A Question on Bayesian Estimation

Hi, guys,
when I tried to estimate a model, the following information is printed in screen:

 ??? Error using ==> chol

Matrix must be positive definite.

Error in ==> gmhmaxlik at 197
dd = transpose(chol(CovJump));

Error in ==> dynare_estimation_1 at 437
[xparam1,PostVar,Scale,PostMean] = …

Error in ==> dynare_estimation at 89

Error in ==> NewModel at 606

Error in ==> dynare at 180
evalin(‘base’,fname) ;

Also, the progress bar of 'Tuning of scale parameter" just get stopped.

Is there anything wrong with the model?


Search the forum, there must be dozens of posts on this. The most common problem is that the data treatment does not match the observables defined in the model. See Pfeifer(2013): “A Guide to Specifying Observation Equations for the Estimation of DSGE Models” sites.google.com/site/pfeiferecon/Pfeifer_2013_Observation_Equations.pdf

Hi, Professor Pfeifer,
Thanks for your answer; I’ve fixed the problem now and got the estimation result. A question relevant to the estimation results is that, there’re two estimation results, the first is RESULTS from POSTERIOR MAXIMIZATION (which comes before MCMC) and the second one is ESTIMATION RESULTS (which is based on MCMC); in the table of my paper, I should report the second result rather than the first one, is it correct?

Thanks again for your kindly help.

Yes, papers doing Bayesian estimation report the parameters from posterior sampling via MCMC

Thanks a lot, Professor Pfeifer!