A fast approach to improve classification performance of ECOC classification systems

Abstract

Error correcting output coding is a well known technique to decompose a multi-class classification problem into a group of two-class problems which can be faced by using a combination of binary classifiers. Each of them is trained on a different dichotomy of the classes. The way the set of classes is mapped on this set of dichotomies may essentially influence the obtained performance. In this paper we present a new tool, the k-NN lookup table to optimize this mapping in a fast way and a fast procedure to change the dichotomies in a proper way. Experiments on artificial and public data sets show that the proposed procedure may significantly improve the ECOC performance in multi-class problems. © 2008 Springer Berlin Heidelberg.