|
Activity Number:
|
607
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 6, 2009 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #304626 |
|
Title:
|
Estimating the number of true null hypotheses in multiple hypothesis testing
|
|
Author(s):
|
Yi-Ting Hwang and Shu-Yu Liao and Hsun-Chih Kuo*+
|
|
Companies:
|
National Chengchi University and National Taipei University
|
|
Address:
|
NO.64,Sec.2,ZhiNan Rd.,Wenshan District, Taipei, 11605, Taiwan
|
|
Keywords:
|
False discovery rate ; familywise error rate ; multiple hypothesis testing ; number of null hypotheses
|
|
Abstract:
|
The overall Type I error can be inflated if many hypotheses are compared simultaneously. False discovery rate (FDR) is recognized as a useful error rates to measure Type I error. Many controlling FDR procedures are proposed and have the ability to control the FDR. Nevertheless, these controlling procedures become too conservative when some hypotheses are from the alternative. Benjamini and Hochberg (2000) proposed an adaptive FDR-controlling procedure that adapts the information of the number of true null hypotheses ($m_{0}$). Benjamini and Hochberg (2000) suggested a graphical approach to construct an estimator of $m_0$, which is shown to overestimate $m_0$. Following a similar construction, this paper proposes new estimators of $m_{0}$. Monte Carlo simulations are used to evaluate accuracy and precision of new estimators and the feasibility of these new adaptive procedures.
|
