Which Is the Best Multiclass SVM Method? An Empirical Study

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.

• Type: Conference
• Year: 2005
• Where: MCS
• Authors: Kaibo Duan, S. Sathiya Keerthi
• Keywords: MCS 2005, Multiclass SVMs, Pattern Recognition, Platt’s Posterior Probabilities, Two-class Svms