Table 4.

Prediction outcomes of hybrid features before and after BTG feature selection algorithm using 5-fold CV on the benchmark training dataset.

Method	Classifier	Acc (%)	Sen (%)	Sp (%)	MCC	AUC
Hybrid features	Bagging	89.90	87.39	92.91	0.79	0.95
	ETC	91.28	90.75	91.94	0.82	0.95
	XGB	89.99	87.81	92.12	0.80	0.95
	CatBoost	91.28	87.39	95.96	0.83	0.94
	DNN	95.39	95.48	95.16	0.90	0.97
Hybrid features + BTG	Bagging	91.28	90.76	91.92	0.82	0.95
	ETC	91.74	89.91	93.94	0.83	0.96
	XGB	92.20	90.75	93.94	0.84	0.96
	CatBoost	92.75	95.51	90.14	0.85	0.96
	DNN	96.84	96.92	98.66	0.93	0.98