paciorek_NSF2020.pdf (471.13 kB)

NIMBLE: Programmable Statistical Modeling for Hierarchical/Graphical Models

Download (471.13 kB)
poster
posted on 28.01.2020 by Christopher Paciorek, Perry de Valpine, Daniel Turek

NIMBLE is a system for building and sharing analysis methods for statistical models, especially for hierarchical models and computationally-intensive methods. NIMBLE is built in R but compiles your models and algorithms using C++ for speed. It includes three components:

  1. A system for using models written in the BUGS model language as programmable objects in R.
  2. An initial library of algorithms for models written in BUGS, including basic MCMC, which can be used directly or can be customized from R before being compiled and run.
  3. A language embedded in R for programming algorithms for models, both of which are compiled through C++ code and loaded into R.

NIMBLE can also be used without BUGS models as a way to compile simple R-like code into C++, which is then compiled and loaded into R with an interface function or object.

Categories

Funding

NSF ACI-1550488

History

Licence

Exports

Categories

Licence

Exports