Dear BUGS users,
I am trying to implement an unobserved components model (UCM) in BUGS. My
model is a simple Gaussian UCM-AR(2):
y(t)=mu(t)+psi(t)
mu(t) ~ Random Walk
psi(t)~ AR(2)
While my estimated values for this model make sense using the maximum
likelihood procedure, I get completely wrong values in BUGS. I must be doing
something wrong but I cannot figure out what it is.
Could someone please give me some advice.
Thanks,
Tarik
The BUGS code looks like this:
model {
for (t in 1:T){
y[t] <- yobs[t];#100*log(yobs[t]);
}
#observation model
for (t in 1:T){
y[t]~dnorm(meany[t],10000);
meany[t] <- mu[t] + psi[t];
}
#state model
for (t in 3:T){
# Random walk trend with drift
mu[t] ~ dnorm(meanmu[t],tau.nu);
meanmu[t] <-mu[t-1]+beta[t-1] ;
# Random walk
beta[t] ~ dnorm(meanbeta[t],tau.zeta);
meanbeta[t] <- beta[t-1];
# Cycle
psi[t] ~ dnorm(meanpsi[t],tau.kappa);
meanpsi[t] <- phi1*psi[t-1]+phi2*psi[t-2];
}
mu[1] ~ dnorm(6,.001);
mu[2] ~ dnorm(mu[1],.001);
beta[1] ~ dnorm(0,0.001);
beta[2] ~ dnorm(0,0.001);
psi[1] ~ dnorm(1,0.001);
psi[2] ~ dnorm(psi[1],0.001);
tau.nu ~ dgamma(0.1,0.1);
tau.zeta ~ dgamma(0.1,0.1);
tau.kappa ~ dgamma(0.1,0.1);
phi1 ~ dnorm(1.5,0.01);
phi2 ~ dnorm(-0.6,0.01);
-------------------------------------------------------------------
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
|