|
Activity Number:
|
10
|
|
Type:
|
Invited
|
|
Date/Time:
|
Sunday, August 2, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
IMS
|
| Abstract - #303192 |
|
Title:
|
The Tau-Path Test: A Generalization of Kendall's Tau Statistic to Test Subpopulation Association
|
|
Author(s):
|
Li Yu*+ and Joseph S. Verducci
|
|
Companies:
|
The Ohio State University and The Ohio State University
|
|
Address:
|
1958 Neil Avenue, Columbus , OH, 43210-1945,
|
|
Keywords:
|
subpopulation monotone association ; Kendall's tau decomposition ; copula mixture
|
|
Abstract:
|
The "tau-path" test method is designed to test strong association in an unknown subpopulation. The method is based on a decomposition of Kendall's tau coefficient of monotone association that sequentially identifies the strongest associations among increasing subsets of observations. The test statistic is derived from the generalized semiparametric distance model, for the sampling distribution of the empirical copula. A computationally efficient procedure, is used to calculate the test statistic. Simulation studies with various copula mixture models show that the proposed test is more powerful than the Mann-Kendall test against alternatives where strong association is limited to a subpopulation. Studies also show promising results in classifying observations into two groups for the two component copula mixtures or order the observations in terms of the population correlation coefficient.
|
