Quantifying simulator discrepancy in discrete-time dynamical simulators

When making predictions with complex simulators it can be important to quantify the various sources of uncertainty. Errors in the structural specification of the simulator, for example due to missing processes or incorrect mathematical specification, can be a major source of uncertainty, but are of...

Full description

Bibliographic Details
Main Authors: Wilkinson, Richard D., Vrettas, Michael, Cornford, Dan, Oakley, Jeremy E.
Format: Article
Published: American Statistical Association 2011
Online Access:https://eprints.nottingham.ac.uk/1524/
_version_ 1848790623184748544
author Wilkinson, Richard D.
Vrettas, Michael
Cornford, Dan
Oakley, Jeremy E.
author_facet Wilkinson, Richard D.
Vrettas, Michael
Cornford, Dan
Oakley, Jeremy E.
author_sort Wilkinson, Richard D.
building Nottingham Research Data Repository
collection Online Access
description When making predictions with complex simulators it can be important to quantify the various sources of uncertainty. Errors in the structural specification of the simulator, for example due to missing processes or incorrect mathematical specification, can be a major source of uncertainty, but are often ignored. We introduce a methodology for inferring the discrepancy between the simulator and the system in discrete-time dynamical simulators. We assume a structural form for the discrepancy function, and show how to infer the maximum likelihood parameter estimates using a particle filter embedded within a Monte Carlo expectation maximization (MCEM) algorithm. We illustrate the method on a conceptual rainfall runoff simulator (logSPM) used to model the Abercrombie catchment in Australia. We assess the simulator and discrepancy model on the basis of their predictive performance using proper scoring rules.
first_indexed 2025-11-14T18:15:33Z
format Article
id nottingham-1524
institution University of Nottingham Malaysia Campus
institution_category Local University
last_indexed 2025-11-14T18:15:33Z
publishDate 2011
publisher American Statistical Association
recordtype eprints
repository_type Digital Repository
spelling nottingham-15242020-05-04T20:24:39Z https://eprints.nottingham.ac.uk/1524/ Quantifying simulator discrepancy in discrete-time dynamical simulators Wilkinson, Richard D. Vrettas, Michael Cornford, Dan Oakley, Jeremy E. When making predictions with complex simulators it can be important to quantify the various sources of uncertainty. Errors in the structural specification of the simulator, for example due to missing processes or incorrect mathematical specification, can be a major source of uncertainty, but are often ignored. We introduce a methodology for inferring the discrepancy between the simulator and the system in discrete-time dynamical simulators. We assume a structural form for the discrepancy function, and show how to infer the maximum likelihood parameter estimates using a particle filter embedded within a Monte Carlo expectation maximization (MCEM) algorithm. We illustrate the method on a conceptual rainfall runoff simulator (logSPM) used to model the Abercrombie catchment in Australia. We assess the simulator and discrepancy model on the basis of their predictive performance using proper scoring rules. American Statistical Association 2011 Article PeerReviewed Wilkinson, Richard D., Vrettas, Michael, Cornford, Dan and Oakley, Jeremy E. (2011) Quantifying simulator discrepancy in discrete-time dynamical simulators. Journal of Agricultural, Biological and Environmental Statistics . ISSN 1085-7117 (Submitted) http://www.amstat.org/publications/jabes.cfm
spellingShingle Wilkinson, Richard D.
Vrettas, Michael
Cornford, Dan
Oakley, Jeremy E.
Quantifying simulator discrepancy in discrete-time dynamical simulators
title Quantifying simulator discrepancy in discrete-time dynamical simulators
title_full Quantifying simulator discrepancy in discrete-time dynamical simulators
title_fullStr Quantifying simulator discrepancy in discrete-time dynamical simulators
title_full_unstemmed Quantifying simulator discrepancy in discrete-time dynamical simulators
title_short Quantifying simulator discrepancy in discrete-time dynamical simulators
title_sort quantifying simulator discrepancy in discrete-time dynamical simulators
url https://eprints.nottingham.ac.uk/1524/
https://eprints.nottingham.ac.uk/1524/