Supporting Folk-Song Research by Automatic Metric Learning and Ranking

In folk song research, appropriate similarity measures can be of great help, e.g. for classification of new tunes. Several measures have been developed so far. However, a particular musicological way of classifying songs is usually not directly reflected by just a single one of these measures. We show how a weighted linear combination of different basic similarity measures can be automatically adapted to a specific retrieval task by learning this metric based on a special type of constraints. Further, we describe how these constraints are derived from information provided by experts. In experiments on a folk song database, we show that the proposed approach outperforms the underlying basic similarity measures and study the effect of different levels of adaptation on the performance of the retrieval system.

[1]  M. Fatih Demirci,et al.  Indexing Music Collections Through Graph Spectra , 2007, ISMIR.

[2]  George A. Tsihrintzis,et al.  MUSIPER: a system for modeling music similarity perception based on objective feature subset selection , 2008, User Modeling and User-Adapted Interaction.

[3]  Andreas Nürnberger,et al.  Personalized Hierarchical Clustering , 2006, 2006 IEEE/WIC/ACM International Conference on Web Intelligence (WI 2006 Main Conference Proceedings)(WI'06).

[4]  Stephan Baumann,et al.  AN ECOLOGICAL APPROACH TO MULTIMODAL SUBJECTIVE MUSIC SIMILARITY PERCEPTION , 2004 .

[5]  Bryan Pardo,et al.  A Query by Humming System that Learns from Experience , 2007, ISMIR.

[6]  Daniel P. W. Ellis,et al.  Support vector machine active learning for music retrieval , 2006, Multimedia Systems.

[7]  Thomas Mandl,et al.  Learning to Cope With Diversity in Music Retrieval , 2003, ISMIR.

[8]  Contents , 2009, Physiology & Behavior.

[9]  Andreas Nürnberger,et al.  Creating a Cluster Hierarchy under Constraints of a Partially Known Hierarchy , 2008, SDM.

[10]  Padraig Cunningham,et al.  Experimenting with music taste prediction by user profiling , 2004, MIR '04.

[11]  Daniel Müllensiefen,et al.  Cognitive Adequacy in the Measurement of Melodic Similarity: Algorithmic vs. Human Judgments , 2004 .

[12]  Keiichiro Hoashi,et al.  Personalization of user profiles for content-based music retrieval based on relevance feedback , 2003, ACM Multimedia.

[13]  J. J. Rocchio,et al.  Relevance feedback in information retrieval , 1971 .

[14]  Remco C. Veltkamp,et al.  A Manual Annotation Method for Melodic Similarity and the Study of Melody Feature Sets , 2008, ISMIR.

[15]  Agnar Aamodt,et al.  Case-Based Reasoning: Foundational Issues, Methodological Variations, and System Approaches , 1994, AI Commun..

[16]  Andreas Nürnberger,et al.  Towards User-Adaptive Structuring and Organization of Music Collections , 2008, Adaptive Multimedia Retrieval.

[17]  Fabio Vignoli,et al.  A Music Retrieval System Based on User Driven Similarity and Its Evaluation , 2005, ISMIR.

[18]  Daniel Gärtner,et al.  Adaptive User Modeling for Content-Based Music Retrieval , 2008, Adaptive Multimedia Retrieval.

[19]  Tim Pohle,et al.  Dynamic Playlist Generation Based on Skipping Behavior , 2005, ISMIR.

[20]  Berry Eggen,et al.  Realization and User Evaluation of an Automatic Playlist Generator , 2003, ISMIR.