|
Activity Number:
|
538
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 2, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #310126 |
|
Title:
|
Mixing Negative Binomial Distribution with Completely Monotonic Functions and Their Applications in Modeling Over-Dispersion Data
|
|
Author(s):
|
Xiaodong Wang*+ and Hanxiang Peng and Ji Zhang
|
|
Companies:
|
sanofi-aventis and The University of Mississippi and sanofi-aventis
|
|
Address:
|
200 Crossing Blvd, Biostatistics, Bridgewater, NJ, 08807,
|
|
Keywords:
|
complete monotonicity ; exchangeability ; negative binomial distribution
|
|
Abstract:
|
In many teratogenic and clinical studies, response binary data are often correlated within each litter or patient. Many statistical methods were developed to analyze the correlated data over 20 years. In this paper, we introduce a new class of parametric parsimonious distributions obtained from mixing negative binomial distribution with completely monotonic functions. By allowing the parameters to depend on covariates, we give a regression procedure that can be used to model the correlated data. We discuss maximum likelihood estimation and give asymptotic normality. The proposed procedure is applied to real data sets. Comparison is made with Poisson distribution model, generalized estimating equation model, and Williams' logistic linear model.
|
