An Intrusion Detection Method Based on Decision Tree-Recursive Feature Elimination in Ensemble Learning

Table 7

Accuracy for each class of different methods.

Author	Dataset	DoS	Probe	R2L	U2R	Average
Our method	KDD CUP 99	0.9976	0.9941	0.9792	0.9974	0.9921
Our method	NSL-KDD	0.9974	0.9920	0.9821	0.9977	0.9923
Hussain	NSL-KDD	1.0000	0.9990	0.7740	0.8860	0.9148
Akashdeep	KDD CUP 99	0.9993	0.9879	0.9190	0.8660	0.9431
Jia	KDD CUP 99	0.9990	0.9818	0.9706	0.8182	0.9424
Jia	NSL-KDD	0.9867	0.9773	0.9694	0.8182	0.9379
Sun	KDD CUP 99	0.9741	0.9313	0.1124	0.2542	0.5680
Andresini	KDD CUP 99	—	—	—	—	0.9249
Jiang	NSL-KDD	0.9621	0.6856	0.6045	0.6132	0.8358