The VGAM package for R


Another important note (October 2015): In VGAM 1.0-0 (and also 0.9-9), changes to link functions such as logit() and loge() were made. For example, reciprocals of the first derivatives are now returned compared to previously. Users are urged to check their results. In the help file, please type ?Links for more information.

Important note (Dec 2014): In VGAM 0.9-5, numerous argument names, their order, and function names are changed. This was due to inconsistencies I detected while writing my book, and it was deemed necessary to standardize things. This particularly relates to family functions implementing discrete and continuous distributions. I have tried to summarize all changes in the NEWS file, however, a few changes might have gone undocumented. Users of previous versions of VGAM are cautioned to check their code. My apologies for this inconvenience. I have always tried to make it plain that the software was always subject to change, prior to the release of VGAM 1.0-0, which coincided with the publication of my book Vector Generalized Linear and Additive Models: With an Implementation in R.

The VGAM package for R fits vector generalized linear and additive models (VGLMs/VGAMs), as well as reduced-rank VGLMs (RR-VGLMs) and quadratic RR-VGLMs (QRR-VGLMs), and can be obtained below. It is a general program for maximum likelihood estimation, and centers on the six S functions vglm(), vgam(), rrvglm(), cqo(), cao() and rcim().

VGAM can fit regression models to the following data types: