A multi-agent personalized ontology profile based user preference profile construction method

This paper proposes a multi-agent personalized ontology profile based user preference profile construction method to comprehensively track the users´ long-term behaviors and short-term behaviors by comprehensively consider the users´ browsing behavior, pace times, as well as using two kinds of different strategies-“Sliding-Time-Window Update Strategy” and “Time-Based-Forgetting Function Update Strategy”, so as to construct a dynamic user preference profile based on the personalized ontology profile. There are four agents in our entire system, with all agents communicating among each other, and each agent having its own responsibility. The simulation result shows that our proposed method has a better performance than the pure sliding window method and the pure forgetting strategy method.