Classification of Markov Encrypted Traffic on Gaussian Mixture Model Constrained Clustering

Algorithm 1

Input: Apply fingerprint.

Step1. Set to the equivalent set according to the restriction set by the IP address of the data packet;

Step2. Calculate the average value of ;

Step3. Calculate each sample in according to , and generate a new partition ;

Step4. Recalculate the mean ;

Step5. Repeat step 3 until the sample mean change does not exceed the threshold;

Step6. Recalculate .

Output: Division of application fingerprints .