stepan
/
dynare
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Removed prior_penalty in dsge_likelihood. 

If options_.prior_trunc is set to zero (the default is strictly positive) then prior_correction is infinite because the prior density is zero (this is not true for the uniform prior)... This does not help the optimizer. Even if we do not fall in this case (because options_.prior_trunc>0 or becuase only uniform priors are used for the bounded parameters) the meaning of this correction is unclear.
time-shift
Stéphane Adjemian (Charybdis) 2012-07-04 13:04:49 +02:00
parent a05b9d6a8a
commit 379972d715
1 changed files with 1 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
matlab/dsge_likelihood.m
View File

@ -145,7 +145,7 @@ function [fval,DLIK,Hess,exit_flag,ys,trend_coeff,info,Model,DynareOptions,Bayes
% In summary, an initial call to the present function, without triggering
% any condition, guarantees that 'penalty' is properly initialized when needed.
persistent penalty prior_penalty
persistent penalty
% Initialization of the returned variables and others...
fval = [];
@ -189,10 +189,6 @@ end
if ~isequal(DynareOptions.mode_compute,1) && any(xparam1<BayesInfo.lb)
k = find(xparam1<BayesInfo.lb);
fval = penalty+sum((BayesInfo.lb(k)-xparam1(k)).^2);
xtemp=xparam1;
xtemp(k)=BayesInfo.lb(k);
prior_correction = max(0,-prior_penalty-priordens(xtemp,BayesInfo.pshape,BayesInfo.p6,BayesInfo.p7,BayesInfo.p3,BayesInfo.p4));
fval = fval+prior_correction;
exit_flag = 0;
info = 41;
if analytic_derivation,
@ -205,10 +201,6 @@ end
if ~isequal(DynareOptions.mode_compute,1) && any(xparam1>BayesInfo.ub)
k = find(xparam1>BayesInfo.ub);
fval = penalty+sum((xparam1(k)-BayesInfo.ub(k)).^2);
xtemp=xparam1;
xtemp(k)=BayesInfo.ub(k);
prior_correction = max(0,-prior_penalty -priordens(xtemp,BayesInfo.pshape,BayesInfo.p6,BayesInfo.p7,BayesInfo.p3,BayesInfo.p4));
fval = fval+prior_correction;
exit_flag = 0;
info = 42;
if analytic_derivation,
@ -777,7 +769,6 @@ DynareOptions.kalman_algo = kalman_algo;
% Update the penalty.
penalty = fval;
prior_penalty = -lnprior;
if analytic_derivation==0 && nargout==2,
lik=lik(start:end,:);