Utilizing Physical and Social Context to Improve Recommender Systems

Context has rarely been incorporated into recommender systems so far, but physical (e.g. a user´s location) or social (e.g. the social network of a user) context can be useful sources for improving recommender systems. In this paper, we first discuss some principles for context-awareness in recommender systems. Then we present our hybrid recommender system for recommending applications to users of mobile devices. Finally, we describe our approach to utilize social networks to enhance collaborative filtering. Our evaluation shows that the social recommender outperforms traditional collaborative filtering algorithms in our scenario.