Table 5. Open in new tab The performance...

Table 5.

The performance of different ML classifiers. LC1 and LC2 are two ensemble classifiers used on LOTAAS 1 and LOTAAS 2, respectively (Tan et al. 2018). Bold type indicates the best performance on HTUR 2 and LOTAAS 1. ^aIndicates the best performance compared with PNCN and LC2.

Data set	Classifier	Accuracy	Recall	Precision	G-mean	F-score	FPR
	C4.5	0.946	0.904	0.635	0.926	0.740	0.051
	MLP	0.947	0.913	0.650	0.931	0.752	0.050
	NB	0.937	0.863	0.579	0.902	0.692	0.057
HTRU 2	SVM	0.871	0.901	0.723	0.919	0.789	0.031
	gh-vfdt	0.978	0.829	0.899	0.907	0.862	0.008
	KNN	0.978	0.825	0.930	0.906	0.875	0.006
	PNCN	0.978	0.831	0.923	0.908	0.874	0.007
	C4.5	0.990	0.948	0.494	0.969	0.623	0.009
	MLP	0.997	0.979	0.753	0.988	0.846	0.002
	NB	0.996	0.959	0.673	0.977	0.782	0.004
	SVM	0.999	0.901	0.966	0.949	0.932	0.001
LOTAAS 1	gh-vfdt	0.998	0.789	0.875	0.888	0.830	0.001
	KNN	0.999	0.947	0.978	0.973	0.961	0.0003
	PNCN	0.999^a	0.956	0.974	0.977	0.964	0.0003^a
	LC1	0.968	0.962	0.961	0.967	0.961	0.028
LOTAAS 2	LC2	0.992	0.987^a	0.993^a	0.991^a	0.990^a	0.005

Data set	Classifier	Accuracy	Recall	Precision	G-mean	F-score	FPR
	C4.5	0.946	0.904	0.635	0.926	0.740	0.051
	MLP	0.947	0.913	0.650	0.931	0.752	0.050
	NB	0.937	0.863	0.579	0.902	0.692	0.057
HTRU 2	SVM	0.871	0.901	0.723	0.919	0.789	0.031
	gh-vfdt	0.978	0.829	0.899	0.907	0.862	0.008
	KNN	0.978	0.825	0.930	0.906	0.875	0.006
	PNCN	0.978	0.831	0.923	0.908	0.874	0.007
	C4.5	0.990	0.948	0.494	0.969	0.623	0.009
	MLP	0.997	0.979	0.753	0.988	0.846	0.002
	NB	0.996	0.959	0.673	0.977	0.782	0.004
	SVM	0.999	0.901	0.966	0.949	0.932	0.001
LOTAAS 1	gh-vfdt	0.998	0.789	0.875	0.888	0.830	0.001
	KNN	0.999	0.947	0.978	0.973	0.961	0.0003
	PNCN	0.999^a	0.956	0.974	0.977	0.964	0.0003^a
	LC1	0.968	0.962	0.961	0.967	0.961	0.028
LOTAAS 2	LC2	0.992	0.987^a	0.993^a	0.991^a	0.990^a	0.005

This Feature Is Available To Subscribers Only