摘要
在现代搜索系统中,由于网络信息的动态性和用户兴趣的迁移性,通常的检索技术已难以满足用户的个性化需求,个性化搜索势在必行.User Profile作为描述用户兴趣和爱好的载体和手段,是个性化搜索领域的重要组成部分之一.本文论述了User Profile的创建、学习、存储、更新及其在个性化搜索系统中发挥的作用等,分析在相关反馈过程中,利用User Profile进行查询扩展的具体过程.针对向量空间模型和概率模型,分别讨论了User Profile的更新问题.最后,展望了User Profile的发展方向,得出"基于本体的User Profile是目前和将来最具潜力的方法"的结论.
In current search systems, because of the dynamic Web information and migratory users＇ interests, general search techniques are difficult to meet the users＇ personalized requirements, and the personalized search techniques are imperative under the situation. As the carrier and method of describing users＇ interests and preferences, User Profile is one of the important parts of personalized search. In this paper, the creation, learning, storage, update, and its applications in personalized search of User Profile are talked over, and the process of using User Profile for query expansion in relevance feedback is analyzed. Moreover, the update of User Profile in Vector Space Model and Probability Model is discussed, respectively. Finally, the evolution of User Profile is explored and the prediction that the User Profile based on ontology is the most potential method in present and future is given.
出处
《小型微型计算机系统》
CSCD
北大核心
2008年第10期1903-1907,共5页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目(60573090,60673139)资助
