R-Package BMS
Bayesian Model Averaging in R
The BMS Package
BMS is a free R package for performing Bayesian Model Averaging within the open-source software R: The tutorials provide some screenshots.
This page refers to the current version BMS 0.3.5.
Install BMS
Note that BMS has temporarily disappeared from CRAN due to a technicality confronting a lack of time.
We aim to get BMS back to CRAN by Aug 2022.
In the meantime you can install BMS by one of the following ways:
- Install BMS manually by first downloading BMS_0.3.5.tar.gz , and then installing it manually in your local R/Rstudio (e.g. menu -> tools -> install packages...)
- Install BMS directly from this site by executing
install.packages("BMS", repos="http://bms.zeugner.eu/getBMS/tarballs")
- Install BMS from github using the package devtools, by the following command:
devtools::install_github("zeugner/BMS")
Soon, BMS shoube be able to be easily installed from inside R, and be available via the CRAN repository.
Introduction to BMS
If you know already something about BMA but not about R, then you might consider the 'FLS' tutorial.
If you know neither BMA nor R well, then the introductory manual "Bayesian Model Averaging with BMS" could be helpful.
Moreover, you might consider the other tutorials, as well as the FAQ and the online help files.
About BMS
The R-package BMS is free Bayesian Model Averaging software that is designed according to three objectives:
- Scope: implements a wide range of (customizable) priors and many ways to obtain posterior results.
- Usability: thoroughly tested; documentation and tutorials aim to be brief and understandable.
- Modularity: coded exclusively in R, structured in modules, without dependence on other contributed packages. This feature aims to provide building blocks for custom projects.
BMS (short for 'Bayesian Model Sampling') was created by Martin Feldkircher and Stefan Zeugner. In case of questions, please feel free to contact us.
The current version of BMS is released under the BSD-3 license. This basically means that you are free to use, distribute and modify the package.
Still, it would be kind if you cite us.