SOME REMARKS ON STRUCTURE SELECTION FOR NONLINEAR MODELS

Abstract
This note is concerned with structure selection of nonlinear models. By structure selection it is meant the choice of the model basis prior to parameter estimation. It is argued that the effect of the sampling period and the noise on the data may in some cases preclude adequate structure selection. Other issues which are briefly discussed include the selection of the sampling period and the effectiveness of information criteria in indicating the best size of a nonlinear model.

This publication has 0 references indexed in Scilit: