Abstract: Through obtaining anonymous users’ browsing path set, on the basis of the new definition of path similarity, building the users’ browsing path similarity matrix, and based on this, designing anonymous users’ browsing path clustering algorithm to obtain the clustering result set and calculate various cluster centers, the typical anonymous users’ browsing path was got. Experimental results show that a typical anonymous users’ browsing path represents different types of users’ Web browsing path, which can be effective reference for Web site information recommendation.

Key words: Web log, anonymous user, similarity, path clustering, typical anonymous users’ browsing path

摘要： 通过获取的匿名用户浏览路径集，依据新的路径相似度定义，建立用户浏览路径相似度矩阵，并在此基础上设计实现了匿名用户浏览路径聚类算法，获得聚类结果集，并计算各类的中心，得到典型匿名用户路径。挖掘结果显示典型匿名用户路径代表了不同类用户网络浏览路径，可有效地作为网站信息推荐的依据。

关键词: Web日志, 匿名用户, 相似度, 路径聚类, 典型匿名用户浏览路径