Improving the Accuracy of Network Intrusion Detection with Causal Machine Learning

Table 7

Comparison of accuracy between SMOTE and CRFS in the CICIDS19 dataset.

Feature selection method	Accuracy under different types of cyberattacks (1, 3, 7, 11)
	1	3	7	11
SMOTE	0.9995	0.9962	0.9109	0.8758
SMOTE (do)	0.9995	0.9894	0.9042	0.8716
Percentage	1	0.9932	0.9926	0.9952