Learning sparse SVM for feature selection on very high dimensional datasets

Tan, M; Wang, L; Tsang, IW

Learning sparse SVM for feature selection on very high dimensional datasets

Tan, M Wang, L Tsang, IW

Permalink

Publication Type:: Conference Proceeding
Citation:: ICML 2010 - Proceedings, 27th International Conference on Machine Learning, 2010, pp. 1047 - 1054
Issue Date:: 2010-09-17

Closed Access

	Filename	Description	Size
	2013004291OK.pdf		450.24 kB	Adobe PDF	View/Open

Copyright Clearance Process

Recently Added
In Progress
Closed Access

This item is closed access and not available.

Full metadata record

Field	Value	Language
dc.contributor.author	Tan, M	en_US
dc.contributor.author	Wang, L	en_US
dc.contributor.author	Tsang, IW https://orcid.org/0000-0001-8095-4637	en_US
dc.date.issued	2010-09-17	en_US
dc.identifier.citation	ICML 2010 - Proceedings, 27th International Conference on Machine Learning, 2010, pp. 1047 - 1054	en_US
dc.identifier.isbn	9781605589077	en_US
dc.identifier.uri	http://hdl.handle.net/10453/29424
dc.description.abstract	A sparse representation of Support Vector Machines (SVMs) with respect to input features is desirable for many applications. In this paper, by introducing a 0-1 control variable to each input feature, Zo-norm Sparse SVM (SSVM) is converted to a mixed integer programming (MIP) problem. Rather than directly solving this MIP, we propose an efficient cutting plane algorithm combining with multiple kernel learning to solve its convex relaxation. A global convergence proof for our method is also presented. Comprehensive experimental results on one synthetic and l0 real world datasets show that our proposed method can obtain better or competitive performance compared with existing SVM-based feature selection methods in term of sparsity and generalization performance. Moreover, our proposed method can effectively handle large-scale and extremely high dimensional problems. Copyright 2010 by the author(s)/owner(s).	en_US
dc.relation.ispartof	ICML 2010 - Proceedings, 27th International Conference on Machine Learning	en_US
dc.title	Learning sparse SVM for feature selection on very high dimensional datasets	en_US
dc.type	Conference Proceeding
utslib.for	0801 Artificial Intelligence and Image Processing	en_US
dc.location.activity	Haifa, Israel	en_US
pubs.embargo.period	Not known	en_US
pubs.organisational-group	/University of Technology Sydney
pubs.organisational-group	/University of Technology Sydney/Faculty of Engineering and Information Technology
pubs.organisational-group	/University of Technology Sydney/Strength - CAI - Centre for Artificial Intelligence
utslib.copyright.status	closed_access
pubs.publication-status	Published	en_US

Abstract:

A sparse representation of Support Vector Machines (SVMs) with respect to input features is desirable for many applications. In this paper, by introducing a 0-1 control variable to each input feature, Zo-norm Sparse SVM (SSVM) is converted to a mixed integer programming (MIP) problem. Rather than directly solving this MIP, we propose an efficient cutting plane algorithm combining with multiple kernel learning to solve its convex relaxation. A global convergence proof for our method is also presented. Comprehensive experimental results on one synthetic and l0 real world datasets show that our proposed method can obtain better or competitive performance compared with existing SVM-based feature selection methods in term of sparsity and generalization performance. Moreover, our proposed method can effectively handle large-scale and extremely high dimensional problems. Copyright 2010 by the author(s)/owner(s).

Please use this identifier to cite or link to this item:

http://hdl.handle.net/10453/29424