I'm running an ecological (log-normal model), two parallel chains and to my
surprise when monitoring convergence for the parameters the Gelman-Rubin
indicates that the parameters have converged while the trace and time series
shows that the two chains are very far from each other thus not converged. Is
this normal? and if so what can be the solution. Below is the code I am using.
model {
for (i in 1:N) {
y[i]~dpois(mu[i])
log(mu[i])<-log(e[i])+alpha+u[i]+v[i]+beta1*nocar[i]+beta2*overcr[i]+beta3*unemp[i]+beta4*lowsc[i]+beta5*age04[i]+beta6*bornoeu[i]+beta7*bornelse[i]+beta8*noqual[i]+beta9*hqual[i]+beta10*lonep[i]+beta11*workagr[i]
#Relative risk
theta[i]<-exp(alpha+u[i]+v[i]+beta1*nocar[i]+beta2*overcr[i]+beta3*unemp[i]+beta4*lowsc[i]+beta5*age04[i]+beta6*bornoeu[i]+beta7*bornelse[i]+beta8*noqual[i]+beta9*hqual[i]+beta10*lonep[i]+beta11*workagr[i])
v[i]~dnorm(0,tau.v)
}
eps<-1.0E-6
u[1:N]~car.normal(adj[],weights[],num[],tau.u)
for (k in 1:sumNumNeigh)
{
weights[k]<-1
}
#Prior on regression coefficients
beta1~dnorm(0.0,1.0E-5)
beta2~dnorm(0.0,1.0E-5)
beta3~dnorm(0.0,1.0E-5)
beta4~dnorm(0.0,1.0E-5)
beta5~dnorm(0.0,1.0E-5)
beta6~dnorm(0.0,1.0E-5)
beta7~dnorm(0.0,1.0E-5)
beta8~dnorm(0.0,1.0E-5)
beta9~dnorm(0.0,1.0E-5)
beta10~dnorm(0.0,1.0E-5)
beta11~dnorm(0.0,1.0E-5)
alpha~dflat()
mean<-exp(alpha)
varu<-1/tau.u
varv<-1/tau.v
tau.u~dgamma(0.1,0.001)
tau.v~dgamma(0.1,0.001)
}
Thank you in advance for your comments.
Oarabile
----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.
-------------------------------------------------------------------
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
|