Table 1.

Summary of configurations of classifiers.

Classifier	Parameters	Value
RF	Number of variables to sample	10
	Maximum number of splits	200
	Predictor selection criterion	“interaction curvature”
	Other parameters	Default
SVM	Model	error-correcting output code multiclass
	Kernel function	“rbf”
	Kernel scale	“auto”
	Standardize	true
	Other parameters	Default
MLP	Number of hidden layers	512
	Size of mini batch	300
	Optimizer	“adam”
	Maximum number of epochs	40
	Other parameters	Default
k-NN	Number of neighbors	3
	Other parameters	Default
NB	Data distributions	Multivariate multinomial distribution
	Other parameters	Default