A new scatter-based multi-class support vector machine

Robert Jenssen; Marius Kloft; Sören Sonnenburg; Alexander Zien; Klaus Robert Müller

doi:10.1109/MLSP.2011.6064625

A new scatter-based multi-class support vector machine

Robert Jenssen, Marius Kloft, Sören Sonnenburg, Alexander Zien, Klaus Robert Müller

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

1 Citation (Scopus)

Abstract

We provide a novel interpretation of the dual of support vector machines (SVMs) in terms of scatter with respect to class prototypes and their mean. As a key contribution, we extend this framework to multiple classes, providing a new joint Scatter SVM algorithm, at the level of its binary counterpart in the number of optimization variables. We identify the associated primal problem and develop a fast chunking-based optimizer. Promising results are reported, also compared to the state-of-the-art, at lower computational complexity.

Original language	English
Title of host publication	2011 IEEE International Workshop on Machine Learning for Signal Processing - Proceedings of MLSP 2011
DOIs	https://doi.org/10.1109/MLSP.2011.6064625
Publication status	Published - 2011
Event	21st IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2011 - Beijing, China Duration: 2011 Sept 18 → 2011 Sept 21

Publication series

Name	IEEE International Workshop on Machine Learning for Signal Processing

Other

Other	21st IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2011
Country/Territory	China
City	Beijing
Period	11/9/18 → 11/9/21

Keywords

multi-class
scatter
μ-SVM

ASJC Scopus subject areas

Human-Computer Interaction
Signal Processing

Access to Document

10.1109/MLSP.2011.6064625

Cite this

A new scatter-based multi-class support vector machine. / Jenssen, Robert; Kloft, Marius; Sonnenburg, Sören et al.
2011 IEEE International Workshop on Machine Learning for Signal Processing - Proceedings of MLSP 2011. 2011. 6064625 (IEEE International Workshop on Machine Learning for Signal Processing).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Jenssen, R, Kloft, M, Sonnenburg, S, Zien, A & Müller, KR 2011, A new scatter-based multi-class support vector machine. in 2011 IEEE International Workshop on Machine Learning for Signal Processing - Proceedings of MLSP 2011., 6064625, IEEE International Workshop on Machine Learning for Signal Processing, 21st IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2011, Beijing, China, 11/9/18. https://doi.org/10.1109/MLSP.2011.6064625

@inproceedings{9822bdd217d348b59c9444a306c06572,

title = "A new scatter-based multi-class support vector machine",

abstract = "We provide a novel interpretation of the dual of support vector machines (SVMs) in terms of scatter with respect to class prototypes and their mean. As a key contribution, we extend this framework to multiple classes, providing a new joint Scatter SVM algorithm, at the level of its binary counterpart in the number of optimization variables. We identify the associated primal problem and develop a fast chunking-based optimizer. Promising results are reported, also compared to the state-of-the-art, at lower computational complexity.",

keywords = "multi-class, scatter, μ-SVM",

author = "Robert Jenssen and Marius Kloft and S{\"o}ren Sonnenburg and Alexander Zien and M{\"u}ller, {Klaus Robert}",

year = "2011",

doi = "10.1109/MLSP.2011.6064625",

language = "English",

isbn = "9781457716232",

series = "IEEE International Workshop on Machine Learning for Signal Processing",

booktitle = "2011 IEEE International Workshop on Machine Learning for Signal Processing - Proceedings of MLSP 2011",

note = "21st IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2011 ; Conference date: 18-09-2011 Through 21-09-2011",

}

TY - GEN

T1 - A new scatter-based multi-class support vector machine

AU - Jenssen, Robert

AU - Kloft, Marius

AU - Sonnenburg, Sören

AU - Zien, Alexander

AU - Müller, Klaus Robert

PY - 2011

Y1 - 2011

N2 - We provide a novel interpretation of the dual of support vector machines (SVMs) in terms of scatter with respect to class prototypes and their mean. As a key contribution, we extend this framework to multiple classes, providing a new joint Scatter SVM algorithm, at the level of its binary counterpart in the number of optimization variables. We identify the associated primal problem and develop a fast chunking-based optimizer. Promising results are reported, also compared to the state-of-the-art, at lower computational complexity.

AB - We provide a novel interpretation of the dual of support vector machines (SVMs) in terms of scatter with respect to class prototypes and their mean. As a key contribution, we extend this framework to multiple classes, providing a new joint Scatter SVM algorithm, at the level of its binary counterpart in the number of optimization variables. We identify the associated primal problem and develop a fast chunking-based optimizer. Promising results are reported, also compared to the state-of-the-art, at lower computational complexity.

KW - multi-class

KW - scatter

KW - μ-SVM

UR - http://www.scopus.com/inward/record.url?scp=82455212637&partnerID=8YFLogxK

U2 - 10.1109/MLSP.2011.6064625

DO - 10.1109/MLSP.2011.6064625

M3 - Conference contribution

AN - SCOPUS:82455212637

SN - 9781457716232

T3 - IEEE International Workshop on Machine Learning for Signal Processing

BT - 2011 IEEE International Workshop on Machine Learning for Signal Processing - Proceedings of MLSP 2011

T2 - 21st IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2011

Y2 - 18 September 2011 through 21 September 2011

ER -

A new scatter-based multi-class support vector machine

Abstract

Publication series

Other

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this