Activity Number:
|
385
|
Type:
|
Contributed
|
Date/Time:
|
Thursday, August 15, 2002 : 8:30 AM to 10:20 AM
|
Sponsor:
|
Section on Physical & Engineering Sciences*
|
Abstract - #300691 |
Title:
|
A Best Subsets Approach for Interpreting Out-of-Control Signals from Hotelling's T2 Chart
|
Author(s):
|
Michael Conerly*+ and Kevin Busby+
|
Affiliation(s):
|
University of Alabama and Corning, Inc.
|
Address:
|
P.O. Box 870226, Tuscaloosa, Alabama, 35487-0226, USA , Cornng, New York, 14831, USA
|
Keywords:
|
Multivariate Quality Control ; Hotelling’s T2 Chart
|
Abstract:
|
Hotelling's T2 chart is used in quality control to monitor the process of several correlated variables. Determining which quality characteristic(s) contribute to an out-of-control signal is a difficult task, with no technique clearly preferred. Historically, practitioners have used univariate charts on each of the variables being monitored; this practice, however, ignores the covariance structure among the variables. Other methods have been proposed to incorporate the correlation structure into the decomposition routine, but these are computer intensive and often difficult to interpret. This study suggests a new method, analogous to best subsets regression, for detecting problematic variables when the T2 chart signals. The proposed method allows for one unique decomposition of the T2 statistic and is easily computed.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2002 program |