Time Serial Model of User Behavior in Content-Based Service System

With the amounts of contents and users increasing, the content-based service systems have to face the overloading problem. Load balancing technology using the information of user behavior is often used to resolve this problem. The major purpose of this paper is to build a time serial model of user behavior in content-based service systems, and propose a dissimilar measurement for this model. Finally the experiment based on DTW algorithm approves the performance of this measurement.