A robust nonlinear identification algorithm using PRESS statistic and forward regression

Author

Hong, X. ; Sharkey, P.M. ; Warwick, K.

Author_Institution

Dept. of Cybern., Univ. of Reading, UK

Volume

14

Issue

2

fYear

2003

fDate

3/1/2003 12:00:00 AM

Firstpage

454

Lastpage

458

Abstract

This paper introduces a new robust nonlinear identification algorithm using the predicted residual sums of squares (PRESS) statistic and forward regression. The major contribution is to compute the PRESS statistic within a framework of a forward orthogonalization process and hence construct a model with a good generalization property. Based on the properties of the PRESS statistic the proposed algorithm can achieve a fully automated procedure without resort to any other validation data set for iterative model evaluation.

Keywords

function approximation; generalisation (artificial intelligence); identification; radial basis function networks; statistical analysis; time series; PRESS statistic; cross validation; forward regression; function approximation; generalization; nonlinear model; orthogonalization; predicted residual sums of squares; radial basis function network; structure identification; time series; Cost function; Iterative algorithms; Least squares approximation; Neural networks; Parameter estimation; Particle measurements; Prediction algorithms; Predictive models; Robustness; Statistics;

fLanguage

English

Journal_Title

Neural Networks, IEEE Transactions on

Publisher

ieee

ISSN

1045-9227

Type

jour

DOI

10.1109/TNN.2003.809422

Filename

1189645