Improving the Accuracy of Network Intrusion Detection with Causal Machine Learning

Table 15

Comparison of accuracy between SMOTE and CRFS in the NSL-KDD dataset.

Feature selection method	Accuracy under different types of cyberattacks (1, 7)
	1	7
SMOTE	0.9951	0.9714
SMOTE (do)	0.9907	0.9701
Percentage	0.9956	0.9987