Table 9. Performance of various machine learning methods on NSL-KDD (Binary classification).

Attack type	Algorithm	Accuracy	Precision	Recall	F1-measure
DoS	KNN	0.997	0.996	0.996	0.996
	SVM	0.993	0.991	0.994	0.992
	Decision tree	0.994	0.991	0.994	0.992
	MLP	0.996	0.991	0.994	0.992
	Random forest	0.995	0.992	0.996	0.997
	Ensemble	0.998	0.998	0.997	0.997
Probe	KNN	0.990	0.986	0.985	0.985
	SVM	0.984	0.969	0.983	0.976
	Decision tree	0.995	0.969	0.983	0.976
	MLP	0.991	0.969	0.983	0.976
	Random forest	0.991	0.996	0.992	0.995
	Ensemble	0.997	0.987	0.989	0.988
R2L	KNN	0.967	0.953	0.954	0.953
	SVM	0.967	0.948	0.962	0.951
	Decision tree	0.979	0.948	0.962	0.949
	MLP	0.973	0.948	0.962	0.955
	Random forest	0.992	0.964	0.837	0.903
	Ensemble	0.973	0.959	0.964	0.961
U2R	KNN	0.997	0.931	0.850	0.878
	SVM	0.996	0.910	0.829	0.848
	Decision tree	0.996	0.910	0.829	0.848
	MLP	0.997	0.910	0.829	0.848
	Random forest	0.971	0.962	0.971	0.970
	Ensemble	0.9972	0.943	0.872	0.895