A Survey on Session-based Recommender Systems

Recommender systems (RSs) have been playing an increasingly important role for informed consumption, services, and decision-making in the overloaded information era digitized economy. In recent years, session-based recommender (SBRSs) emerged as a new paradigm of RSs. Different from other RSs such content-based collaborative filtering-based that usually model long-term yet static user preferences, SBRSs aim to capture short-term but dynamic preferences provide more timely accurate recommendations sensitive evolution their session contexts. Although intensively studied, neither unified problem statements nor in-depth elaboration SBRS characteristics challenges are available. It is also unclear what extent addressed overall research landscape is. This comprehensive review addresses above aspects by exploring depth entities (e.g., sessions), behaviours users’ clicks on items), properties length). We propose general statement SBRSs, summarize diversified data define taxonomy categorize representative research. Finally, we discuss opportunities this exciting vibrant area.