Inferring efficient hierarchical taxonomies for MIR tasks: Application to musical instruments

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

A number of approaches for automatic audio classification are based on hierarchical taxonomies since it is acknowledged that improved performance can be thereby obtained. In this paper, we propose a new strategy to automatically acquire hierarchical taxonomies, using machine learning methods, which are expected to maximize the performance of subsequent classification. It is shown that the optimal hierarchical taxonomy of musical instruments (in the sense of inter-class distances) does not follow the traditional and more intuitive instrument classification into instrument families.

Original languageEnglish
Title of host publicationISMIR 2005 - 6th International Conference on Music Information Retrieval
PublisherQueen Mary, University of London
Pages324-328
Number of pages5
ISBN (Print)9780955117909
Publication statusPublished - 1 Jan 2005
Event6th International Conference on Music Information Retrieval, ISMIR 2005 - London, United Kingdom
Duration: 11 Sept 200515 Sept 2005

Publication series

NameISMIR 2005 - 6th International Conference on Music Information Retrieval

Conference

Conference6th International Conference on Music Information Retrieval, ISMIR 2005
Country/TerritoryUnited Kingdom
CityLondon
Period11/09/0515/09/05

Keywords

  • Clustering
  • Hierarchical taxonomy
  • Musical instrument
  • Probabilistic distance

Fingerprint

Dive into the research topics of 'Inferring efficient hierarchical taxonomies for MIR tasks: Application to musical instruments'. Together they form a unique fingerprint.

Cite this