Implementation of an Algorithm for Fitting a Class of Generalized Logistic Models

A general class of models is proposed for describing the dependence of binary data on explanatory variables, extending the scope of the standard logistic model. A simple and intuitive algorithm for fitting parameters is presented, based on the properties of generalized linear models; implementation is facilitated through the flexibility of GLIM. Asymptotic variance expressions are derived for the parameters and the fitted values. User defined GLIM macros for fitting the model are included and an example is presented.