Big Likelihood and Marginal likelihood computation
Posted: Fri Jun 20, 2008 4:10 pmDear all,
I have one question about the computation of marginal data density in Dynare. When the likelihood of the DSGE is very big (because if we have logL = 3000, exp(logL) gives Inf in matlab) so that the mdd will tend to zero, how does Dynare compute the mdd taking this fact into account?
Thanks in advance for your reply.
Aqua.
I have one question about the computation of marginal data density in Dynare. When the likelihood of the DSGE is very big (because if we have logL = 3000, exp(logL) gives Inf in matlab) so that the mdd will tend to zero, how does Dynare compute the mdd taking this fact into account?
Thanks in advance for your reply.
Aqua.