|
Activity Number:
|
428
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 5, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #305776 |
|
Title:
|
Inference for a Difference in Correlated Counts on an Interval
|
|
Author(s):
|
Justin W. Davis*+ and David H. Annis
|
|
Companies:
|
University of Missouri and SportsQuant
|
|
Address:
|
Office of Research, Columbia, MO, 65212,
|
|
Keywords:
|
Poisson Distribution ; Skellam Distribution ; Gene Sequencing
|
|
Abstract:
|
In this talk we explore ways of testing for a difference in counts of some event occurring within an interval, where the counts are observed under two different conditions in paired manner and possibly with covariates. Examples of this include the number of reads of a DNA sequence occurring on an interval within a person treated under two different conditions, or a sporting event where the interval is time and the counts are points scored. This builds on existing work on Skellam's distribution for the difference of correlated Poisson random variables. We explore the performance of a likelihood ratio test under various sample sizes and the behavior of an approximation of the test.
|
