A NOVEL WEB USAGE MINING METHOD - Mining and Clustering of DAG Access Patterns Considering Page Browsing Time
Koichiro Mihara, Masahiro Terabe, Kazuo Hashimoto · 2008
In this paper, we propose a novel method to analyze web access logs. The proposed method defines a web access pattern as a DAG with page browsing time, and extracts the patterns using the closed frequent embedded DAG mining algorithm, DIGDAG. The proposed method succeeds in extracting as small number of patterns as necessary minimum, and enables more efficient analysis by clustering the extracted results.