Title :
Bayesian model parameter estimation of systems subject to random input and output measurement error
Author :
Perttunen, Cary D.
Author_Institution :
Dept. of Electr. Eng., Louisville Univ., KY, USA
Abstract :
A method, based on Bayesian statistics, for estimating the parameters of a general multiple-input single-output model subject to random input and output measurement error of any distributional form is proposed. First, the method for Bayesian parameter estimation is given. An estimate and a probability region for the estimate of the parameters is found. Then, the necessary extensions to allow for a measurement error in the input are formulated. An application of this to the general linear model with measurement error in both the input and output is given. Finally, the parameters of the junction depletion capacitance model of a diode are estimated, and probability regions for these parameters are found.<>
Keywords :
Bayes methods; measurement errors; parameter estimation; semiconductor device models; semiconductor diodes; statistical analysis; system theory; Bayesian statistics; diode; junction depletion capacitance model; measurement error; multiple-input single-output model; parameter estimation; Bayes procedures; Measurement errors; Parameter estimation; Semiconductor device modeling; Semiconductor diodes; Statistics;
Conference_Titel :
Systems Engineering, 1989., IEEE International Conference on
Conference_Location :
Fairborn, OH, USA
DOI :
10.1109/ICSYSE.1989.48660
