Analysis Methods for Computer Experiments: How to Assess and What Counts?
Authors:
- Hao ChenHCHao Chen
- Jason L. LoeppkyJLJason L. Loeppky
- Jerome SacksJSJerome Sacks
- William J. WelchWWWilliam J. Welch
Open Access
- 1 February 2016
- journal article
- research article
- Published by Institute of Mathematical Statistics in Statistical Science
- Vol. 31 (1) , 40-60
Abstract
Statistical methods based on a regression model plus a zero-mean Gaussian process (GP) have been widely used for predicting the output of a deterministic computer code. There are many suggestions in the literature for how to choose the regression component and how to model the correlation structure of the GP. This article argues that comprehensive, evidence-based assessment strategies are needed when comparing such modeling options. Otherwise, one is easily misled. Applying the strategies to several computer codes shows that a regression model more complex than a constant mean either has little impact on prediction accuracy or is an impediment. The choice of correlation function has modest effect, but there is little to separate two common choices, the power exponential and the Matérn, if the latter is optimized with respect to its smoothness. The applications presented here also provide no evidence that a composite of GPs provides practical improvement in prediction accuracy. A limited comparison of Bayesian and empirical Bayes methods is similarly inconclusive. In contrast, we find that the effect of experimental design is surprisingly large, even for designs of the same type with the same theoretical properties.Keywords
This publication has 37 references indexed in Scilit:
- Quantile-Based Optimization of Noisy Computer Experiments With Tunable PrecisionTechnometrics, 2012
- Bayesian Treed Gaussian Process Models With an Application to Computer ModelingJournal of the American Statistical Association, 2008
- Response to James M. LucasTechnometrics, 1996
- Effect of outdoor airborne particulate matter on daily death counts.Environmental Health Perspectives, 1995
- Measurement Error and Spatial Variability Effects on Characterization of Volatile Organics in the SubsurfaceEnvironmental Science & Technology, 1995
- Parameter space exploration of an ocean general circulation model using an isopycnal mixing parameterizationJournal of Marine Research, 1994
- Arctic sea ice variability: Model sensitivities and a multidecadal simulationJournal of Geophysical Research: Oceans, 1994
- Bayesian Design and Analysis of Computer Experiments: Use of Derivatives in Surface PredictionTechnometrics, 1993
- Bayesian Prediction of Deterministic Functions, with Applications to the Design and Analysis of Computer ExperimentsJournal of the American Statistical Association, 1991
- Design and Analysis of Computer ExperimentsStatistical Science, 1989