摘要： 随着网络上的知识不断暴涨，用户要求快速、有效地获取网络资源，该文提出了一种通过产生用户动态偏好来达到个性化搜索的方法，通过对搜索结果作一种基于内容的按名分类算法来为用户创建一个以RDF的形式表达的动态偏好。通过对RDF偏好文件的聚类来发现用户社区，产生相应的用户社区偏好。在搜索过程中，通过对该用户及其所属社区的偏好来分析。对搜索结果按重要性排序，达到了用户个性化搜索的目的。

关键词: 动态偏好, RDF, 聚类, 用户社区

Abstract: With the explosion of data on networks, it has been one of the important issues in Web search application that how to get Web resources to the user’s mind efficiently. We describe a new way of generating dynamic user profiles for personalization in Web resource search in this paper. It uses content-based Web search results classification to create a user profile, which is described by a RDF file. Such user characterizations are then used to find user groups based on cluster for those RDF files. Then dynamic profile and the traits of the corresponding user group in Web resources search can be used to provide a better service for the user.

Key words: Dynamic profile, RDF, Cluster, User Community

中图分类号: 

• TP393