Context-aware music recommendation in mobile smart devices

Abstract

Mobile smart devices are now being used by many people as the primary means of listening to music. Based on the observation that a user's preferences for particular music pieces change with the contexts she/he faces, this paper proposes an approach to music recommendation that takes the target context into account by exploiting data gathered from a smart device. In addition, the proposed approach improves recommendations by including newly-added music pieces and newly-faced contexts, avoiding repeated recommendations of the same music piece, and reflecting a user's positive and negative feedback over time. We show the effectiveness of the proposed approach through extensive user study.