Back To Index Previous Article Next Article Full Text

Statistica Sinica 28 (2018), 899-919

COMPUTER EXPERIMENTS: PREDICTION ACCURACY,
SAMPLE SIZE AND MODEL COMPLEXITY REVISITED
Ofir Harari 1, Derek Bingham 1 , Angela Dean 2 and Dave Higdon 3
1 Simon Fraser University, 2 The Ohio State University
and 3 Biocomplexity Institute of Virginia Tech

Abstract: We revisit the problem of determining the sample size for a Gaussian process emulator and provide a data analytic tool for exact sample size calculations that goes beyond the 𝓃 = 10𝒹 rule of thumb and is based on an IMSPE-related criterion. This allows us to tie sample size and prediction accuracy to the anticipated roughness of the simulated data, and to propose an experimental process for computer experiments, with extension to a robust scheme.

Key words and phrases: Computer experiments, Gaussian processes, sample size calculation.

Back To Index Previous Article Next Article Full Text