|
What Tools/Software are Available for Applying Bayesian
Metrology?
|
The benefits of Bayesian methodolgy will not be fully
realized unless software and tools for Bayesian analysis
are readily available to NIST scientists and engineers.
Currently available software includes the following.
-
BUGS (Bayesian inference Using Gibbs Sampling)
is a program that analyzes complex statistical
models using Markov chain Monte Carlo methods.
It is jointly developed by the MRC Biostatistics
Unit, Cambridge, UK and the Imperial College School
of Medicine at St. Mary's, London.
- The bounds on bias (BOB) methodology has been
incorporated into the
Dataplot
software program via the
CONSENSUS MEANS and the
CONSENSUS MEANS PLOT commands. The
BOB methodology was developed by the NIST Statistical
Engineering Division.
- The
Flexible Bayesian Modeling and Markov Chain Sampling
software, provided by Radford Neal of the University
of Toronto, is a collection of C programs for Bayesian
regression and classification models based on nueral
networks and Gaussian processes, and Bayesian mixture
models. This software is written in C for Unix and
Linux platforms.
- The
Hydra MCMC library,
provided by Gregory Warnes of the University
of Washington, is a Java based library for implementing
Markov Chain Monte Carlo (MCMC) samplers using
Metropolis, Metroplis-Hastings, Gibbs, and adaptive
and multiple chain/parallel methods.
- Alan Genz of Washington State University, has
provided the
BAYESPACK library. BAYESPACK is a suite of
Fortran subroutines for performing high-dimensional
numerical integration relevant to solving many types of
Bayesian problems.
|
