|
Activity Number:
|
405
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 6, 2008 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Quality and Productivity
|
| Abstract - #301657 |
|
Title:
|
Heterogeneous Variance Models and Their Applications in Parameter Designs
|
|
Author(s):
|
Fassil Nebebe*+
|
|
Companies:
|
Concordia University
|
|
Address:
|
1455 de Maisonneuve Blvd. West, Montreal, QC, H3G 1M8, Canada
|
|
Keywords:
|
Control factors identification ; crossed array designs ; PerMIA ; model mis-specification
|
|
Abstract:
|
We consider in this paper the issues of estimation and variance model building in the analysis of robust parameter designs. The model examined accommodates the possible simplification resulting from the use of a "Performance Measure Independent of Adjustment (PerMIA)" and is applicable in both crossed array and combined array designs. We consider a likelihood based parametric method through the use of a very flexible class of distributions for the error terms. The application of the idea of PerMIA typically involves a smaller set of "control factors" that affect the PerMIA (the remaining design factors are adjustment factors that affect only the mean). Thus a central issue is the identification of the control factors. We address the issue of identifying the control factors in building an equation for the PerMIA, and examine the effects of model misspecification.
|
