Online Intelligent Music Recommendation: The Opportunity and Challenge for People Well-Being Improvement

Abstract

In recent decades, fast development of contemporary digital entertainment and Internet technology has dramatically changed how people produce and consume music. This demands design and development of smart online music recommendation. In this paper, four important research (information retrieval and machine learning related) directions including content descriptor generation, personalization, playlist optimization and performance evaluation are identified and discussed to achieve effective online music recommendation. Based on a comprehensive analysis of the directions, we critically review and discuss the opportunities and challenges for IR research and emerging implications for real world practice. To further demonstrate its promises, we present case study on applying VenueMusic system to support people well-being improvement.