|
QICID: 24363
Title: Using Orthogonal Arrays in the Sensitivity Analysis of Computer Models
Copyright: ASQ; American Statistical Association
Author: Morris, Max D.; Moore, Leslie M.; McKay, Michael D.
Organization: Iowa State University; Los Alamos National Laboratory
Subject: Computer programs, Design of experiments (DOE), Variance components, First order design, Uncertainty, Sampling plans, Orthogonal array (OA);
Series: Technometrics, Vol. 50, No. 2, May 2008, pp. 205-215
This ARTICLE is available for FREE
to ASQ members
with the appropriate membership type and/or magazine subscriptions. If you are signed-in you should be able to download articles you are entitled to receive. If you have questions about your membership please contact Customer Care at 800-248-1946 or help@asq.org
Abstract: [This abstract is based on the authors' abstract.]
Permuted column sampling plans that are popular in sensitivity analysis of computer models are examined. Theses plans support estimation of first-order sensitivity coefficients, but estimates are biased when random column permutation is used to construct the sampling arrays. Deterministic column permutations may be used to eliminate this estimation bias. It is shown that any permuted column sampling plan that eliminates estimation bias can be characterized by an orthogonal array of strength 2. Approximate standard errors of the first-order sensitivity indices are given for this sampling plan, which is demonstrated with two examples.
Number of pages: 11
Price for ASQ Members: $5.00
Price for List/Forum/Division: $10.00
All electronic articles are sent as PDFs via e-mail. To view the documents, you will need Adobe
Reader (free download).
Orders placed during business hours are usually filled within one business day.
If you have questions please e-mail our Customer Care center at help@asq.org.
Browse QIC articles chronologically
previous next
|
