A Clustering Based Procedure for Learning the Hidden Unit Parameters in Elliptical Basis Function Networks
Radial basis function (RBF) neural networks can be used for a wide range of applications as they can be regarded as universal approximators and their training is faster than that of multilayer perceptrons. A more general version of these neural networks are referred to as elliptical basis function (EBF) networks. In this paper a robust method for EBF parameter estimation is proposed, based on hyperellipsoidal clustering and on multivariate sign and rank concepts. A simulation study on a classification problem has shown that this method represents a valid learning scheme, particularly in presence of outlying data.
KeywordsRadial Basis Function Radial Basis Function Network Hide Unit Gaussian Basis Function Spatial Median
Unable to display preview. Download preview PDF.
- BISHOP, M.C. (1995): Neural Networks for Pattern Recognition, Clarendon Press, Oxford.Google Scholar
- CERIOLI, A. (2001): Elliptical Clusters and the K-means Algorithm. In: Book of Short Papers of Cladag 2001, Meeting of the Classification and Data Analysis Group of the Italian Statistical Society.Google Scholar
- LAY, S.R. and HWANG, J.N. (1993): Robust construction of Radial Basis Function Networks for classification, IEEE International Conference on Neural Networks, 1859–1864.Google Scholar
- PILLATI, M. and CALO, D.G. (2001): A Robust Clustering Procedure for Centre Location in RBF Networks. In C. Provasi (Ed.): Modelli complessi e metodi computazionali intensivi per la stima e la previsione, Cleup Editrice, Padova.Google Scholar