Abstract

Todays' WUM techniques allow to perform the mining process based on lists of words, stems, and visitors' sessions. Although, results are (generally) far from visitors' real goals or motivations when browsing a web site. Thus, extraction of useful modifications of site organization or contents are difficult to obtain. It is needed a way to enhance the WUM process, to allow better results, closer to visitors' real preferences and goals. It was developed a Semantic WUM process, which uses a concept-based approach to add semantics into the mining process. The solution proposed, was applied to a real web site to produce offline enhancements of contents and structure. The method was compared with four different WUM methods. Afterwards, the quality of enhancements was evaluated using a survey to 100 subjects, proving the effectiveness of the proposal. © 2008 IEEE.