Mining of generalized web browsing patterns

Web mining of browsing patterns including simple sequential patterns and sequential patterns with browsing times has been studied recently However, most of these works focus on mining browsing patterns of web pages directly In this work, we introduce the problem of mining generalized browsing patterns on cross-levels of a taxonomy comprised of web pages. An algorithm called Apriori-Level is proposed to discover these cross-level generalized browsing patterns. Example demonstrating the proposed approach is given. Comparison with the Generalized Sequential Patterns (GSP) algorithm [16] shows that our approach generates fewer candidate sequences. The proposed algorithm thus promotes the efficiency of discovering coarser granularity of web browsing patterns.