|
Activity Number:
|
67
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 6, 2006 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
Section on Quality and Productivity
|
| Abstract - #307103 |
|
Title:
|
Sequential Analysis on Misspecified Distributions
|
|
Author(s):
|
Theresa Utlaut*+ and Kevin Anderson
|
|
Companies:
|
Intel Corporation and Intel Corporation
|
|
Address:
|
29430 Dutch Canyon Road, Scappoose, OR, 97056,
|
|
Keywords:
|
sequential ; SPRT
|
|
Abstract:
|
Sequential analysis is ideal in environments where resources are limited or delayed decisions result in penalties. The sequential probability ratio test (SPRT) has been shown by Wald and others to guarantee the minimum average sample number to make a decision. As with many statistical tests, the SPRT is based on an underlying assumed distribution. The presentation will investigate the impact of the errors and average sample number on the SPRT when the distributional assumptions have been both slightly and grossly violated. It will demonstrate the use of misspecified SPRT's in a semiconductor manufacturing application, and recommend some potential approaches for successful decision-making in the face of misspecification.
|
