I’m using a MCMC method to estimate a model in dynare.
The algorithm sorts the values of a jumping normal distribution for the draws of the estimation centered in the last acceptance value, how can I extract from my code the average obtained of the normal distribution in each draw of my estimation, meaning, How can I discover the highest value of the posterior distribution?
I would like to obtained this highest average value of the normal distribution to use in the next estimation to try to get a convergence.
Thanks in advance,
What do you mean with “highest average value”? And which normal distribution are you referring to?
I’m using the Metropolis Hastings method. So, I choose a proposal distribution (prior) with an average and variance covariance matrix for the first draw. For the next ones, I will use as average the last value that I accepted looking at the precedent draw and so on. I would like to identify the highest value for the average that I accepted for my distribution to use for the next draw as average for my distribution and repeat this for the next draws.