A novel multi-class SVM classifier based on DDAG

摘要

This paper presents a new method of constructing multi-class SVM classifier, which is based on the structure of Decision Directed Acyclic Graph (DDAG) and using active constraint for each SVM classifier. For k-class problem, it combines k(k-1)/2 two-class SVM classifiers, one for each pair of classes. In order to speed up the training and decision process of the classifier, we make three changes on the standard two-class classifiers, i.e. large margin, 2-norm squared for the error for the soft margin and active constraint. While in the testing phase, we use a rooted binary directed acyclic graph which has k(k-1)/2 internal nodes and k leaves. Computational experiment indicates that this is a simple and fast approach to generate multi-class SVM classifiers.