【作者】 许欢庆； 王永成； 孙强；

【Author】 XU Huan-qing, WANG Yong-cheng, SUN Qiang (Department of Computer Science, Shanghai Jiao Tong University, Shanghai 200030, China)

【机构】 上海交通大学计算机系；

【摘要】 用户兴趣建模是互联网个性化信息服务的关键技术。本文中,提出一种基于加权概念网络的用户兴趣建模方法。该方法利用动态学习算法,挖掘蕴含在用户反馈文档中的概念及其概念关系,建立加权概念网络的用户模型,从而捕捉和表述用户兴趣偏好。基于加权概念网络用户兴趣模型,提出了检索提问个性化理解,以及文档个性化重评价的实现方法。为了检验提出方法的建模性能,设计了信息过滤仿真试验。测试结果表明:加权概念网络有较好的用户建模性能。更多还原

【Abstract】 User modeling is one of the crucial techniques of personalized information service. In this paper, we propose a new approach for user modeling based on Weighted Concept Network. This approach presents user’s preference using concepts and concept relations implied in documents proposed by the user’s relevance feedback. Incremental learning mechanism was applied to structure the Weighted Concept Network for user’s interests. The personalized query expansion and re_ranking algorithm based on Weighted Concept Network were implemented. The simulated experiment about information filtering was designed. Experimental results indicate that this approach has better performance in user modeling.更多还原