Towards context-aware recommendations: Strategies for exploiting multi-criteria communities

Nowadays, recommender systems are becoming popular since they help users alleviate the information overload problem by offering personalized recommendations. Most systems apply collaborative filtering to predict individual preferences based on opinions of like-minded people through their ratings on items. Recently, context-aware recommender systems (CARSs) exploit additional context data such as time, place and so on for providing better recommendations. However, the large majority of CARSs use only ratings as a criterion for building communities, and ignore other available data allowing users to be grouped into communities. In this paper, we present a novel approach for exploiting multi-criteria communities to generate context-aware recommendations. The underlying idea of three proposed algorithms CRMC, CRESC and CREOC is that for each context, communities from the most suitable criteria followed by the learning phase are incorporated into the recommendation process. Experimental results show that our approach can improve the quality of recommendations.