|
Activity Number:
|
2
|
|
Type:
|
Invited
|
|
Date/Time:
|
Sunday, August 2, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistical Computing
|
| Abstract - #303029 |
|
Title:
|
Ramps: An R Package for Unified Geostatistical Modeling of Complex Spatio-Temporal Data
|
|
Author(s):
|
Brian J. Smith*+ and Jun Yan
|
|
Companies:
|
The University of Iowa and University of Connecticut
|
|
Address:
|
200 Hawkins Drive, C22 GH, Iowa City, IA, 52242,
|
|
Keywords:
|
Geostatistics ; Data Fusion ; Bayesian Inference ; Statistical Computing ; Markov chain Monte Carlo ; R
|
|
Abstract:
|
We introduce and demonstrate the R package "ramps" which implements reparameterized and marginalized posterior sampling (RAMPS) for complex Bayesian geostatistical models. Our package allows joint modeling of areal and point-source data arising from the same underlying spatial process. The reparameterization of variance parameters facilitates a slice sampling algorithm based on simplexes, which can be useful in general when multiple variances are present. The implementation takes advantage of sparse matrix operations in the "Matrix" package and can provide substantial savings in computing time for large data sets. Support is provided for numerous spatial and spatiotemporal correlation structures, user-defined correlation structures, and non-spatial random effects. An intuitive interface enables users to analyze data sets and plot results will little programming effort.
|
