Hi everyone,
I got this results when I ran the code for Bayesian estimation. How to understand this result? What kind of information we can get from this result?
Thank you,
Alex
mode_compute=6 does iterated runs of the Metropolis-Hastings algorithm to get the mode and the covariance matrix in the posterior. You are seeing two iterations of that. The first two outputs show the norm of the change in the parameter mean and covariance. Next, the current mode is displayed and the improvement relative to the previous iteration. That change should be rather small as otherwise, you are not yet close to the actual mode. Finally, the optimal scaling parameter for tuning the acceptance rate is displayed.
Thank you Prof. Pfeifer for answering my question.
In the third line, it is the current mode. for the mode, is it correct to say the bigger, the better? Also, In most cases, the mode is a negative number. However, I do get positive number when I was trying different senarios. Can we say the positive current mode is better than a negative one. Thank you Prof. Pfeifer.
Yes, the higher, the better. That’s the point of the mode. It’s supposed to be the highest point. Whether it’s negative or positive does not really matter. It’s the logarithm of a product of densities. If the densities are smaller than 1, you will be summing negative numbers.
Thank you Prof. Pfeifer for answeing my question.