spStack

Overview

spStack is an R package that delivers Bayesian inference for point-referenced spatial data by assimilating posterior inference over a collection of candidate models using stacking of predictive densities. Currently, it supports point-referenced Gaussian, Poisson, binomial and binary outcomes. Users can supply candidate values of spatial process parameters and certain auxiliary model parameters, based on which the collection of models will be created. spStack utilizes the Bayesian conjugate linear modelling framework for Gaussian data and the generalized conjugate multivariate distribution theory for non-Gaussian exponential family data. Learn more in vignette("spStack"). Technical details of the methodology are available in Zhang, Tang, and Banerjee 2024 and Pan, Zhang, Bradley, and, Banerjee 2024.

Installation

For a quick installation of the development version, run the following command in R.

# Install development version from GitHub
# install.packages("pak")
pak::pak("SPan-18/spStack-dev")

If installing from CRAN, use the following.

install.packages("spStack")

Installing from source

To install the package from source, download the tarball spStack_X.X.XX.tar.gz file. After setting the working directory at the file location, either issue R CMD install spStack_X.X.XX.tar.gz in the terminal, or run the following command in R to install the package.

install.packages("spStack_X.X.XX.tar.gz", type = "source", repos = NULL)

Note that the package is written in C++ with calls to FORTRAN routines and hence contains a Makevars file for cross-platform portability. So, it is important to set the correct path to FORTRAN libraries as well as BLAS and LAPACK on your computer. For example, if you are working on MacOS, create a file ~.R/Makevars and set global configurations for the libraries to link with R. The following is an example of such a Makevars file.

# Set Fortran library paths
FLIBS = -L/opt/homebrew/opt/gcc/lib/gcc/14 -lgfortran -lquadmath -lm

# BLAS and LAPACK libraries (using Accelerate framework on macOS)
BLAS_LIBS = -L/System/Library/Frameworks/Accelerate.framework/Versions/Current/ -framework Accelerate
LAPACK_LIBS = -L/System/Library/Frameworks/Accelerate.framework/Versions/Current/ -framework Accelerate

It tells R to use the Accelerate framework, that comes pre-installed with Mac for BLAS and LAPACK functions. If you do not have gfortran, simply run brew install gcc on the terminal which will install the gcc compiler and gfortran comes bundled with gcc. If gcc is installed using Homebrew, then the path should be the same as above, otherwise the path for gfortran needs to set correctly.

Usage

Once successfully installed, load the library in R.

library(spStack)

View the vignette by running vignette("spStack") and see example code to implement predictive stacking for different Bayesian hierarchical spatial models.

Acknowledgement

The graphics used in the logo has been obtained from the page Gaussian RF (Boris Kozintsev, 1999). It represents a realization of an isotropic Gaussian random field under the Matérn correlation function with decay and smoothness parameters 5 and 2, respectively.