Statistical Tools for Chemical Kinetics

Document Type

Presentation Abstract

Presentation Date

10-14-2004

Abstract

Modelling of chemical kinetics often leads to complex dynamical systems whose parameters are to be identified from experimental data. The accuracy of the model predictions should be estimated by statistical methods that take into account the noise in the data as well as possible limitations in modelling. While the models typically are strongly nonlinear, classical statistics is restricted to linear theory and may thus lead to misleading results. New MCMC (Markov chain Monte Carlo) methods allow a proper reliability analysis even for nonlinear models. Here we present applications of this Bayesian methodology to parameter estimation and optimal design of experiments. Examples are given from chemical kinetics as well as from biological modelling of algal mass occurrences in lake systems.

Additional Details

Thursday, 14 October 2004
4:10 p.m. in Jour 304

This document is currently not available here.

Share

COinS