A hybrid cost-sensitive ensemble for heart disease prediction

Table 12 Wilcoxon signed-rank test: classifiers with feature subset versus Classifiers with 13 features on Cleveland dataset

	RF	LR	SVM	ELM	KNN	Ensemble
E
Z-value\(^{1}\)	− 2.803	− 2.803	− 2.803	− 2.803	− 2.803	− 2.803
p value\(^{2}\)	\(0.002^{**}\)	\(0.002^{**}\)	\(0.002^{**}\)	\(0.002^{**}\)	\(0.002^{**}\)	\(0.002^{**}\)
Precision
Z-value	− 2.599	− 2.701	− 2.497	− 2.293	− 2.803	− 2.803
p value	\(0.006^{**}\)	\(0.004^{**}\)	\(0.010^{*}\)	\(0.020^{*}\)	\(0.002^{**}\)	\(0.002^{**}\)
Recall
Z-value	− 1.988	− 2.701	− 2.293	− 1.674	− 1.988	− 2.803
p value	\(0.049^{*}\)	\(0.004^{**}\)	\(0.020^{*}\)	0.132	\(0.049^{*}\)	\(0.002^{**}\)
G-mean
Z-value	− 2.497	− 2.701	− 2.497	− 2.293	− 1.784	− 2.803
p value	\(0.010^{*}\)	\(0.004^{**}\)	\(0.010^{*}\)	\(0.020^{*}\)	0.084	\(0.002^{**}\)
MC
Z-value	− 2.293	− 2.701	− 2.497	− 1.784	− 2.599	− 2.803
p value	\(0.020^{*}\)	\(0.004^{**}\)	\(0.010^{*}\)	0.084	\(0.006^{**}\)	\(0.002^{**}\)
Specificity
Z-value	− 2.803	− 2.803	− 2.802	− 2.702	− 2.701	− 2.803
p value	\(0.002^{**}\)	\(0.002^{**}\)	\(0.002^{**}\)	\(0.004^{**}\)	\(0.004^{**}\)	\(0.002^{**}\)
AUC
Z-value	− 2.701	− 2.701	− 2.803	− 2.090	− 2.803	− 2.803
p value	\(0.004^{**}\)	\(0.004^{**}\)	\(0.002^{**}\)	\(0.037^{*}\)	\(0.002^{**}\)	\(0.002^{**}\)

ISSN: 1472-6947