varlasso package

This repo includes an R package for fitting Bayesian Cector Autoregressive (VAR) models with optional shrinkage priors on coefficient estimates

Documentation for the package is here: https://atsa-es.github.io/varlasso/

This work extends the MARSS package (which provides similar models in a maximum likelihood setting). MARSS: https://cran.r-project.org/web/packages/MARSS/index.html

And more material can be found on our Applied Time Series website, https://atsa-es.github.io/

Installation

To install the package, use

# install.packages("remotes")
remotes::install_github("atsa-es/varlasso")

Citations

A paper describing applications of these methods to VAR problems is in PeerJ,

Ward, E.J., K.N. Marshall, and M.D. Scheuerell. 2022. Regularizing priors for Bayesian VAR applications to large ecological datasets, PeerJ

And the citation for this repository is