Statistics on reliability of neural network estimates

Author

Bhide, V.M. ; Piovoso, M.J. ; Kosanovich, K.A.

Author_Institution

E.I. du Pont de Nemours & Co, Wilmington, DE, USA

Volume

3

fYear

1995

fDate

21-23 Jun 1995

Firstpage

1877

Abstract

It has been demonstrated that artificial neural networks can be used to infer estimates of variables infrequently measured in many applications. These estimates have been used in closed-loop control applications, however, the reliability of the estimates have not been used to improve the controller´s performance. This work focuses on the generation of an appropriate statistic for the reliability of the estimate, the confidence interval. This statistic is calculated from an empirical sampling distribution obtained using the bootstrap technique. A demonstration of the bootstrap method in the context of an ANN to estimate of a distillation process bottoms´ composition is provided. A discussion on the use of the bootstrap estimate and its confidence interval to the practical problem of controller tuning and process performance follows

Keywords

chemical industry; computer bootstrapping; distillation; neural nets; parameter estimation; process control; reliability; statistical analysis; bootstrap technique; closed-loop control; confidence interval; controller tuning; distillation process; neural network estimates; reliability; statistical analysis; Artificial neural networks; Chemical engineering; Feedback control; Input variables; Neural networks; Predictive models; Process control; Sampling methods; Statistical distributions; Statistics;

fLanguage

English

Publisher

ieee

Conference_Titel

American Control Conference, Proceedings of the 1995

Conference_Location

Seattle, WA

Print_ISBN

0-7803-2445-5

Type

conf

DOI

10.1109/ACC.1995.531212

Filename

531212