Content-Based Music Discovery
Music recommendation systems have become a valuable aid for managing large music collections and discovering new music. Our content-based recommendation system employs signal-based features and semantic music attributes generated using machine-based learning algorithms. In addition to playlist generation and music recommendation, we are exploring new usability concepts made possible by the analysis results. Functionality such as the mufin vision sound universe enables the user to discover his own music collection or even unknown catalogues in a new, more intuitive way.
Keywordsmusic visualization recommendation cloud clustering semantic attributes auto-tagging
Unable to display preview. Download preview PDF.
- 1.Bahanovich, D., Collopy, D.: Music Experience and Behaviour in Young People. University of Hertfordshire, UK (2009)Google Scholar
- 2.Celma, O.: Music Recommendation and Discovery in the Long Tail. PhD-Thesis, Universitat Pompeu Fabra, Spain (2008)Google Scholar
- 3.Nielsen Soundscan: State of the industrie (2007), http://www.narm.com/2008Conv/StateoftheIndustry.pdf (July 22, 2009)