A Graph-Based Web Usage Mining Considering Page Browsing Time

With the increase of large web sites which have complex link structures, web access logs have caught attention as a clue for web site administrators to understand user’s needs and demands. While conventional statistical analysis is used for most of the cases, web usage mining is an emerging attempt to apply data-mining based technique to web access log analyses. However, statistical and data-mining based analyses have been independently applied, and no method has been reported to correlate their results yet. This paper introduces a novel web usage mining method to combine the statistical analysis of page browsing time and the graph based data mining technique in order to extract users’ typical browsing behaviors.