Research Article
Preprocessing Method for Encrypted Traffic Based on Semisupervised Clustering
| Input: , , | | Output: Best feature subset (BFS), Clustering result (CR) | (1) | select a cluster algorithm | (2) | calculate initial confusion matrix for | | obtain FPRini, TPRini | (3) | for backward selection and exclude do | (4) | calculate confusion matrix for | (5) | if FPR < FPRlast: | (6) | exclude | (7) | FPRlast = FPR, TPRlast = TPR | (8) | if FPR = FPRlast and TPR ≥ TPRlast: | (9) | exclude | (10) | FPRlast = FPR, TPRlast = TPR | (11) | if FPR > FPRlast: | (12) | retain | (13) | end | (14) | obtain feature subset as BFS, CR for and | (15) | return BFS, CR. |
|