Table 15.

Experimental results on Hungarian dataset with 10 features

Mean ± SD	RF	LR	SVM	ELM	KNN	Proposed ensemble
E (%)	72.73 ± 6.29	73.85 ± 7.06	72.72 ± 6.78	69.94 ± 8.26	60.09 ± 10.59	79.87 ± 7.32
Precision (%)	72.72 ± 8.17	73.38 ± 8.14	71.78 ± 8.31	69.18 ± 10.08	53.77 ± 13.27	80.89 ± 7.89
Recall (%)	49.00 ± 16.03	52.92 ± 14.85	44.30 ± 17.06	44.39 ± 20.61	37.77 ± 18.40	66.38 ± 14.13
G-mean	62.75 ± 11.18	65.96 ± 10.60	60.44 ± 12.46	58.39 ± 14.78	45.48 ± 14.87	75.75 ± 9.22
MC (%)	109.40  ± 31.01	103.24 ± 32.31	118.24 ± 33.24	123.00 ± 42.83	148.60 ± 48.07	74.08 ± 32.11
Specificity (%)	82.62 ± 5.75	83.40 ± 5.22	85.57 ± 5.62	80.65 ± 8.26	59.28 ± 13.55	87.31 ± 3.60
AUC (%)	67.38 ± 10.99	68.59 ± 10.98	65.43 ± 10.99	61.67 ± 13.98	50.81 ± 15.55	77.64 ± 8.31

The average $+-$ sd on 10-folds CV. The best result is bolded