Abstract Details
Activity Number:
|
586
|
Type:
|
Invited
|
Date/Time:
|
Thursday, August 7, 2014 : 8:30 AM to 10:20 AM
|
Sponsor:
|
Section on Statistical Computing
|
Abstract #310915
|
|
Title:
|
Beyond the Black Box: Flexible Programming of Hierarchical Modeling Algorithms for BUGS-Compatible Models Using NIMBLE
|
Author(s):
|
Perry de de Valpine*+ and Daniel Turek and Christopher J. Paciorek and Rastislav Bodik and Duncan Temple Lang
|
Companies:
|
University of California, Berkeley and University of California, Berkeley and University of California and University of California, Berkeley and University of California
|
Keywords:
|
BUGS ;
MCMC ;
Particle Filter ;
hierarchical models ;
Monte Carlo
|
Abstract:
|
We introduce a domain-specific language embedded in R for programming hierarchical model algorithms for models declared in the BUGS language. Various software packages allow flexible specification of hierarchical models and provide black box algorithms such as specific types of MCMC, particle filter (PF), Laplace approximation, or others. However, many new and old algorithms remain inaccessible for practical use without re-writing them for each model. Moreover, having different packages for different algorithms makes it difficult to combine methods or try different methods on the same problem. The new R-based NIMBLE language allows flexible programming of algorithms with access to the model structure declared by BUGS code. BUGS code is processed into model-specific C++ code, compiled, and interfaced with R. Functions in NIMBLE can use the model structure to allow automatic specialization to the details of any model. Once specialized to a model, functions can be processed into C++, compiled, and interfaced with R. With R's CRAN package system, developers will be able to distribute new algorithms written in NIMBLE. Examples with MCMC, PF, and more will be presented.
|
Authors who are presenting talks have a * after their name.
Back to the full JSM 2014 program
|
2014 JSM Online Program Home
For information, contact jsm@amstat.org or phone (888) 231-3473.
If you have questions about the Professional Development program, please contact the Education Department.
The views expressed here are those of the individual authors and not necessarily those of the JSM sponsors, their officers, or their staff.
Copyright © American Statistical Association.