Dataset selection for training one-class support vector machines

Li, Y; Maguire, L

Abstract

This paper proposes an efficient training strategy for one-class support vector machines. The strategy exploits the feature of a trained one-class SVM which uses points only residing on the exterior region of data distribution as support vectors. Thus the proposed training set reduction method selects the so-called extreme points which sit on the boundary of data distribution, through local geometry and k-nearest neighbors. Experimental results on synthetic and real-world data demonstrate that the proposed training strategy can reduce training set of support vector machines considerably while the obtained model maintains generalization capability to the level of a model trained on the full training set.

Citation

Li, Y., & Maguire, L. (2009, December). Dataset selection for training one-class support vector machines. Presented at International Conference on Computational Intelligence and Software Engineering, Wuhan, China

• Presentation Conference Type: Other
• Conference Name: International Conference on Computational Intelligence and Software Engineering
• Conference Location: Wuhan, China
• Start Date: Dec 11, 2009
• End Date: Dec 13, 2009
• Publication Date: Dec 11, 2009
• Deposit Date: Jul 27, 2015
• Publisher: Institute of Electrical and Electronics Engineers
• Publisher URL: http://dx.doi.org/10.1109/CISE.2009.5366620
• Related Public URLs: http://ieeexplore.ieee.org/xpl/mostRecentIssue.jsp?punumber=5362500
• Additional Information: Event Type : Conference