Quasi-likelihood functions, generalized linear models, and the Gauss—Newton method
To define a likelihood we have to specify the form of distribution of the observations, but to define a quasi-likelihood function we need only specify a relation between the mea...