An Investigation of feature models for music genre classification using the support vector classifier

Anders Meng, John Shawe-Taylor

AbstractThe purpose of this work is two-fold:
1. To investigate the multivariate Gaus-
sian model and multivariate autoregressive
model for modelling short time features
(e.g. mel frequency cepstral coefficients)
over a segment of audio.
2. Investigate how these two models can be
formulated in a kernel framework.
KeywordsProduct probability kernel, Convolution kernel, multivariate autoregressive model, SVM, Support Vector machine
TypeMisc [Poster]
Journal/Book/ConferenceInternational Conference on Music Information Retrieval
Year2005
AddressLondon
NotePoster presented at the conference
Electronic version(s)[pdf]
BibTeX data [bibtex]
IMM Group(s)Intelligent Signal Processing


Back  ::  IMM Publications