Weighing the role of multi-criteria communities for recommender systems

Recommender systems (RSs) have been designed to deal with the information overload problem by providing users with personalised recommendations, and now are becoming increasingly popular. Most RSs are based on collaborative filtering which is a technique predicting users' preferences by using opinions of like-minded users through their ratings on items. Recently, context-aware recommender systems (CARSs) have been developed to exploit additionally contextual information such as time, place, weather and so forth for providing better recommendations. However, the majority of CARSs work on ratings as a unique criterion for building communities and ignore other available data. This paper focuses on integrating multi-criteria communities into CARSs to enhance the context-aware recommendation quality. The integration of multi-criteria communities could allow users to take advantage of different natures of communities rather than exploiting only single-criterion rating ones. The experiments show that our proposed method outperforms comparative context-aware approaches.