Skip to main content

gnm: an R Package for Generalized Nonlinear Models

News: One-day course, "Introduction to Generalized Nonlinear Models" to be given by Heather Turner at ETH Zurich, 11 May 2015. Details and registration


gnm is a package for R, which provides facilities for specifying and fitting (by maximum likelihood) a broad class of generalized nonlinear models. These models are like generalized linear models (linear regression, logistic regression, log-linear models, etc.) but may also include one or more nonlinear terms in the predictor function (i.e., the right hand side of the regression equation). Some special cases are models involving multiplicatively structured interactions, such as the UNIDIFF and row-column association models from sociology and the AMMI and GAMMI models from crop science; stereotype models for ordered categorical response; Lee-Carter models for trends in age-specific mortality, and diagonal-reference models for dependence on a square 2-way classification.

The package authors, Heather Turner and David Firth, received the 2007 John M. Chambers Statistical Software Award for their work on gnm, which was supported by the ESRC Professorial Fellowship RES-051-27-0055.

Articles in SCGN and R News (the links are to PDF files) give concise introductions to the package, the latter from a more technical viewpoint.

For a detailed overview of the package, with several examples of gnm applications, see the manual for version 0.9-6: Generalized nonlinear models in R: An overview of the gnm package (pdf file, 4.1MB).

A short course, "Introduction to generalized nonlinear models in social research", was given by the authors at the ESRC Oxford Spring School in 2005 and the course slides are available to download (pdf file, 300K). Note that these slides were prepared for gnm version 0.7-2 and some slides are now out of date.

The gnm software is made available free, under GPL version 2. It comes with ABSOLUTELY NO WARRANTY.

Package contents

The package is available at CRAN sites for installation into R in the standard way.

Version history

0.6-1: first public release (2005-06-28)
For subsequent history see the NEWS file in the inst subdirectory of the package.

Related links: