Mining history of changes to web access patterns

摘要: Recently, a lot of work has been done in web usage mining [2]. Among them, frequent Web Access Pattern (WAP) is the most well researched issue [1]. The idea to transform logs into sequences events with user identifications and timestamps, then extract association sequential patterns from data certain metrics. WAPs have applied wide range applications such as personalization, system improvement, site modification, business intelligence, characterization However, existing techniques focus only on WAP snapshot data, while dynamic real life. While are useful many applications, knowledge hidden behind historical changes which reflects how change, also critical adaptive web, maintenance, etc.In this paper, we propose novel approach discover WAPs. Rather than focusing occurrence WAPs, frequently changing access patterns. We define type knowledge, Frequent Mutating (FM-WAP), based FM-WAP process consists three phases. Firstly, represented set trees partitioned sequence groups ( subsets trees) according user-defined calendar pattern, where each group forest. Consequently, log by forests called history. Then, among history detected stored global Finally, extracted traversal Extensive experiments show that our proposed can produce efficiently good scalability.