基于用户访问兴趣的Web日志挖掘  

Web log mining based on user's accessing interest

作　　者：何跃[1] 马丽霞[1] 腾格尔[1] 

机构地区：[1]四川大学工商管理学院,成都610064

出　　处：《系统工程理论与实践》2012年第6期1353-1361,共9页Systems Engineering-Theory & Practice

基　　金：国家自然科学基金(70771067)

摘　　要：Web日志挖掘可以发现访问者兴趣和需求,提出了一种改进的以访问时间、点击次数以及访问路径共同刻画用户的访问兴趣的Web日志挖掘算法.首先以Web日志为基础构建相关矩阵,使用平均访问时间相似度和访问路径相似度共同度量用户访问兴趣的相似程度,最后采用直接聚类去除相交项的聚类算法将相似用户和相关URL聚类.实证分析结果表明该算法能较好地解释用户的实际访问兴趣,从而为网站提供相应的运营建议.The Web log mining could obtain the web users＇ browsing interesting and their requirements, so an improved Web log mining algorithm which based on access time,browsing times and browsing paths is proposed.This study use Web log to create related matrixes,and then measure similarity of users＇ interest by considering both similarity of average browsing time and similarity of browsing paths.Finally, direct clustering algorithm is used to cluster the users of similar browsing interest and URL.The empirical study is based on real world Web log data.The result of the proposed algorithm can provide some useful recommendation on the website operation.

关 键 词：WEB日志挖掘 用户兴趣 相似度 聚类 

分 类 号：TP393.092[自动化与计算机技术—计算机应用技术]