A hierarchical method for user´s behavior characteristics visualization and special user identification

With the high-speed development of WEB 2.0, the number of web applications continues to grow, making users´ behavior become increasingly complex and difficult to monitor. In this paper, we develop a new method for user´s behavior characteristics visualization and special user identification. Firstly, we divide all the web applications into 12 kinds and each kind includes several specific applications. Based on this classification, we develop a hierarchical behavior spectrum to visualize the user´s behavior easily and capture the user´s behavior characteristics very well. Secondly, we develop a method by using KL Divergence theory to measure the similarity of different users´ behavior and identify the special users whose behavior is pivotal for network management. The experimental results based on actual traffic traces show that the method proposed in this paper can visualize the users´ behavior easily and the accuracy rate of the special user identification is over 75%.