Table 3.

Table describing the median value of AUC, the optimal number of features, and AUC for best/worst features selection methods.

	AUC (median ± SD)	Optimal feature number	Best/worst feature selection method (AUC)
Naive Bayes	0.64 ± 0.028	5	ReliefFdistance (0.72)/EqualHellinger (0.62)
Random forest	0.63 ± 0.012	15	ReliefFdistance (0.69)/ImpurityHellinger (0.61)
K-Nearest neighbor	0.60 ± 0.23	20	ReliefFdistance (0.64)/gain ratio (0.55)