First Experiments on Ensembles of Radial Basis Functions
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.
KeywordsRadial Basis Function Radial Basis Function Neural Network Ensemble Method Error Reduction Training Pattern
Unable to display preview. Download preview PDF.
- 6.Freund, Y., Schapire, R.: Experiments with a New Boosting Algorithm. In: Proceedings of the Thirteenth International Conference on Machine Learning, pp. 148–156 (1996)Google Scholar
- 8.Auda, G., Kamel, M.: EVOL: Ensembles Voting On-Line. In: Proc. of the World Congress on Computational Intelligence, pp. 1356–1360 (1998)Google Scholar
- 9.Liu, Y., Yao, X.: A Cooperative Ensemble Learning System. In: Proc. of the World Congress on Computational Intelligence, pp. 2202–2207 (1998)Google Scholar
- 10.Jang, M., Cho, S.: Ensemble Learning Using Observational Learning Theory. In: Proceedings of the International Joint Conference. on Neural Networks, vol. 2, pp. 1281–1286 (1999)Google Scholar