Spectrum-based design of sinusoidal RBF neural networks

Author

András, Péter

Author_Institution

Dept. of Psychol., Univ. of Newcastle, Newcastle upon Tyne, UK

Volume

2

fYear

2002

fDate

6/24/1905 12:00:00 AM

Firstpage

1421

Lastpage

1426

Abstract

This paper introduces and describes the spectrum-based design of radial basis function (RBF) neural networks. The RBF networks used in this paper work with damped sinusoidal nonlinear activation functions. The concept of the associated data spectrum is introduced, and it is shown how to apply this spectrum to find the number of hidden neurons and their internal parameters for a neural network solution of the related data processing problem. A time series prediction application is presented. The relation of the proposed method to the support vector machine method and the application of the method to select appropriate basis functions for a problem with given data are discussed

Keywords

forecasting theory; learning automata; network synthesis; radial basis function networks; spectra; time series; associated data spectrum; basis functions selection; damped sinusoidal nonlinear activation functions; data processing; hidden neurons; internal parameters; sinusoidal radial basis function neural networks; spectrum-based design; support vector machine method; time series prediction application; Approximation methods; Artificial neural networks; Data processing; Neural networks; Neurons; Psychology; Radial basis function networks; Support vector machines;

fLanguage

English

Publisher

ieee

Conference_Titel

Neural Networks, 2002. IJCNN '02. Proceedings of the 2002 International Joint Conference on

Conference_Location

Honolulu, HI

ISSN

1098-7576

Print_ISBN

0-7803-7278-6

Type

conf

DOI

10.1109/IJCNN.2002.1007725

Filename

1007725