Preprocessing Method for Encrypted Traffic Based on Semisupervised Clustering

Algorithm 1

SPEC.

	Input:X, , k,
	Output: SFSPEC–the ranked feature list
(1)	construct , the similarity set from X (and Y);
(2)	construct graph G from ;
(3)	build W, D and L from G；
(4)	for each feature vector do
(5)	;
(6)	end
(7)	ranking SFSPEC in ascending order for
(8)	return SFSPEC.