Justifying Results of the Music Recommender System HORST

Björn Forcher, Sahar Javadi, Christian Reuschling, Stephan Baumann, Andreas Dengel

In: Timour Klouche , Eduardo Miranda (Hrsg.). 9th Conference on Interdisciplinary Musicology. Conference on Interdisciplinary Musicology (CIM-14) 9th December 4-6 Berlin Germany Springer 2014.


The Music Recommender System Horst suggests musical entities, for instance bands or albums, which takes the users taste of music into account. The recommendation algorithm is based on semantic search technology and provides similar good results to traditional approaches such as collaborative or content-based filtering. A major benefit of the use of semantic technology is that results can be explained or justified by showing the connections between the interests of the users and the recommendations of the system by means of semantic networks. However, in a preliminary user survey we found out that justifications of this kind do not always satisfy the user's needs. In this paper, the effects of three factors of semantic networks are examined in a user experiment which influence user satisfaction: size of the network, edge label changes and diversity of edge directions.

Deutsches Forschungszentrum für Künstliche Intelligenz
German Research Center for Artificial Intelligence