[R-sig-Geo] help with glsm.mcmc
Sullivan, Mary M
MMSullivan at ric.edu
Fri Apr 18 19:44:35 CEST 2008
I am a novice user of geoRglm and am running into a problem with ghe glsm.mcmc command. My data are binomial (with response a vector of 0 and 1), about 625 cases, with covariates.
My attempts to use the command result in the following:
> lds.2<- list(trend = ~x + y + urban01 + nymphsc + dist2edges + dist2rdss + totedges + H100s + H320s + H400s , beta = 0, cov.model = "exponential", cov.pars = c(1.2,23), nugget = 0.0036, family="binomial")
> mcmc.2.test <-mcmc.control(S.scale=0.2, n.iter=1000)
> test.2.tune <-glsm.mcmc(sampdat.geo, model=lds.2, mcmc.input=mcmc.2.test)
Error in glsm.mcmc(sampdat.geo, model = lds.2, mcmc.input = mcmc.2.test) :
size of beta is incompatible with trend specified
In reading help, I am referred to prepare.likfit.glsm and likfit.glsm, but prepare.likfit.glsm needs the output from the glsm.mcmc, which is what I can't seem to get! Would appreciate help on putting in a beta value. FYI, the cov.pars and nugget came from fitting the exponential variogram.
Mary M. Sullivan, Ed. D.
Professor of Mathematics and Educational Studies
Rhode Island College
600 Mt. Pleasant Ave. Providence, RI 02908
mmsullivan at ric.edu
More information about the R-sig-Geo