First experiments on ensembles of radial basis functions

Abstract

Building an ensemble of classifiers is an useful way to improve the performance with respect to a single classifier. In the case of neural networks the bibliography has centered on the use of Multilayer Feedforward. However, there are other interesting networks like Radial Basis Functions (RBF) that can be used as elements of the ensemble. Furthermore, as pointed out recently the network RBF can also be trained by gradient descent, so all the methods of constructing the ensemble designed for Multilayer Feedforward are also applicable to RBF. In this paper we present the results of using eleven methods to construct an ensemble of RBF networks. We have trained ensembles of a reduced number of networks (3 and 9) to keep the computational cost low. The results show that the best method is in general the Simple Ensemble. © Springer-Verlag 2004.