R-Forge Logo

Welcome to Hyper-g Priors for GAM selection project!

This R-package implements objective Bayesian model selection with penalised splines, using a so-called hyper-g prior for the parameters in the generalised additive model.

Description: The R-package hypergsplines implements objective Bayesian variable selection in generalized additive models (Normal, Binomial and Poisson). As objective parameters prior a (generalized) hyper-g prior is used. As objective model prior so-called multiplicity correction priors are used. More details on the methodology can be found in the technical report (2011).
Download: Currently the R-package hypergsplines is not available from CRAN, because it is still under continuing development. The current versions are available from R-Forge (this site). If you are running a recent R version, you can obtain a binary snapshot of the development version as
You can also manually download the binary snapshot or the source tarball. You can have a look at the recent changes to the package here. The project summary page can be found here.
Documentation: The R-package features a vignette which introduces the most important functions in a logistic additive regression example. After installing the package, you can open the vignette from within R with the command
vignette("examples", package="hypergsplines").
You can also have a look at the ISBA 2012 poster presenting the methodology.
Developers: Daniel Sabanes Bove, Division of Biostatistics, Institute of Social and Preventive Medicine, University of Zurich, Switzerland