Using evolutionary multiobjective techniques for imbalanced classification data

Abstract

The aim of this paper is to study the use of Evolutionary Multiobjective Techniques to improve the performance of Neural Networks (NN). In particular, we will focus on classification problems where classes are imbalanced. We propose an evolutionary multiobjective approach where the accuracy rate of all the classes is optimized at the same time. Thus, all classes will be treated equally independently of their presence in the training data set. The chromosome of the evolutionary algorithm encodes only the weights of the training patterns missclassified by the NN. Results show that the multiobjective approach is able to consider all classes at the same time, disregarding to some extent their abundance in the training set or other biases that restrain some of the classes of being learned properly. © 2010 Springer-Verlag Berlin Heidelberg.