Assessments metrics for multi-class imbalance learning: A preliminary study

Abstract

In this paper we study some of the most common global measures employed to measure the classifier performance on the multi-class imbalanced problems. The aim of this work consists of showing the relationship between global classifier performance (measure by global measures) and partial classifier performance, i.e., to determine if the results of global metrics match with the improved classifier performance over the minority classes. We have used five strategies to deal with the class imbalance problem over five real multi-class datasets on neural networks context. © 2013 Springer-Verlag Berlin Heidelberg.