A multi-agent personalized ontology profile based query refinement approach for information retrieval

This paper proposes a multi-agent query refinement approach to comprehensively track the users´ behaviors. Based on the approach, this paper creates the personalized ontology profile, with which it refines and expands the users´ initial query. We use four agents in our entire system. First, we use Client Agent to verify user identities by creating a union user account and to monitor whether a device is ready to run a personalized ontology profile creation task; then we use a Personalized Ontology Profile Agent to create a personalized ontology profile which not only can consider the Knowledge structure but also consider the users´ behavior. Finally, we use a three level-strategy to expand the query automatically, and send back the retrieval results to the user. Based on the users´ feedback, we reformulate the expanded query. We compare our method with the conceptual retrieval method as well as WordNet-based expansion, and we prove that our method has better average precision ratio.