Navigating the vast expanse of the World Wide Web requires personalized experiences. This research explores a novel approach to extracting typical user profiles from web access logs using a relational competitive fuzzy clustering method. Introducing the concept of a 'user session' as a sequence of web accesses, the study defines a new distance measure to capture website organization. The Competitive Agglomeration clustering algorithm is extended for relational data, creating CARD, which handles complex similarity measures. The algorithm successfully analyzed web server access logs, unveiling typical session profiles, suggesting the method may be used for e-commerce personalization or improved website architecture and highlights its relevance in the age of big data and personalized experiences.
This research in International Journal on Artificial Intelligence Tools aligns with the journal's focus on AI applications. By introducing CARD for relational data, the work fits the journal’s scope. The methodology used to analyze web server access logs contributes to user profiling in information systems. Thus, emphasizing its importance within the journal’s research direction.