基于用户浏览兴趣的网页预取策略  

A Strategy for Pre-Fetching Based on Usage Information

作　　者：高凯[1] 王永成[1] 李刚[1] 

机构地区：[1]上海交通大学计算机科学与工程系,上海200240

出　　处：《上海交通大学学报》2006年第3期499-501,506,共4页Journal of Shanghai Jiaotong University

基　　金：国家高技术研究发展计划(863)项目(2002AA119050)

摘　　要：提出一种基于用户历史浏览情况对未来相关网页进行预取的算法.该算法能自适应调整用于生成会话的参数,并使用Session-tree结构完成对相关页面的预取.在实现中考虑到用户在特定事件和时间下的历史浏览行为,使用奖励因子和快速排序策略以提高效率.试验分析结果表明,基于自适应会话调整及群体共同兴趣剖像的预取算法能较好地完成对相关页面的预取.This paper proposed an approach to predictive prefetching, which consists of an adaptive session generation method to extract navigational behaviors from Web log and a Session-tree structure to organize sessions for fast retrieval. It presents an algorithm to segment historical information of individuals＇ browsing behaviors into sessions based on the similarity, and the session tree can retrieve those subsequences both by events and timestamps efficiently. The experimental results show that this method which is based on the adaptive session generation and the common interest profile can work well.

关 键 词：数据挖掘 网页预取 会话 兴趣剖像 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]