Feature Selection Based on Cross-Correlation for the Intrusion Detection System

Table 3

Comparison of the MIFS, CFA, and CCFS methods for DT, SVM, KNN, and NB classifiers on the NSL-KDD dataset with 20 features.

NSL-KDD (20 features)
Classification method	MIFS				CFS				CCFS
	DT	SVM	KNN	NB	DT	SVM	KNN	NB	DT	SVM	KNN	NB
Accuracy	91.21	87.80	89.79	87.35	93.12	90.15	91.50	89.24	93.47	90.69	93.09	90.87
Precision	93.52	88.17	89.66	89.27	94.82	91.18	92.46	90.34	94.87	92.39	94.39	92.27
Recall	93.45	87.21	89.61	89.27	94.82	91.18	92.46	91.18	95.79	92.12	94.35	92.27
F1-score	93.49	87.69	89.64	89.27	94.82	91.18	92.46	90.76	95.33	92.25	94.37	92.27