Please use this identifier to cite or link to this item: https://doi.org/10.1109/ICIEA.2009.5138876
Title: Combining a global SVM and local nearest-neighbor classifiers driven by local discriminative boundaries
Authors: Xiong, W.
Ong, S.H. 
Le, T.T. 
Lim, J.H.
Liu, J.
Foong, K. 
Keywords: Adaptive metric
Boundary driven
Combination
Local
Nearest neighbors
Support vector machines
Issue Date: 2009
Source: Xiong, W.,Ong, S.H.,Le, T.T.,Lim, J.H.,Liu, J.,Foong, K. (2009). Combining a global SVM and local nearest-neighbor classifiers driven by local discriminative boundaries. 2009 4th IEEE Conference on Industrial Electronics and Applications, ICIEA 2009 : 3597-3600. ScholarBank@NUS Repository. https://doi.org/10.1109/ICIEA.2009.5138876
Abstract: Nonlinear support vector machines (SVMs) rely on the kernel trick and tradeoff parameters to build nonlinear models to classify complex problems and balance misclassification and generalization. The inconvenience in determining the kernel and the parameters has motivated the use of local nearest neighbor (NN) classifiers in lieu of global classifiers. This substitution ignores the advantage of SVM in global error minimization. On the other hand, the NN rule assumes that class conditional probabilities are locally constant. Such an assumption does not hold near class boundaries and in any high dimensional space due to the curse of dimensionality. We propose a hybrid classification method combining the global SVM and local NN classifiers. Local classifiers occur only when the global SVM is likely to fail. Furthermore, local NN classifiers adopt an adaptive metric driven by local SVM discriminative boundaries. Improved performance has been demonstrated compared to partially similar. © 2009 IEEE.
Source Title: 2009 4th IEEE Conference on Industrial Electronics and Applications, ICIEA 2009
URI: http://scholarbank.nus.edu.sg/handle/10635/69635
ISBN: 9781424428007
DOI: 10.1109/ICIEA.2009.5138876
Appears in Collections:Staff Publications

Show full item record
Files in This Item:
There are no files associated with this item.

SCOPUSTM   
Citations

2
checked on Dec 12, 2017

Page view(s)

33
checked on Dec 9, 2017

Google ScholarTM

Check

Altmetric


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.