Skip to content

[R--gR] geoR/glsm.mcmc

1 message · Gabrielle Kelly

#
Dear group members,
I have tried using the function likfit in geoRglm and got some output. Howvever my data is 
Bernoulli so it is more appropriate to use glsm.mcmc. I used the following code: 
----------------------------------------------------------------------------------------------------------------- 
testc<-glsm.mcmc(coords=data3geo$coords,data=data3geo$data, 
units.m=data3geo$realisations, model=list(family="binomial", 
link="logit", trend=~x1+x2+x3+x4, 
beta=c(-.05,0.3,0.3,0.3,0.2), 
cov.pars=c(0.5,0.5),cov.model="spherical",nugget=0.3), 
mcmc.input=mcmc.control(S.scale=0.2, thin=1)) 
------------------------------------------------------------------- 
However I got the following error message: 
-------------------------------------------------------------------- 
Error in if (all(round(fisher.l) == 0)) fisher.l <- data * 0.025 : 
 missing value where TRUE/FALSE needed 

Perhaps this is elementary - I would be grateful if someone could point me in the right 
direction.
Thanks in advance.
Gabrielle Kelly, Ph.D.
UCD School of Mathematical Sciences
University College Dublin, Ireland