Dear Bugs Users
I am fitting models at different levels of complexity (i.e. hierarchical
and non-hierarchical) and I am interested in comparing the DIC values.
However, I am not sure why pD is much higher than the known level of
parameters when I do a non-hierarchical parameterization.
The specific model fits a size-selective probability of capture (3
parameter logistic curve) to 22 lakes that have been sampled with nets.
When I do a non-hierarchical model there are 3 selectivity parameters *
22 lakes = 66 paramters, but the pD in the output is approximately 83
(pD = var(deviance/2)). When I run a global model (1 selectivity curve
for all lakes) the pD is reasonably close to the true parameter number
of 3. When I run the hierarchical model the pD is about 66 which is
higher than expected as well.
If anyone has an explanation that would be great.
Thanks in advance,
Paul
The lake specific model is below.
Incidently if I use the logit prior for pmax the number of parameters
drops, but convergence is poor.
model{
for(j in 1:22){
#logit_pmax[j] ~ dunif(-10,10) ;
#pmax[j] <- exp(logit_pmax[j]) / (1 + exp(logit_pmax[j]))
pmax[j] ~ dbeta(1,1)
alpha[j] ~ dunif(-20,0)
beta[j] ~ dbeta(1,1)
#beta[j] ~ dunif(-2,2)
}
for(i in 1:N){
p[i] <- pmax[lake[i]] * (exp(alpha[lake[i]] + beta[lake[i]]*l[i])/(1
+ exp(alpha[lake[i]] + beta[lake[i]]*l[i])))
r[i] ~ dbin(p[i],n[i])
}
}
-------------------------------------------------------------------
This list is for discussion of modelling issues and the BUGS software.
For help with crashes and error messages, first mail [log in to unmask]
To mail the BUGS list, mail to [log in to unmask]
Before mailing, please check the archive at www.jiscmail.ac.uk/lists/bugs.html
Please do not mail attachments to the list.
To leave the BUGS list, send LEAVE BUGS to [log in to unmask]
If this fails, mail [log in to unmask], NOT the whole list
|