A New Classifier for Multi-Class Problems Based on Negative Selection Algorithm

Abstract

A novel classification approach based on the principle of self and non-self discrimination by T cells in biological immune system is proposed in the paper. In order to classify the multi-class problems, the concepts of self and non-self in negative selection algorithm were redefined. The classifier consisted of different kinds of detector sets obtained from the algorithm. Each detector set is applicable for classification in a way that one class is distinguished from the others. The classifier is tested in the experiments on UCI dataset. The results show that our algorithm is useful for classification problems and comparable with other traditional classification methods.