The NEWER system: How to exploit a neuro-fuzzy strategy for Web recommendation

Web recommendation is a promising technology aimed to predict the needs of users by suggesting them information or services retained interesting according to their preferences. Web recommendation finds in Soft Computing techniques a valid tool to handle with the uncertainty and the ambiguity characterizing the Web and all phases of user interactions with Web sites. the main rationale behind this success seems to be the complementary nature of Soft Computing paradigms that properly combined enable the development of hybrid schemes exploiting the potential of each single paradigm. In this paper, we present NEWER, a neuro-fuzzy Web recommendation system that dynamically suggests interesting pages to the current user. NEWER employs a neuro-fuzzy approach in order to determine categories of users sharing similar interests and to extract a recommendation model in the form of fuzzy rules expressing associations between user categories and relevances of pages. the derived model is used by an online recommendation module to dynamically suggest interesting links. Comparative accuracy results show the effectiveness of NEWER.