基于智能Agent的多维权值信息检索模型

The Agent-Based Information Retrieval Model with Multi-weight Ranking Algorithm

为了弥补目前信息检索系统中存在的资源耗费大、实效性不高、用户满意度低等缺点,该文提出并实现一种新型的基于智能Agent的多维权值信息检索模型。该模型主要思想是在信息检索系统中应用了智能Agent技术,通过将检索任务分担到用户客户机、检索服务器和被检索主机的方式以达到提高检索实效性和节省系统和网络资源的目的;该文还通过综合考虑用户检索偏好等特征以及信息本身的重要度和检索匹配程度,提出了一种多维权值排序算法MWRA,提高检索的排序能力,从而为用户提供符合其个性化特征的检索结果。该文首先描述了基于智能Agent的多维权值信息检索模型,然后深入分析了多维权值排序算法MWRA,最后对模型的性能进行了比较和分析。实验结果和性能分析结论表明,基于智能Agent的多维权值信息检索模型的性能尤其在排序能力等方面有明显的提高。

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.