论文部分内容阅读
用户兴趣建模是提升P2P文件共享系统上信息推荐、个性化检索等信息服务水平的重要途径。本文从节点用户共享的和用户查询的信息资源两方面出发,采用K中心点聚类方法和增量聚类方法,给出了基于双层聚类方法的P2P文件共享系统的用户兴趣建模方法。第一层采用K中心点聚类方法分析和识别节点用户共享的信息资源的内容主题,从而获得用户的部分兴趣主题,建立初始用户兴趣模型;在第一层聚类结果基础上,第二层采用增量聚类方法,考察节点用户的查询日志及其反馈的相关结果,识别用户的兴趣主题,从而完善初始用户兴趣模型;并给出了该用户兴趣建模的更新策略。实验结果表明,本文给出的P2P文档共享系统的用户兴趣模型能提高P2P文档共享系统的检索性能。
User interest modeling is an important way to improve the level of information service such as information recommendation and personalized search in P2P file sharing system. Based on K-center clustering method and incremental clustering method, this paper presents a method of user interest modeling in P2P file sharing system based on two-layer clustering method, which is shared by both node users and users. . The first layer uses K center clustering method to analyze and identify the content topics of information resources shared by node users, so as to obtain some of the user’s interest topics and establish an initial user interest model. Based on the first layer of clustering results, the second layer Adopting incremental clustering method, it investigates the query log of node users and the related results of their feedback, and identifies the user’s interest topic so as to perfect the initial user interest model; and gives the update strategy of the user interest modeling. The experimental results show that the user interest model of P2P document sharing system presented in this paper can improve the retrieval performance of P2P document sharing system.