Overdispersed Generalized Linear Models

Abstract Generalized linear models have become a standard class of models for data analysts. However, in some applications, heterogeneity in samples is too great to be explained by the simple variance function implicit in such models. Utilizing a two parameter exponential family which is overdispersed relative to a specified one-parameter exponential family enables the creation of classes of overdispersed generalized linear models (OGLMs) which are analytically attractive. We propose fitting such models within a Bayesian framework employing noninformative priors in order to let the data drive the inference. Hence, our analysis approximates likelihood-based inference but with possibly more reliable estimates of variability for small sample sizes. Bayesian calculations are carried out using a Metropolis-within-Gibbs sampling algorithm. An illustrative example using a data set involving damage incidents to cargo ships is presented. Details of the data analysis are provided including comparison with the standard generalized linear models analysis. Several diagnostic tools reveal the improved performance of the OGLM.

[1]  Moshe Shared,et al.  On Mixtures from Exponential Families , 1980 .

[2]  Purushottam W. Laud,et al.  On Bayesian Analysis of Generalized Linear Models Using Jeffreys's Prior , 1991 .

[3]  D. Cox,et al.  Parameter Orthogonality and Approximate Conditional Inference , 1987 .

[4]  Alan E. Gelfand,et al.  Bayesian statistics without tears: A sampling-resampling perspective , 1992 .

[5]  N. Breslow,et al.  Approximate inference in generalized linear mixed models , 1993 .

[6]  B. Efron.  THE GEOMETRY OF EXPONENTIAL FAMILIES , 1978 .

[7]  J. F. C. Kingman,et al.  Information and Exponential Families in Statistical Theory , 1980 .

[8]  H. Jeffreys,et al.  The Theory of Probability , 1896 .

[9]  A. Gelfand,et al.  Bayesian Model Choice: Asymptotics and Exact Calculations , 1994 .

[10]  B. Lindsay.  Exponential family mixture models (with least-squares estimators) , 1986 .

[11]  R. W. Wedderburn,et al.  On the existence and uniqueness of the maximum likelihood estimates for certain generalized linear models , 1976 .

[12]  B. Efron.  Double Exponential Families and Their Use in Generalized Linear Regression , 1986 .

[13]  Hong Chang,et al.  Model Determination Using Predictive Distributions with Implementation via Sampling-Based Methods , 1992 .

[14]  Adrian F. M. Smith,et al.  Bayesian Inference for Generalized Linear and Proportional Hazards Models Via Gibbs Sampling , 1993 .

[15]  Alan E. Gelfand,et al.  A Note on Overdispersed Exponential Families , 1990 .

[16]  L. Pettit,et al.  Measuring the effect of observations on Bayes factors , 1990 .

[17]  L. M. M.-T..  Theory of Probability , 1929, Nature.

[18]  P. McCullagh,et al.  Generalized Linear Models , 1972, Predictive Analytics.

[19]  Lisa M. Ganio,et al.  Diagnostics for Overdispersion , 1992 .