Page 1 of 1
Log data density [Laplace approximation] is NaN.
Posted:
Fri May 28, 2010 9:38 amby coot
Hello.
When I run a DSGE model (tried compute_mode from 1 to 7 [except 2]), I get errors, usually something related to
Log data density [Laplace approximation] is NaN.
what is the problem ??
Re: Log data density [Laplace approximation] is NaN.
Posted:
Fri May 28, 2010 9:58 amby jpfeifer
Please post the mod-file.
Re: Log data density [Laplace approximation] is NaN.
Posted:
Fri May 28, 2010 5:56 pmby coot
Solved the problem. I missed somethings...
But I get strange posterior mode and distribution graphs. Any ideea or comment would be ok...
One more thing, my model fits the data almost perfect, and all measurement errors are 0
Re: Log data density [Laplace approximation] is NaN.
Posted:
Sat May 29, 2010 6:19 amby jpfeifer
Again, without the mod-file, the data-file and some information what the variables in your posted PDF mean, it's hard to judge what's going on. Did you correctly specify the measurement equation, taking care of stationarity? And how many observables and shocks do you have?
Re: Log data density [Laplace approximation] is NaN.
Posted:
Sat May 29, 2010 12:08 pmby coot
I've solved the problems. I'll post the files when I run out of ideas. The model has about 70 equations, 15 observable variables, it is the one developed in Adolfson et. al 2005. Now there is no reason for me to post the files because I find every time some mistakes to correct them. But when I am finished or out of ideas I'll post it.
