User Preferences Initialization and Integration in Critique-Based Mobile Recommender Systems

Recommender systems are decision support tools aimed at assisting users in finding products that best suit their preferences. The success of a recommendation session depends significantly on how, at the beginning of the interaction, the system initializes its representation of user’s preferences. In mobile recommender systems, guessing an initial representation of user’s preferences is even more difficult because of some limitations of mobile devices as well as characteristics of mobile users. In this paper we propose an approach for user preferences initialization that exploits a range of available knowledge sources related to the user. In this approach personalized recommendations can be generated using both a persistent and a context-dependent user model.