|
Activity Number:
|
219
|
|
Type:
|
Other
|
|
Date/Time:
|
Tuesday, August 8, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
The ASA
|
| Abstract - #305238 |
|
Title:
|
The Modeling and Analysis of Data from Computer Experiments
|
|
Author(s):
|
Thomas Santner*+
|
|
Companies:
|
The Ohio State University
|
|
Address:
|
Department of Statistics, Columbus, OH, 43210-1247,
|
|
Keywords:
|
computer experiments ; gasp models
|
|
Abstract:
|
In this second talk we will consider statistical models that one might use to analyze data from computer experiments and that form the basis for sequential design strategies. We discuss the advantages and disadvantages of response surface modeling and then introduce the flexible class of Gaussian Stochastic Process (GASP) models. Because GASP models are popular in practice but unfamiliar to many statisticians we will provide an introduction to these models and give some intuition about the parameters in these models. We will discuss prediction and other inference and indicate how such models can be fit using statistical software. Examples will be used throughout.
|
