|
Activity Number:
|
368
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 4, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Quality and Productivity
|
| Abstract - #303307 |
|
Title:
|
Supersaturated Designs: Are Our Results Significant?
|
|
Author(s):
|
David Edwards*+ and Robert W. Mee
|
|
Companies:
|
Virginia Commonwealth University and The University of Tennessee
|
|
Address:
|
1001 W. Main Street, Richmond, VA, 23284,
|
|
Keywords:
|
all-subsets regression ; effect sparsity ; forward selection ; global model test ; randomization test
|
|
Abstract:
|
Supersaturated designs (SSDs) are designs that examine more than n-1 factors in n runs. Although relatively easy to find literature involving the construction of SSDs, there is much less available regarding their use and analysis. Whether using forward selection or all-subsets regression, it is common to select models from SSDs that explain a large percentage of the total variation. Naive p-values can persuade the user that included factors are indeed active. We propose the use of a global model randomization test in conjunction with all-subsets to more appropriately select candidate models of interest. An approximation to this test when the number of factors is too large for many repetitions of all-subsets to be feasible is also suggested. Finally, we propose a randomization test for reducing the number of terms in candidate models with small global p-values.
|
