Session-Based Recommender Systems

Session-based recommendation is concerned with the problem of tailoring item suggestions according to the short-term needs and assumed intents of the user. The input in this recommendation scenario consists of an often very short sequence of user interactions that are observed in an ongoing usage session, and in many cases longer-term preferences of the users are not available. Such problems are highly relevant in practice because (i) recommendations should often be made also to anonymous and first-time users and because (ii) the users’ intents can change from session to session. In this chapter, we first elaborate on practical application scenarios for session-based recommender systems, provide a characterization of the problem class, and outline key challenges. Afterwards, we review technical approaches to session-based recommendation and report common practices of evaluating such systems. The chapter ends with a discussion of open challenges and an outlook on future directions in the area.