基于关联规则的多层次Web聚类法研究

Multi-level Clustering of Web Pages and Users Based on Association Rules

提出了基于关联规则的多层次、超图分割聚类方法,用于对Web网页和用户进行有效聚类。该方法借助网站层次图,可以根据实际需要,在各个层次上进行聚类分析;在挖掘出满足一定Support的关联规则的基础上进行聚类,仅将高度相关的网页和用户聚在同一类,而将关联性较小的网页排除在聚类外。

A clustering method of Web pages and users based on association rules and hypergraph partitioning is proposed. Clustering can be done on multi-levels according to the specific needs with the help of the hierarchy diagrams of Web sites. The Web pages and users that are highly related are clustered together, excluding those which are not closely related.