I set mh_replic=10000. Acceptance rate happened to be 20%.
Then, I guess that the number of final draws is 2000.
But, actually, the number of draws in “filename_mh_blc.mat” files in metropolis folder is 10000.
Thank you for reply in advance.
I set mh_replic=10000. Acceptance rate happened to be 20%.
Then, I guess that the number of final draws is 2000.
But, actually, the number of draws in “filename_mh_blc.mat” files in metropolis folder is 10000.
Thank you for reply in advance.
You are misunderstanding the working of the Metropolis-Hastings algorithm. It provides correlated draws from the posterior (see Chib/Greenberg 1995 for a good introduction). When a draw is rejected, the old draw becomes the new draw. That’s why there are always as many draws as specified.