NIMBLE Software

I am a co-founder and core developer of NIMBLE, a general computational statistics system (R package nimble). NIMBLE stands for Numerical Inference for statistical Models using Bayesian and Likelihood Estimation. See more at NIMBLE’s web site.

At the time of this writing, NIMBLE has been used in around 600 peer-reviewed papers and 80 dissertations, numbers that are sure to grow faster than this page is updated.

NIMBLE is most commonly used for hierarchical Bayesian modeling with Markov chain Monte Carlo (MCMC) algorithms. However it is extensible for both models and algorithms, and it supports other methods such as maximum likelihood algorithms (e.g. Monte Carlo Expectation Maximization), Laplace approximation, and sequential Monte Carlo (aka particle filtering). NIMBLE has been used in fields ranging from analysis of NBA (basketball) shooting patterns to spatial epidemiology to astronomy, and of course ecology.

NIMBLE is under active development. It is free, open-source software, and contributions from others are very much welcome.

NIMBLE has been supported by several grants from the US National Science Foundation.

Key Features

Flexible Model Language: Build custom hierarchical models using BUGS/JAGS-like syntax
Extensible Algorithms: Customize MCMC samplers and other algorithms
Compiled Code: C++ compilation for computational efficiency
R Integration: Seamless integration with the R statistical computing environment

Getting Started

To install NIMBLE:

install.packages("nimble")

For more information: