R Package · INLA · Spatial Statistics
fbesag
A non-stationary extension of the classical Besag model for Bayesian disease mapping and spatial smoothing over heterogeneous sub-regions.
A non-stationary extension of the classical Besag model for Bayesian disease mapping and spatial smoothing over heterogeneous sub-regions.
fbesag is a flexible, non-stationary spatial random effects model built on top of the INLA framework. It extends the classical Besag (ICAR) model by allowing the precision parameter to vary across sub-regions, making it well-suited for modeling spatially heterogeneous phenomena such as disease risk or environmental exposures.
Precision varies across spatial sub-regions, capturing local heterogeneity.
Fully compatible with the R-INLA framework via the cgeneric interface.
Applied to public health surveillance across irregular areal domains.
Install directly from GitHub. Requires a working R toolchain and the devtools package.
library("devtools")
install_github("esmail-abdulfattah/fbesag")
If this fails, run
run_fbesag_option_1.R
from the inst/ folder.
Download cgeneric.h and fbesag.c
from inst/, then compile:
gcc -Wall -fpic -g -O -c -o fbesag.o fbesag.c
gcc -shared -o fbesag.so fbesag.ogcc -Wall -fpic -g -O -c -o fbesag.o fbesag.c
gcc -shared -o fbesag.dll fbesag.o
Then run wrapper.R from inst/.
On Windows, replace fbesag.so with fbesag.dll inside get_fbesag().
No compilation required. Slightly slower than the C backend, but works on any platform.
Run run_fbesag_option_2.R
from the inst/ folder.
fbesag has been applied to disease mapping in Brazil and is suitable for:
We are actively looking for collaborators interested in:
King Abdullah University of Science and Technology (KAUST)
esmail.abdulfattah@kaust.edu.sa