On a Class of Multivariate Negative Binomial Distributions.
Abstract
The negative binomial distribution arises in probability as the distribution of the waiting time to achieve a specified number of successes in a sequence of Bernoulli trials. In addition, it has been widely used in statistics as a model for a variety of data involving counts. Multivariate analogues of the negative binomial distribution are of interest as joint distributions of waiting times in Bernoulli trials and for modelling data involving pairs or m-tuples of possibly dependent counts. This paper considers m-variate distributions concentrated on Z(+) superscript m is identical to (Z(+)) superscript m where Z(+) = (0,1,2,...) and having negative binomial univariate marginal distributions. Such distributions will be called multivariate/m-variate negative binomial distributions. It is not hard to see that distributions of this form do not, for any fixed m, exhaust the class of all m-variate negative binomial distributions. Several types of counterexample can be constructed.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1980
- Accession Number
- ADA095445
Entities
People
- R. K. Milne
Organizations
- University of North Carolina at Chapel Hill