The effect of different basis functions on a radial basis function network for time series prediction: A comparative study

Abstract Many applications using radial basis function (RBF) networks for time series prediction utilise only one or two basis functions; the most popular being the Gaussian function. This function may not always be appropriate and the purpose of this paper is to demonstrate the variation of test set error between six recognised basis functions. The tests were carried out on the Mackey–Glass chaotic time series, Box–Jenkins furnace data and flood prediction data sets for the Rivers Amber and Mole, UK. Each RBF network was trained using a two-stage approach, utilising the k-means clustering algorithm for the first stage and singular value decomposition for the second stage. For this type of network configuration the results indicate that the choice of basis function (and, where appropriate, basis width parameter) is data set dependent and evaluating all recognised basis functions suitable for RBF networks is advantageous.

[1]  C. Micchelli.  Interpolation of scattered data: Distance matrices and conditionally positive definite functions , 1986 .

[2]  James L. McClelland,et al.  Parallel distributed processing: explorations in the microstructure of cognition, vol. 1: foundations , 1986 .

[3]  Chng Eng Siong,et al.  Gradient radial basis function networks for nonlinear and nonstationary time series prediction , 1996, IEEE Trans. Neural Networks.

[4]  T. Kohonen.  Self-organized formation of topographically correct feature maps , 1982 .

[5]  Teuvo Kohonen,et al.  Self-organized formation of topologically correct feature maps , 2004, Biological Cybernetics.

[6]  Michio Sugeno,et al.  A fuzzy-logic-based approach to qualitative modeling , 1993, IEEE Trans. Fuzzy Syst..

[7]  Michel Verleysen,et al.  Model Selection with Cross-Validations and Bootstraps - Application to Time Series Prediction with RBFN Models , 2003, ICANN.

[8]  D. Broomhead,et al.  Radial Basis Functions, Multi-Variable Functional Interpolation and Adaptive Networks , 1988 .

[9]  Mahesan Niranjan,et al.  Neural networks and radial basis functions in classifying static speech patterns , 1990 .

[10]  John Moody,et al.  Fast Learning in Networks of Locally-Tuned Processing Units , 1989, Neural Computation.

[11]  Jyh-Shing Roger Jang,et al.  ANFIS: adaptive-network-based fuzzy inference system , 1993, IEEE Trans. Syst. Man Cybern..

[12]  Christopher M. Bishop,et al.  Neural networks for pattern recognition , 1995 .

[13]  Christian W. Dawson,et al.  An artificial neural network approach to rainfall-runoff modelling , 1998 .

[14]  B. G. Batchelor,et al.  Method for location of clusters of patterns to initialise a learning machine , 1969 .

[15]  Kurt Hornik,et al.  Multilayer feedforward networks are universal approximators , 1989, Neural Networks.

[16]  J. Mason,et al.  Algorithms for approximation , 1987 .

[17]  Christian W. Dawson,et al.  Inductive Learning Approaches to Rainfall-Runoff Modelling , 2000, Int. J. Neural Syst..

[18]  George E. P. Box,et al.  Time Series Analysis: Forecasting and Control , 1977 .

[19]  Jooyoung Park,et al.  Universal Approximation Using Radial-Basis-Function Networks , 1991, Neural Computation.

[20]  A. W. Jayawardena,et al.  Use of Radial Basis Function Type Artificial Neural Networks for Runoff Simulation , 1998 .

[21]  R. L. Hardy.  Multiquadric equations of topography and other irregular surfaces , 1971 .

[22]  L. Glass,et al.  Oscillation and chaos in physiological control systems. , 1977, Science.

[23]  M. J. D. Powell,et al.  Radial basis functions for multivariable interpolation: a review , 1987 .

[24]  Dale Schuurmans,et al.  Automatic basis selection techniques for RBF networks , 2003, Neural Networks.