Online Incremental Learning for High Bandwidth Network Traffic Classification

Algorithm 1

Classification model initialization.

(1) : Pre-collected flow instances (array of features)

(2) : True labels for

(3) : Distance of to origin

(4) : Number of micro-model

(5) : Number of Pre-collected flow instances

(6) : Number of cluster in a micro-model

(7) for  in range of  do

(8)  Calculate

(9) end for

(10) Quicksort

(11) Distribute into micro-model

(12) for each micro-model do

(13)  Generate cluster

(14)  Summarize cluster into Clustering Feature, CF

(15)  Store clusters in time-series and set timestamp to 0

(16) end for