Abstract:
In this article, a brand-new information retrieval model based on Agent technology is proposed, which is to counteract some significant deficiencies existing in current information retrieval systems, such as resource consuming much, information updating delayed and so on. The main idea of this model is to apply Agent technology into the information retrieval system in order to provide users a information retrieval model of new pattern, which could update on time, save resource through distributing the information retrieval task among clients, retrieval severs and information owners. And MWRA(Multi-weight Ranking Algorithm) is also proposed in this article to improve the ranking capability of the information retrieval system, which is based on several facts, including the inclination of users, the importance of information and the matching of query. In this article, we firstly introduced the agent-based information retrieval model. Then the multi-weight ranking algorithm, which included in the model, is analyzed. Finally, the performance of the model is discussed and the prototype system is tested through a series of examinations, the result of which addresses the Agent-based information retrieval model with MWRA is better in ranking and other capabilities.