Activity Number:
|
414
- Model Building and Selection
|
Type:
|
Contributed
|
Date/Time:
|
Tuesday, August 1, 2017 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Section on Statistical Computing
|
Abstract #324542
|
|
Title:
|
A Data-Driven Approach to the Problem of Multiple Testing
|
Author(s):
|
Nasrine Bendjilali* and Boualem Bendjilali and Wei-Min Huang
|
Companies:
|
Rowan University and RVCC and Lehigh University
|
Keywords:
|
multiple testing procedures ;
False Discovery Rate
|
Abstract:
|
We present a new class of data-driven large-scale multiple testing procedures that are adaptive to general dependence structure of the test statistics. The use of clustering techniques will reveal features in the data that will enhance discoveries while maintaining a control of a suitable type I error rate. Preliminary simulation study demonstrates robust performance of the proposed procedure under various dependence structures of test statistics with accurately controlling various type I error rates, including the False Discovery Rate (FDR). The proposed new procedure shows a significant improvement in power relative to its competitors. Theoretical properties and power performance analysis will be presented.
|
Authors who are presenting talks have a * after their name.