Leveraging multi-dimensional user models for personalized next-track music recommendation

The automated generation of playlists given a user's last played tracks is a common feature of modern music platforms. Existing approaches to this "next-track music recommendation" problem often focus solely on the user's recent listening behavior or current situational context and do not consider long-term preferences. In this work, we explore the value of including different types of information that reflect long-term preferences into the playlist generation process. Although empirical evaluations show that the most recently played tracks should generally govern the next-track selection process, considering long-term preferences can help to improve the quality of the playlists in different dimensions.1