Identification of Multiple‐input Transfer‐function Noise Models: A Regression Approach — Part I: Theory

Evaluates an identification strategy, based on ridge regression, for mutliple‐input models. The corresponding computer algorithm was implemented on a VAX–8800 computer at the Computing Centre at A˚bo Akademi. The evaluation of the ridge‐regression method was carried out by simulations of different transfer‐function noise model structures. The models are essentially the same as those of Edlund, but a far greater number of replications, 1,000, is used in each of the 21 cases tested. Furthermore, uses actually identified and estimated ARIMA models of the residuals in the identification procedure of impulse response weights, unlike Edlund, who used only theoretical noise models in filtering the input and output series. A short discussion of the underlying theory is presented in Part I. The procedures and results of the empirical testing will be published in Part II in Kybernetes, Vol. 22 No. 7, 1993.