A new scatter-based multi-class support vector machine

Jenssen, R; Kloft, M; Sonnenburg, S; Zien, A; Müller, K-R

doi:10.1109/MLSP.2011.6064625

Local TagsRelease HistoryDetailsSummary

A new scatter-based multi-class support vector machine

Jenssen, R., Kloft, M., Sonnenburg, S., Zien, A., & Müller, K.-R. (2011). A new scatter-based multi-class support vector machine. In T. Tan (Ed.), 2011 IEEE International Workshop on Machine Learning for Signal Processing. Piscataway, NJ, USA: IEEE.

Item is Released

show all

Basic

hide

Item Permalink: https://hdl.handle.net/21.11116/0000-0010-5657-C Version Permalink: https://hdl.handle.net/21.11116/0000-0010-8407-1

Genre: Conference Paper

Files

show Files

Locators

show

Creators

hide

Creators:
Jenssen, R, Author
Kloft, M, Author
Sonnenburg, S¹, Author
Zien, A, Author
Müller, K-R, Author

Affiliations:
1Rätsch Group, Friedrich Miescher Laboratory, Max Planck Society, ou_3378052

Content

hide

Free keywords: -

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.

Details

hide

Language(s):

Dates: Date issued: 2011-10

Publication Status: Issued

Pages: -

Publishing info: -

Table of Contents: -

Rev. Type: -

Identifiers: DOI: 10.1109/MLSP.2011.6064625

Degree: -

Event

hide

Title: IEEE International Workshop on Machine Learning for Signal Processing (MLSP 2011)

Place of Event: Beijing, China

Start-/End Date: 2011-09-18 - 2011-09-21

Legal Case

show

Project information

show

Source 1

hide

Title: 2011 IEEE International Workshop on Machine Learning for Signal Processing

Source Genre: Proceedings

Creator(s):
Tan, T, Editor

Affiliations:
-

Publ. Info: Piscataway, NJ, USA : IEEE

Pages: 6 Volume / Issue: - Sequence Number: - Start / End Page: - Identifier: ISBN: 978-1-4577-1621-8