Publication of Holger Fröhlich

Holger Fröhlich and Andreas Zell

Feature Subset Selection for Support Vector Machines by Incremental Regularized Risk Minimization

Abstract

In this paper we present a novel feature selection algorithm for SVMs which works by decreasing the regularized risk in an iterative manner by using a combination of a backward elimination procedure together with an exchange algorithm. It is applicable to linear as well as to nonlinear problems. We test this new algorithm on toy and real life data sets and show its good performance in comparision to state-of-the-art feature selection methods.

Download

[pdf]

BibTeX

@conference{Froe04,
	AUTHOR="H. Fröhlich and A. Zell",
	TITLE="{Feature Subset Selection for Support Vector Machines by Incremental Regularized Risk Minimization}",
	ORGANIZATION="IEEE Int. Joint Conf. on Neural Networks (IJCNN)",
	VOLUME 3,
	PAGES="2041 - 2046",
	YEAR=2004
}