Preprocessing Method for Encrypted Traffic Based on Semisupervised Clustering

Table 3

Comparison of the clustering results of the four clustering algorithms.

Clustering algorithm	Testing samples		Labeled benign samples		Labeled malicious samples		Features in BFS
	Benign	Malicious	Benign	Malicious	Benign	Malicious
K-means	733	5967	95	202	0	278	82
DBSCAN	1811	4889	204	93	0	278	25
GMM	4115	2385	229	68	129	149	93
UPGMA	2	6698	0	297	0	278	39