Improving the Accuracy of Network Intrusion Detection with Causal Machine Learning

Table 18

Performance of different classifiers in CICIDS19 dataset under different types of cyberattacks.

Algorithm	Detection accuracy under different types of cyberattacks (1, 3, 7, 11)
	1	3	7	11
RS-KNN-CFS	0.9985	0.9953	0.9735	0.8917
TPE-KNN-CFS	0.9954	0.9942	0.9687	0.8793
RS-KNN-IGBS	0.9938	0.4577	0.4157	0.2887
TPE-KNN-IGBS	0.9864	0.3633	0.3024	0.2697
RS-RF-CFS	0.9986	0.9948	0.9676	0.8951
TPE-RF-CFS	0.9987	0.9947	0.9529	0.8921
RS-RF-IGBS	0.9928	0.4561	0.4170	0.2963
TPE-RF-IGBS	0.9883	0.4534	0.4033	0.2965
BRS	0.9985	0.9461	0.7869	0.7732
CMLN	0.9995	0.9993	0.9856	0.9852