Tutorial: Introduction to Generalized Nonlinear Models in R

 
Heather Turner and David Firth, Department of Statistics, University of Warwick, UK.

Abstract

In many practical applications we wish to model the expected value of a response that is non-Gaussian by a nonlinear function of covariates. Generalized nonlinear models provide a framework to do this.

Generalized nonlinear models may be viewed as an extension to generalized linear models — allowing a nonlinear predictor — or an extension to nonlinear least squares — accommodating non-Gaussian data.

In this tutorial we shall provide an overview of generalized nonlinear models and describe how they may be specified, fitted and evaluated using the gnm package. We will also present several examples of how generalized nonlinear models are used in practice.

Outline

  • Introduction to generalized nonlinear models.
  • Overview of the gnm package:
    • specification of nonlinear terms;
    • controlling the fitting procedure;
    • evaluating the fitted model.
  • Several examples from a range of applications, likely to include:
    • models for contingency tables;
    • multiplicative models;
    • models for multinomial data;
    • models for mortality data.
  • Summary.

Audience

People who wish to find out what generalized nonlinear models are and whether such models might be useful in their field of application, and people who want an introduction to the gnm package.


Background Knowledge

It will be assumed that participants have some familiarity with generalized linear models (e.g. logistic regression models, log-linear models, etc) and/or nonlinear least squares models.

Slides

You can download the slides of the course