Sparse multinomial kernel discriminant analysis (sMKDA)

Harrison, R.F. and Pasupa, K. (2009) Sparse multinomial kernel discriminant analysis (sMKDA). Pattern Recognition, 42 (9). pp. 1795-1802. ISSN 0031-3203

Full text available as:

Preview

Text harrisonrf5.pdf Download (656Kb)

Abstract

Dimensionality reduction via canonical variate analysis (CVA) is important for pattern recognition and has been extended variously to permit more flexibility, e.g. by "kernelizing" the formulation. This can lead to over-fitting, usually ameliorated by regularization. Here, a method for sparse, multinomial kernel discriminant analysis (sMKDA) is proposed, using a sparse basis to control complexity. It is based on the connection between CVA and least-squares, and uses forward selection via orthogonal least-squares to approximate a basis, generalizing a similar approach for binomial problems. Classification can be performed directly via minimum Mahalanobis distance in the canonical variates. sMKDA achieves state-of-the-art performance in terms of accuracy and sparseness on 11 benchmark datasets.

Item Type:	Article
Copyright, Publisher and Additional Information:	© 2009 Elsevier. This is an author produced version of a paper subsequently published in Pattern Recognition. Uploaded in accordance with the publisher's self-archiving policy
Keywords:	Linear discriminant analysis; Kernel discriminant analysis; Multi-class; Multinomial; Least-squares; Optimal scaling; Sparsity control
Academic Units:	The University of Sheffield > Faculty of Engineering (Sheffield) > Department of Automatic Control and Systems Engineering (Sheffield)
Depositing User:	Miss Anthea Tucker
Date Deposited:	23 Jul 2009 10:53
Last Modified:	08 Feb 2013 16:58
Published Version:	http://dx.doi.org/10.1016/j.patcog.2009.01.025
Status:	Published
Publisher:	Elsevier
Refereed:	Yes
Identification Number:	10.1016/j.patcog.2009.01.025
URI:	http://eprints.whiterose.ac.uk/id/eprint/9012

Actions (login required)

View Item