Which is the best multiclass SVM method? An empirical study (2005)
| Venue: | Proceedings of the Sixth International Workshop on Multiple Classifier Systems |
| Citations: | 19 - 0 self |
BibTeX
@INPROCEEDINGS{Duan05whichis,
author = {Kai-bo Duan and S. Sathiya Keerthi},
title = {Which is the best multiclass SVM method? An empirical study},
booktitle = {Proceedings of the Sixth International Workshop on Multiple Classifier Systems},
year = {2005},
pages = {278--285}
}
Bookmark
 |
 |
 |
 |
 |
 |
OpenURL
Abstract
Abstract. Multiclass SVMs are usually implemented by combining several two-class SVMs. The one-versus-all method using winner-takes-all strategy and the one-versus-one method implemented by max-wins voting are popularly used for this purpose. In this paper we give empirical evidence to show that these methods are inferior to another one-versusone method: one that uses Platt’s posterior probabilities together with the pairwise coupling idea of Hastie and Tibshirani. The evidence is particularly strong when the training dataset is sparse. 1
Citations
| 6694 | Statistical Learning Theory - Vapnik - 1998 |
| 933 | A training algorithm for optimal margin classifiers - Boser, Guyon, et al. - 1992 |
| 503 | Probabilistic outputs for support vector machines and comparison to regularized likelihood methods - Platt - 1999 |
| 369 | A comparison of methods for multiclass support vector - Hsu |
| 210 | Classification by pairwise coupling - Hastie, Tibshirani - 1998 |
| 193 | Large Margin DAGs for Multiclass Classification - Platt, Cristianini, et al. - 2000 |
| 13 | Probabilistic discriminative kernel classifiers for multi-class problems - Roth - 2001 |
| 12 | R.C.: Probability estimates for multi-class classification by pairwise coupling - Wu, Lin, et al. - 2004 |
| 4 | R.C.: A note on Platts probabilistic outputs for support vector machines - Lin, Lin, et al. - 2007 |
| 2 | G.: Solving multiclass problem via error-correcting output code - Dietterich, Bakiri - 1995 |
| 1 | A.: In defence of one-versus-all classificaiton - Rifkin, Klautau |
