|
Activity Number:
|
111
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 30, 2007 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistical Computing
|
| Abstract - #310363 |
|
Title:
|
Resampling-Based Multiple Testing Procedure
|
|
Author(s):
|
Nasrine Bendjilali*+ and Wei-Min Huang
|
|
Companies:
|
Lehigh University and Lehigh University
|
|
Address:
|
23626 Magic Mountain PKWY, Valencia, CA, 91355,
|
|
Keywords:
|
false discovery rate ; p-value ; multiple hypotheses testing
|
|
Abstract:
|
This article introduces a method for multiple hypotheses testing which is based on the jump size between the potential "smallest null" p-value and the potential "largest alternative" p-value with the structure of resampling methods. The size of the jump is used to determine the stopping rule of the procedure and the resampling incorporates the covariance structure of the data without the need for any distributional assumptions. The proposed procedure seems to have fairly good control of FDR (the false discovery rate). Preliminary study also shows promising gain in power.
|
