Abstract
In this paper, we propose a boundary method to speed up constructing the optimal hyperplane of support vector machines. The boundary, called key vector set, is an approximate small superset of support vector set which is extracted by Parzen window density estimation in the feature space. Experimental results on Checkboard data set show that the proposed method is more efficient than some conventional methods and requires much less memory.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
REFERENCES
V. Vapnik (1995), The Nature of Statistical Learning Theory. Springer-Verlag, New York.
S.F. Zheng (2004), Unsupervised clustering based reduced support vector machines. Computer Engineering and Application, 14, pp. 74–6 (in Chinese).
M.H. Yang and N. Ahuja (2000 June), A geometric approach to train support vector machines. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 2000 (CVPR 2000), Hilton Head Island, pp. 430–7.
N. Kwak and C.H. Choi (2002 December), Input feature selection by mutual information based on Parzen window. IEEE Transaction on Pattern Analysis and Machine Intelligence, 24, 12, pp. 1667–71.
Z.Q. Bian and X.G. Zhang (2000), Pattern Recognition, 2nd ed., Tsinghua University Press, Beijing.
URL (1999), Check Board, http://www.cs.wisc.edu/~olvi/data/check1.txt.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2006 Springer
About this paper
Cite this paper
Wang, Y., Zhou, C., Huang, Y., Liang, Y., Yang, X. (2006). A BOUNDARY METHOD TO SPEED UP TRAINING SUPPORT VECTOR MACHINES. In: LIU, G., TAN, V., HAN, X. (eds) Computational Methods. Springer, Dordrecht. https://doi.org/10.1007/978-1-4020-3953-9_31
Download citation
DOI: https://doi.org/10.1007/978-1-4020-3953-9_31
Publisher Name: Springer, Dordrecht
Print ISBN: 978-1-4020-3952-2
Online ISBN: 978-1-4020-3953-9
eBook Packages: EngineeringEngineering (R0)