Table 4.

Parameter setting for BPNN.

Parameter	Value	Meaning
n	Features selected by correlation-based ensemble feature selector	Number of input nodes
H	(2n + 1)	Number of hidden nodes
H layer	1	Hidden layer
O	Linear	Output

Initial weights and bias are randomly assigned with small random variables ranging from −0.5 to 0.5, and the learning rate is kept as 0.5.