Document Detail

Subspace-based support vector machines for pattern classification.
MedLine Citation:
PMID:  19592215     Owner:  NLM     Status:  MEDLINE    
In this paper, we discuss subspace-based support vector machines (SS-SVMs), in which an input vector is classified into the class with the maximum similarity. Namely, for each class we define the weighted similarity measure using the vectors called dictionaries that represent the class, and optimize the weights so that the margin between classes is maximized. Because the similarity measure is defined for each class, for a data sample the similarity measure to which the data sample belongs needs to be the largest among all the similarity measures. Introducing slack variables, we define these constraints either by equality constraints or inequality constraints. As a result we obtain subspace-based least squares SVMs (SSLS-SVMs) and subspace-based linear programming SVMs (SSLP-SVMs). To speedup training of SSLS-SVMs, which are similar to LS-SVMs by all-at-once formulation, we also propose SSLS-SVMs by one-against-all formulation, which optimize each similarity measure separately. Using two-class problems, we clarify the difference of SSLS-SVMs and SSLP-SVMs and evaluate the effectiveness of the proposed methods over the conventional methods with equal weights and with weights equal to eigenvalues.
Takuya Kitamura; Syogo Takeuchi; Shigeo Abe; Kazuhiro Fukui
Related Documents :
15603165 - How to measure information carried by a modulated vocal signature?
18779085 - Decision manifolds--a supervised learning algorithm based on self-organization.
17627045 - A thousand words in a scene.
14555625 - Automatic discovery and classification of bioinformatics web sources.
2753575 - Dispositional correlates of hypnosis: a phenomenological approach.
9825355 - Assessment of the coupling between rtapex and rr interval as an index of temporal dispe...
Publication Detail:
Type:  Journal Article     Date:  2009-07-02
Journal Detail:
Title:  Neural networks : the official journal of the International Neural Network Society     Volume:  22     ISSN:  1879-2782     ISO Abbreviation:  Neural Netw     Publication Date:    2009 Jul-Aug
Date Detail:
Created Date:  2009-08-11     Completed Date:  2009-11-02     Revised Date:  -    
Medline Journal Info:
Nlm Unique ID:  8805018     Medline TA:  Neural Netw     Country:  United States    
Other Details:
Languages:  eng     Pagination:  558-67     Citation Subset:  IM    
Graduate School of Engineering, Kobe University, Kobe, Japan.
Export Citation:
APA/MLA Format     Download EndNote     Download BibTex
MeSH Terms
Databases, Factual
Least-Squares Analysis
Pattern Recognition, Automated*
Principal Component Analysis

From MEDLINE®/PubMed®, a database of the U.S. National Library of Medicine

Previous Document:  Retrospective reports of behavioral inhibition and young adults' current symptoms of social anxiety,...
Next Document:  Synchronization of nonidentical chaotic neural networks with time delays.