Connectionist modeling vs. Bayesian procedures for sparse data pharmacokinetic parameter estimation

Author

Shadmehr, Reza ; D´Argenio, David Z.

Author_Institution

Univ. of Southern California, Los Angeles, CA, USA

fYear

1989

fDate

9-12 Nov 1989

Firstpage

2058

Abstract

A connectionist model (adaptive neural network) is developed for estimating the pharmacokinetic properties of a drug from plasma concentrations measured during the course of therapy. The back-propagation algorithm was used to determine the weights in a three-layered network model from simulated sets of kinetic parameters and drug concentrations. The estimation performance of the connectionist model is shown to compare well to that of maximum-likelihood and Bayesian estimators

Keywords

Bayes methods; parameter estimation; physiological models; 3-layered network model; Bayesian estimator; adaptive neural network; back-propagation algorithm; connectionist model; estimation performance; kinetic parameters; maximum-likelihood estimator; plasma concentration; sparse data pharmacokinetic parameter estimation; Adaptive systems; Bayesian methods; Drugs; Kinetic theory; Maximum likelihood estimation; Medical treatment; Neural networks; Plasma measurements; Plasma properties; Plasma simulation;

fLanguage

English

Publisher

ieee

Conference_Titel

Engineering in Medicine and Biology Society, 1989. Images of the Twenty-First Century., Proceedings of the Annual International Conference of the IEEE Engineering in

Conference_Location

Seattle, WA

Type

conf

DOI

10.1109/IEMBS.1989.96593

Filename

96593