摘要
自适应网站能够根据用户需要快速灵活地改变自身,动态适应不断变化的用户需求和应用环境。本文基于图的频繁闭项集从站点一定时期内的日志中挖掘出闭相关页面集,据此提供在线动态推荐为用户导航,改善了传统的合作推荐存在的稀疏性和扩展性问题,在不增大网站服务器负荷的情况下提高对所有用户的信息服务质量。最后,分析了两种测量推荐系统性能的方法并对系统进行评价。
Adaptive web cites can automatically and flexibly improve their organization to fit users' requirements and application environments that are always changing continually. According to concept of frequent-close-item set based on graph, this paper provides method of on-line dynamic recommendation based on close correlative page set mined from websites' logs of a period. This method improves problems of sparsity and expansibility of traditional collaborative recommendation and enhances quality of service without increasing servers' load. Finally this paper analyzes two methods of measuring performance of recommendation system and evaluates the system.
出处
《计算机科学》
CSCD
北大核心
2006年第3期75-79,共5页
Computer Science
关键词
网站自适应
相关页面集
闭相关页面集
动态推荐
Adaptive web cite, Correlative page set, Close correlative page set, Dynamic recommendation