|
Activity Number:
|
43
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, July 31, 2016 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
|
Abstract #319652
|
|
|
Title:
|
A Bayesian Decision Theoretic Methodology for Analyzing Gene Expression Data Under Skewed Alternatives
|
|
Author(s):
|
Naveen Bansal*
|
|
Companies:
|
Marquette University
|
|
Keywords:
|
False Discovery Rate ;
Multiple Hypotheses ;
Skew Normal
|
|
Abstract:
|
Analyzing gene expression data involve testing thousands of multiple hypotheses. Any test based on familywise error rate yields conservative results. Thus a false discovery rate (FDR) is used to produce satisfactory tests. The directional false discovery rate (DFDR) is used to test null hypotheses against directional alternatives. It is well known that the directional hypotheses testing with DFDR controlled error rate can be performed by first performing two-tailed tests with FDR controlled error rate and then making directional decision based on the sign of the test statistics. This result requires that the alternatives are symmetric; in particular, it assumes that the left-sided and the right-sided alternatives are equally likely. We will give some examples where this is not the case. And we will discuss a new statistical methodology based on a Bayesian decision theoretic criterion when the alternatives are skewed.
|
Authors who are presenting talks have a * after their name.
Back to the full JSM 2016 program
|
