Dear all,
I have two questions about the MCMC algrithm in Dynare.
-
I want to estimate the postior distribution by 500,000 times, which is common practice in the literature. Because my computer
is not good enough to do it in one time, I let my computer do it accumulatively, by running MCMC 100,000 times at each run by 5 times and using the “load_mh_file” option in the estimation command. Is it correct to do so? If it generates the same answer as when I run MCMC by 500,000 times at once? -
When I do this accumulative computation, sometimes (not always) I got the following error message:
MH: I’m loading past metropolis-hastings simulations…
MH: … It’s done. I’ve loaded 200000 simulations.
??? Error using ==> load
Unable to read file ./NKPC16_Bays8/metropolis/NKPC16_Bays8_mh33_blck1.mat: No such file or directory.
Error in ==> random_walk_metropolis_hastings at 60
load(’./’ MhDirectoryName ‘/’ ModelName ‘_mh’ int2str(NewFile(b)) …
Error in ==> dynare_estimation at 920
feval(options_.posterior_sampling_method,‘DsgeLikelihood’,options_.proposal_distribution,xparam1,invhess,bounds,gend,data);
Error in ==> NKPC16_Bays8 at 269
dynare_estimation(var_list_);
Error in ==> dynare at 102
evalin(‘base’,fname) ;
Can anyone help to figure out what is wrong with it?
Thank you in advance and looking forward to having your answer
Roc