An Investigation of Feature Models for Music Genre Classification using the Support Vector Classifier

Anders Meng, John Shawe-Taylor

    Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

    43 Downloads (Pure)

    Abstract

    In music genre classification the decision time is typically of the order of several seconds however most automatic music genre classification systems focus on short time features derived from 10-50ms. This work investigates two models, the multivariate gaussian model and the multivariate autoregressive model for modelling short time features. Furthermore, it was investigated how these models can be integrated over a segment of short time features into a kernel such that a support vector machine can be applied. Two kernels with this property were considered, the convolution kernel and product probability kernel. In order to examine the different methods an 11 genre music setup was utilized. In this setup the Mel Frequency Cepstral Coefficients (MFCC) were used as short time features. The accuracy of the best performing model on this data set was 44% as compared to a human performance of 52% on the same data set.
    Original languageEnglish
    Title of host publicationInternational Conference on Music Information Retrieval
    Publication date2005
    Pages604-609
    Publication statusPublished - 2005
    Event6th International Conference on Music Information Retrieval - London, United Kingdom
    Duration: 11 Sep 200515 Sep 2005
    Conference number: 6
    http://ismir2005.ismir.net/

    Conference

    Conference6th International Conference on Music Information Retrieval
    Number6
    CountryUnited Kingdom
    CityLondon
    Period11/09/200515/09/2005
    Internet address

    Keywords

    • Support Vector Machine
    • Product Probability Kernel
    • Convolution Kernel
    • Music Genre
    • Feature Integration

    Fingerprint Dive into the research topics of 'An Investigation of Feature Models for Music Genre Classification using the Support Vector Classifier'. Together they form a unique fingerprint.

    Cite this