172
Stride	1	0.8310	0.8321
	2	0.8089	0.8156
Number of neurons	25	0.8191	0.8232
	81	0.8310	0.8321
	169	0.8189	0.8236
Learning rate	0.01	0.8250	0.8296
	0.001	0.8310	0.8321
	0.0001	0.7763	0.7802
Dropout probability	0.1	0.8310	0.8321
	0.2	0.8196	0.8228
	0.3	0.8180	0.8227
Batch size	200	0.8166	0.8231
	250	0.8310	0.8321
	300	0.8135	0.8209
Activation function	ReLU_ReLU	0.8132	0.8224
	ReLU_Sigmoid	0.8127	0.8210
	ReLU_Tanh	0.8127	0.8242
	Sigmoid_ReLU	0.8224	0.8296
	Sigmoid_Sigmoid	0.8245	0.8301
	Sigmoid_Tanh	0.8271	0.8308
	Tanh_ReLU	0.8253	0.8297
	Tanh_Sigmoid	0.8245	0.8309
	Tanh_Tanh	0.8310	0.8321

FCNN model obtains the optimal AUC value, based on the different hyper-parameters combinations.