【摘要】 设计了一种通用的、基于访问序列的网页推荐系统.通过在线与离线系统相分离的系统架构,并对用户访问序列进行预先聚类,所设计的系统具有较高的推荐效率.对一个互联网上真实网站的实验研究,发现所设计的导航系统仅需要收集较少的用户信息就可以取得较好的推荐效果.更多还原

【Abstract】 Websites have become the key resources of all sorts of information.But,with the scales growing larger and structures becoming more complicated for websites today,it is even harder for users to locate relevant information.This essay proposed a general webpage recommendation system based on user access patterns.The offline-online separated architecture,coupling with the pre-clustering of the visiting sequences,make the system very efficient in recommendation.Experimental results on an Internet website illustrate that the proposed system have satisfied system performance with minimum amount of information demands.更多还原