Estudio comparativo entre máquinas de soporte vectorial multiclase, redes neuronales artificiales y sistema de inferencia neuro-difuso auto organizado para problemas de clasificación

Abstract

In this paper we provide an explanation of the structure and how works a self-organized neuro-fuzzy inference system (SONFIS) in detail, we used it over three classification problems (Fisher iris, Breast Cancer and Human Activities) to then compare its results to well-known universal classifiers like artificial neural networks (ANN) and multiclass support vector machines (SVM), a brief description of each one is shown. The results show that SONFIS has a similar/better performance than ANN and SVM with the plus of generating a rule base that helps to understand the inner structure of the problem.