基于人工蚁群的Web会话聚类  

Web sessions clustering based on artificial ants colonies

作　　者：周海斌[1] 

机构地区：[1]广西大学招生就业指导中心,广西壮族自治区南宁530004

出　　处：《重庆邮电学院学报（自然科学版）》2006年第5期657-659,共3页Journal of Chongqing University of Posts and Telecommunications(Natural Sciences Edition)

摘　　要：将改进的蚁群聚类算法应用于Web使用挖掘中,可对Web事务进行聚类,以便了解Web用户的兴趣以及它们之间的联系,从而为用户提供个性化的服务。同时定义一个Web会话为一个带权值的多维向量,也定义了两个会话间的相似度度量。实验表明在广西大学网站抽取的会话数据集上执行蚁群聚类算法得到的聚类是稳定的。结果显示该算法执行得很好,能找到没有噪音的聚类。In this paper, the author puts clustering algorithm into the application of Web usage mining for the clustering of Web affairs. By this way, the users＇ interests and the relations between them can be known, and then the personalized services will be provided. At the same time, the author defines a Web session as a weighted multi-dimension vector and also develops a similarity measure between two sessions. The experiment shows that the clusters gained by ant colony clustering algorithm are stable on a data set made of real sessions extracted from the Web site of the University of Guangxi. The result shows that find non-noisy clusters. this algorithm performs well and it is able to find non-noisy clusters.

关 键 词：WEB 使用挖掘 人工蚁群 聚类 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]