I am doing a Bayesian analysis using a spatially varying coefficient model and
I am concerned about how I can extract these spatially varying coefficients for
all 3 variables as given in example below.
It is simple to get these spatially varying coefficients from the spTDyn package
as shown below. However, the spTDyn package takes a huge amount of time
to run the model when we get a couple of more variables, and
the iteration is in thousands. I thought calling the spTDyn in the Bsptime function
computes the results much faster than spTDyn by itself? Could there be any
simple method to get these coefficients? I will then map each of the seven
spatially varying coefficients separately.
I could use the spSVC function from the spBayes package that could finish much faster than the other packages but extracting the spatially varying coefficients is still a problem. Any tips for the spSVC function would be appreciated as well.
library(spTDyn) library(bmstdr) library(spBayes) data("nysptime") f <- y8hrmax~ xmaxtemp + sp(xmaxtemp)+ xwdsp + sp(xwdsp) + xrh + sp(xrh) # using the spTDyn package model1 <- GibbsDyn(formula = f, data = nysptime, coords = ~utmx + utmy, spatial.decay = decay(distribution = Gamm(2,1),tuning = 0.06)) # obtain parameter summary with spatially varying coefficients summary(model1, coefficient = "spatial") # calling the spTDyn package in Bsptime function in bmstdr package model2 <- Bsptime(package = "sptDyn", model = "GP", formula = f, data = nysptime, coordtype = "utm", coords = 4:5)
Thanks for your time and help in this regard.