I am currently trying to use MCMC technqiues to produce mean estimates of animal tracking locations. I am using JAGS (Just Another Gibbs Sampler) in R. Now, I am trying to find out what the most suitable parameters for nr. iterations after the burn-in, burn-in and thinning intervals are.
My idea is to use the Rafter Lewis diagnostic with the raftery.diag() function from the package "coda". However, I am a bit confused about the output. According to the documentation, the output of the function contains a 3d array with M = suggested burn-in, N = suggested number of iterations and Nmin = suggested number of iterations based on zero autocorrelation.
My two questions are:
What does "Error" "3746" in the resmatrix mean and how can I solve this error?
How can I find the suggested thinning interval in the raftery.diag() output?