A new scatter-based multi-class support vector machine

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

Research output: Chapter in Book/Report/Conference proceedingConference 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 languageEnglish
Title of host publicationIEEE International Workshop on Machine Learning for Signal Processing
DOIs
Publication statusPublished - 2011 Dec 5
Externally publishedYes
Event21st IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2011 - Beijing, China
Duration: 2011 Sep 182011 Sep 21

Other

Other21st IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2011
CountryChina
CityBeijing
Period11/9/1811/9/21

    Fingerprint

Keywords

  • μ-SVM
  • multi-class
  • scatter

ASJC Scopus subject areas

  • Human-Computer Interaction
  • Signal Processing

Cite this

Jenssen, R., Kloft, M., Sonnenburg, S., Zien, A., & Muller, K. (2011). A new scatter-based multi-class support vector machine. In IEEE International Workshop on Machine Learning for Signal Processing [6064625] https://doi.org/10.1109/MLSP.2011.6064625