文档介绍：3
1 1
2 2
2
计算机科学 2008Vol 35 № 3

基于用户浏览行为聚类 Web 用户)
陈敏苗夺谦段其国
(同济大学电子与信息工程学院上海 201804) (教育部嵌入式系统和服务计算重点实验室)

摘要本文结合 Web 用户浏览行为的特点,提出了一种新的路径相似度的计算方法,在计算相似度时不仅把用户
的浏览模式仅作为一种序列模式来考虑,还充分考虑了用户在网上浏览的时间因素。然后,把粗糙度的概念引入
Leader 聚类算法中,提出粗糙 Leader 聚类算法。最后,使用标准数据集进行了试验,证明基于此种相似度计算方法,
应用粗糙 Leader 算法聚类 Web 用户的有效性。
3 关键词 Web 日志挖掘,聚类,相似度,粗糙度

Clustering Web Users Based on Users’Browsing Action
CH EN Min MIAO Duo Qian DUAN Qi Quo
(Department puter Science and Engineering , Tongji University , Shanghai 201804)
( The Key Laboratory of Embedded System and puting ,Ministry of Education)

Abstract A novel method to get similitude actions of Web users is proposed in this paper after taking into account the
characteristics of users’browsing actions. The new similarity is defined according to not only the browsing pages but
also the time when users browse Web pages. Then , the concept of rough approximations is introduced in Leader cluster
algorithm and rough Leader cluster algorithm is suggested. Finally , the performance of the rough Leader cluster algo
rithm is tested and analyzed by benchmark based on the novel method puting the similarities of the web users’
access patterns.
Keywords Web usage mining , Clustering , Similarity , Rough approximations

中,提出粗糙 Leader 聚类算法。最后,使用标准数据集进行
1 引言
了试验,证明基于此种相似度计算方法,应用粗糙 Leader 算
作为 Web 智能(Web Intelli