JPL Technical Report Server

Multiclass reduced-set support vector machines

Show simple item record Tang, Benyang Mazzoni, Dominic 2008-05-02T16:35:46Z 2008-05-02T16:35:46Z 2006-07-25
dc.identifier.citation International Conference of Machine Learning, Pittsburgh, Pennsylvania, July 25, 2006 en_US
dc.identifier.clearanceno 06-2158
dc.description.abstract There are well-established methods for reducing the number of support vectors in a trained binary support vector machine, often with minimal impact on accuracy. We show how reduced-set methods can be applied to multiclass SVMs made up of several binary SVMs, with significantly better results than reducing each binary SVM independently. Our approach is based on Burges' approach that constructs each reduced-set vector as the pre-image of a vector in kernel space, but we extend this by recomputing the SVM weights and bias optimally using the original SVM objective function. This leads to greater accuracy for a binary reduced-set SVM, and also allows vectors to be "shared" between multiple binary SVMs for greater multiclass accuracy with fewer reduced-set vectors. We also propose computing pre-images using differential evolution, which we have found to be more robust than gradient descent alone. We show experimental results on a variety of problems and find that this new approach is consistently better than previous multiclass reduced-set methods, sometimes with a dramatic difference. en_US
dc.description.sponsorship NASA/JPL en_US
dc.language.iso en_US en_US
dc.publisher Pasadena, CA : Jet Propulsion Laboratory, National Aeronautics and Space Administration, 2006. en_US
dc.subject support vector machines en_US
dc.subject reduced set methods en_US
dc.subject Kernel pre-image en_US
dc.subject multiclass en_US
dc.subject differential evolution en_US
dc.title Multiclass reduced-set support vector machines en_US
dc.type Preprint en_US

Files in this item

This item appears in the following Collection(s)

Show simple item record



My Account