eql {JointGLM} | R Documentation |
Compute the Extended Quasi-Likelihood and the Extended Quasi-Deviance for a Joint Model.
eql(mod.mean, mod.disp, df.adj = TRUE)
mod.mean, mod.disp |
Two fitted GLMs/GAMs correponding to the mean and the dispersion components respectively. |
df.adj |
Logical. If TRUE - the default, adjustment on degree of freedom is applied for the EQD. |
The Extended Quasi-Likelihood (EQL) generalizes the Quasi-Likelihood as it behaves as a likelihood function on both mean and dispersion parameters. The Extended Quasi-Deviance is defined as
sum_i d_i + sum_i phi_i log( 2π phi_i V(y_i) )
where d_i is the deviance contribution of the i-th observation, phi_i is the fitted values of dispersion parameter for the i-th observation, V is the variance function of the mean GLM and y_i is the i-th observation.
This function returns the Extended Quasi-Likelihood and the Extended Quasi-Deviance for a joint model.
Mathieu Ribatet and Bertrand Iooss
McCullagh P. and Nelder J. A. (1987). Generalized Linear Models Second Edition. London: Chapman and Hall. ISBN 0-412-31760-5.