Table 3.

Comparison of means of classification recall, F1 score and AUC for the different parameters tested for multilayer perceptron (MLP).

Parameter	Value	Average recall	Average F1 score	Average AUC
Hidden layer	t	0.9746 a	0.9746 ab	0.9995 ab
	i	0.9757 a	0.9757 a	0.9996 a
	a	0.9717 b	0.9717 b	0.9994 b
	Significance	**	**	**
Activation function	tanh	0.9855 a	0.9855 a	0.9998 a
	relu	0.9837 a	0.9837 a	0.9998 a
	Logistic	0.9527 b	0.9527 b	0.9990 b
	Sign.	***	***	***
Warm start	True	0.9740	0.9739	0.9995
	False	0.9739	0.9739	0.9994
	Significance	n.s.	n.s.	n.s.

Dissimilar lowercase letters within the different parameter values represent statistically different means, using Tukey's range test at a significance level p = 0.05.

n.s., not significant (p ≥ 0.05);

^**p < 0.01.

^***p < 0.001.

AUC, area under the receiver operating characteristic curve. t, the number of neurons is the sum of the number of attributes and classes. i, the number of neurons is the number of attributes. a, the number of neurons is half the amount of t. tanh, hyperbolic tangent function. relu, rectified linear unit function. logistic, logistic sigmoid function.