Demo paper: The BBC Desktop Jukebox music recommendation system: A large scale trial with professional users
暂无分享,去创建一个
Recommending music for professional use presents challenges that are substantially different from those faced by systems targeting recreational listeners and other classes of end users. This paper describes a trial system and survey for assessing the utility of content and metadata-based recommendation technology targeting television and radio producers. First, we briefly assess the applicability of existing recommendation technologies. We then describe the trial system and the applied recommendation methodologies used in the context of a music database exceeding one million tracks. Finally we draw conclusions from a small user study conducted with professional programme producers.
[1] Paul Lamere,et al. Music recommendation and discovery revisited , 2011, RecSys '11.
[2] Daniel P. W. Ellis,et al. Support vector machine active learning for music retrieval , 2006, Multimedia Systems.
[3] Mark Levy,et al. Lightweight measures for timbral similarity of musical audio , 2006, AMCMM '06.