ABSTRACT
This chapter provides a brief introduction to generalized linear models (GLiMs) methodology, and illustrates some environmetric applications. It deals with a short review of the concepts and terminology of the general linear model. One new concern that arises when assessing GLiM adequacy is the quality of the link assumption. Logistic regression analysis of incidental tumor data can be extended to include other explanatory factors as covariables, giving it great flexibility. The basic precept of any GLiM is to extend the linear model in two ways: generalize to nonnormal parent distributions such as binomial, Poisson, or gamma; and/or generalize to nonlinear functions that link the unknown means of the parent distribution with the predictor variables. The more flexible techniques of generalized linear modeling may be applied also to count data, with the linear model fit on some positive scale such as log-linear or absolute value.
