Application and Evaluation of Clustering Methods Using Swarm Intelligence to Traffic Classification

One approach to classify the Internet traffic is to use the statistical information of traffic. In this approach, the traffic is classified based on the statistical information by the machine learning. One of the popular machine learning techniques is clustering. In recent years, many clustering methods have been proposed. Among them, the bio-inspired clustering methods can classify the data accurately with a small computational complexity/ However, the applicability of such new clustering methods to the traffic classification is not sufficiently discussed. In this paper, we evaluate one of the bio-inspired clustering methods called AntTree when it is applied to the traffic classification. The results show that the AntTree can classify Web application traffic accurately even when it is applied to the online classification.