Generalized linear models with varying dispersion

SUMMARY Generalized linear models are further generalized to include a linear predictor for the dispersion as well as for the mean. It is shown how the convenient structure of generalized linear models can be carried over to this more general setting by considering the mean and dispersion structure separately. Mean and dispersion submodels are formulated for this, the dependent variable for the dispersion submodel being the deviance components of the mean submodel. The fact that both submodels are essentially generalized linear models themselves is used to derive simple expressions for the likelihood equations and for asymptotic tests. Estimation algorithms are proposed which have good convergence properties. The results apply mainly to the normal, inverse Gaussian and gamma distributions but can be extended to discrete distributions by using quasi-likelihoods. The methods developed are applied to a well-known data set.

[1]  P. McCullagh.  Quasi-Likelihood Functions , 1983 .

[2]  S. Weisberg,et al.  Diagnostics for heteroscedasticity in regression , 1983 .

[3]  Richard Morton,et al.  A generalized linear model with nested strata of extra-Poisson variation , 1987 .

[4]  James M. Ortega,et al.  Iterative solution of nonlinear equations in several variables , 2014, Computer science and applied mathematics.

[5]  Ole E. Barndorff-Nielsen,et al.  Exponential Models with Affine Dual Foliations , 1983 .

[6]  Daryl Pregibon,et al.  Review: P. McCullagh, J. A. Nelder, Generalized Linear Models , 1984 .

[7]  W. Douglas Stirling.  Heteroscedastic Models and an Application to Block Designs , 1985 .

[8]  O. Barndorff-Nielsen,et al.  Reproductive Exponential Families , 1983 .

[9]  P. McCullagh,et al.  Generalized Linear Models , 1984 .

[10]  D. A. Williams,et al.  Extra‐Binomial Variation in Logistic Linear Models , 1982 .

[11]  R. Park.  Estimation with Heteroscedastic Error Terms , 1966 .

[12]  R. W. Wedderburn.  Quasi-likelihood functions, generalized linear models, and the Gauss-Newton method , 1974 .

[13]  J. A. Nelder,et al.  Quasi-Likelihood and GLIM , 1985 .

[14]  A. Harvey.  Estimating Regression Models with Multiplicative Heteroscedasticity , 1976 .

[15]  M. West,et al.  Dynamic Generalized Linear Models and Bayesian Forecasting , 1985 .

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

[17]  Herbert C. Rutemiller,et al.  Estimation in a Heteroscedastic Regression Model , 1968 .

[18]  M. Aitkin.  Modelling variance heterogeneity in normal regression using GLIM , 1987 .

[19]  J. Nelder,et al.  An extended quasi-likelihood function , 1987 .

[20]  S. Weisberg,et al.  Applied Linear Regression (2nd ed.). , 1986 .

[21]  T. Breurch,et al.  A simple test for heteroscedasticity and random coefficient variation (econometrica vol 47 , 1979 .

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

[23]  J. L. Jensen,et al.  Saddlepoint Formulas for Reproductive Exponential Models , 1985 .

[24]  B. Jørgensen.  Exponential Dispersion Models , 1987 .