I know it is a stupid question…but I really can not find the marginal data density code in manual or user guide.is it in the “estimate”?I don not think it works when I code “marginal_density()” after "estimate"command.
I will be really grateful if someone can give a example.
btw,Is it the same as marginal likelihood?I believe many people had compared two models by this index,really need a help!
yaliallen
The log marginal data density will be computed automatically by the estimation-command if you use Bayesian estimation. As written in the manual, it is stored in oo_.MarginalDensity.
Yes, this it the statistic used for model comparison and is also referred to a the marginal likelihood or model evidence.
[quote=“jpfeifer”]The log marginal data density will be computed automatically by the estimation-command if you use Bayesian estimation. As written in the manual, it is stored in oo_.MarginalDensity.
Yes, this it the statistic used for model comparison and is also referred to a the marginal likelihood or model evidence.[/quote]
thank you mr pfeifer for helping me. but i cannot find where to see this output. if i should type a additional code “oo_.MarginalDensity” under estimation? I try this,but it shows:
[quote]??? Reference to non-existent field ‘MarginalDensity’.
Error in ==> LJWSIGEQZ at 1805
oo_.MarginalDensity;
Error in ==> dynare at 120
evalin(‘base’,fname) ;[/quote]
or
[quote]Starting Dynare (version 4.3.3).
Starting preprocessing of the model file …
ERROR: LJWSIGEQZ.mod:127.69-71: syntax error, unexpected NAME[/quote]
if not,as [Marginal Data Density) said,I didn’t say the same result with my code.
really want to know what goes wrong…Thank u!
attachments is my code and data
LJWSIGEQZ.mod (3.08 KB)
data07.m (294 Bytes)
The marginal data density is a Bayesian concept. You are estimating the model with ML and not with Bayesian techniques. That won’t work.