Suggestion of interesting Web pages by a Neuro-fuzzy system

In this paper we explore the use of a neuro-fuzzy strategy to develop a Web personalization system that dynamically suggests interesting page categories for the current user. As a preliminary step, log files are analyzed to identify user access patters. Then, groups of users which exhibit a common browser behavior (i.e. user categories) are discovered by applying a fuzzy clustering algorithm to the user access patterns. Finally, a hybrid approach based on the combination of the fuzzy reasoning and the neural paradigm is applied in order to derive fuzzy associations between user categories and Web pages categories to be suggested to users. The derived knowledge is ultimately used by an online recommendation module to dynamically point out Web pages in the category judged interesting for the current user. Some preliminary experimental results are presented on a specific Web site.