Case Based Reasoning for Information Personalization: Using a Context-Sensitive Compositional Case Adaptation Approach

In this paper, we present an intelligent information filtering strategy that is a hybrid of item-based Collaborative Filtering (CF) and Case Based Reasoning (CBR) methods. Information filtering is implemented in two phases: in phase I, we have developed a multi-feature item-based CF strategy that allows creating a detailed context for filtering the information and retrieving N information objects based on user's interests and also preferred by similar users with similar tastes. In phase II, we use the N retrieved items as input to the CBR information filtering system and apply CBR-based compositional adaptation technique to selectively collect distinct information components of the N retrieved past items pairs to produce a composite recommendation that better addresses the initial user's interests and needs. We show that the hybrid of context-based similarity and compositional adaptation techniques improves significantly the quality of the recommendations presented to the user in terms of accurate and precise personalized information content.