Improved support vectors for classification through preserving neighborhood geometric structure constraint

Xuchu Wang; Yanmin Niu

doi:10.1117/1.3610982

1 August 2011 Improved support vectors for classification through preserving neighborhood geometric structure constraint

Xuchu Wang, Yanmin Niu

Author Affiliations +

Optical Engineering, Vol. 50, Issue 8, 087202 (August 2011). https://doi.org/10.1117/1.3610982

Abstract

The support vector machine (SVM) has become a popular classifier in pattern recognition, computer vision, and other fields. Traditional SVM may result in a nonrobust solution for classifying complex data because its separating hyperplane only reflects the marginal distance information of isolated support vectors, while discarding some useful class structural information. In this paper a new support vector classifier with neighborhood preserving constraint is proposed to enhance the support vectors by preserving the local geometric structure on the manifold of within-class samples. This structure can be represented as a weighted graph matrix and regulated by adding a preprocessing transform in standard SVM. Experimental results validate its effectiveness with comparison to related methods on several synthetic and real-world data sets and show its competence, especially for classifying high dimensional data in a small sample size case.

©(2011) Society of Photo-Optical Instrumentation Engineers (SPIE)

Citation Download Citation

Xuchu Wang and Yanmin Niu "Improved support vectors for classification through preserving neighborhood geometric structure constraint," Optical Engineering 50(8), 087202 (1 August 2011). https://doi.org/10.1117/1.3610982

Published: 1 August 2011

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available