Web用户聚类中访问路径差异性度量方法的研究  

RESEARCH ON ACCESS PATHS' DIFFERENCE METRIC IN WEB USER CLUSTERING

作　　者：牛晓晨[1] 王劲松[1] 邢高峰[1] 

机构地区：[1]西安邮电大学计算机学院,陕西西安710121

出　　处：《计算机应用与软件》2013年第3期196-199,共4页Computer Applications and Software

摘　　要：用户访问路径是Web用户聚类时度量用户兴趣度的参数之一。结合日志记录的特点,从参数的数学特征出发,设计一种基于二进制数表示的访问路径矩阵来度量用户的兴趣度,并给出新的相异度矩阵构造法。实验表明,所提出的访问路径矩阵是可行的,生成的相异度矩阵在表现用户访问路径差异上是准确的。User＇s access path is one of the parameters for measuring user＇s interest degree when Web users clustering.In this paper,in combination with log recorders＇ features and proceeding from mathematical characteristics of parameter,we design an access path matrix represented by the binary to measure user＇s interest degree,and present the construction method of dissimilarity matrix.Experiment shows that the access path matrix proposed in the paper is feasible,and the dissimilarity matrix generated is accurate in representing the difference of user＇s access path.

关 键 词：WEB用户聚类 访问路径矩阵 相异度矩阵 日志文件 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]