Research Repository

An inexact penalty method for the semiparametric Support Vector Machine classifier

Lai, Daniel and Mani, N and Palaniswami, Marimuthu (2006) An inexact penalty method for the semiparametric Support Vector Machine classifier. In: 2006 International Joint Conference on Neural Networks, Sheraton Vancouver Wall Centre Hotel, Vancouver, BC, Canada, July 16-21, 2006|. IEEE, Piscataway, New Jersey, pp. 333-338.

Full text for this resource is not available from the Research Repository.

Abstract

The support vector machine (SVM) classifier has been a popular classification tool used for a variety of pattern recognition tasks. In this study, we compare the performance of a semiparametric SVM classifier derived using an inexact penalty method on the original SVM formulation. This semiparametric form can be easily solved using a sequential decomposition method. We compare the accuracy of the semiparametric SVM against the standard SVM classifier trained using the SMO algorithm. The results indicate that in some cases the semiparametric SVM can give better generalization results than a standard SVM. We also demonstrate several cases where our iterative algorithm solves the SVM problem faster than the SMO.

Item Type: Book Section
ISBN: 0780394909, 9780780394896
Uncontrolled Keywords: ResPubID19070, pattern classification, classification tool, inexact penalty method, pattern recognition tasks, semiparametric support vector machine, classifier, sequential decomposition method, Gaussian processes, H infinity control, iterative algorithms, kernel, machine learning, neural networks, pattern recognition, support vector machine classification, support vector machines, systems engineering and theory
Subjects: FOR Classification > 0801 Artificial Intelligence and Image Processing
Faculty/School/Research Centre/Department > Centre for Ageing, Rehabilitation, Exercise & Sport Science (CARES)
Depositing User: VUIR
Date Deposited: 16 Dec 2013 00:52
Last Modified: 16 Mar 2015 05:02
URI: http://vuir.vu.edu.au/id/eprint/5909
DOI: 10.1109/IJCNN.2006.246700
ePrint Statistics: View download statistics for this item

Repository staff only

View Item View Item

Search Google Scholar