Yves Deville 2023-05-22
bever is a R package funded by IRSN/Behrig.
This package is devoted to some Bayesian approaches in Extreme Value (EV) modelling, with focus on Markov Chain Monte-Carlo inference. The package does not implement by itself the Bayesian estimation or the inference of any EV model. These tasks are left to the revdbayes package or to some other MCMC sampler as can be implemented by using JAGS or Stan. The bever package aims at producing some classical results or plots such as Return Level tables or plots. Its main goal is to allow a comparison of the frequentist and Bayesian inference methods when applied to EV models.
The name bever relates to Bayes Extreme-Value. In some modern or ancient European languages, the word relates to the river or/and to one of its famous inhabitant. Hydrology may be one field of application.
This package is still in its early stage of development
-
The use of block maxima with block duration differing from 1 is not well tested.
-
For now it is possible to draw the predictive curve on the same graphics as the RL plot only for GEV models, not for Poisson-GP models. The reason is that some analysis is still required for the transformation of plotting positions in “Renext style” to use them in “block style”.
In an R session use
library(remotes)
install_github("IRSN/bever", dependencies = TRUE)This should install the package and make it ready to use.
Mind that by default this does not build the vignette shipped with the package (long-form documentation). To build the vignette, use instead
install_github("IRSN/bever", dependencies = TRUE, build_vignettes = TRUE)The installation will then take a longer time but the vignette will be accessible from the help of the package (link above the “Help Pages” section).
You can also select a specific branch or a specific commit by using the
suitable syntax for install_github. For instance to install the branch
develop use
install_github("IRSN/bever@develop", dependencies = TRUE)See the remotes package documentation for more details.
You can also clone the repository to install the package. If you do not
have yet a local bever repository, use git clone to clone the
bever repository
git clone https://github.com/IRSN/beverThis will create a bever sub-directory of the current directory,
i.e. the directory from which the git command was issued. Of course this
can work only if you have the authorisation to clone.
Move to the parent directory of your cloned repository and use the following command from a terminal to create a tarball source file
R CMD build beverThis will produce a source tarball bever_x.y.z where x, y and z
stand for the major, minor and patch version numbers. Then you can
install from a command line
R CMD INSTALL bever_x.y.zNote that you must also have all the packages required by bever installed.
You can also use the RStudio IDE to install the package.