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

- store oo_.FilteredVariablesKStepAheadVariances only if options_.filter_covariance. this can save a lot of memory for large models 

- oo_.FilteredVariables must be stored independently to a previous MCMC. evaluate_smoother must in any case provide full info in oo_
- options_.filtered_vars does not always imply filtered variables are computed. options_.nk is more robust.
time-shift
Marco Ratto 2017-01-10 14:21:35 +01:00 committed by Stéphane Adjemian (Charybdis)
parent 510bbe200e
commit 4b2d139c7a
2 changed files with 5 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
matlab/prior_posterior_statistics.m
View File

@ -337,7 +337,7 @@ if options_.smoother
end
end
if options_.filtered_vars
if ~isempty(options_.nk) && options_.nk > 0
pm3(endo_nbr,gend,ifil(4),B,'One step ahead forecast (filtered variables)',...
'',varlist,M_.endo_names_tex,M_.endo_names,...
varlist,'FilteredVariables',DirectoryName,'_filter_step_ahead');

8
matlab/store_smoother_results.m
View File

@ -136,7 +136,7 @@ if ~isempty(options_.nk) && options_.nk ~= 0
end
% add constant
oo_.FilteredVariablesKStepAhead = aK(options_.filter_step_ahead,i_endo_in_dr_matrices,:)+constant_all_variables;
if ~isempty(PK) %get K-step ahead variances
if ~isempty(PK) && options_.filter_covariance %get K-step ahead variances
oo_.FilteredVariablesKStepAheadVariances = ...
PK(options_.filter_step_ahead,i_endo_in_dr_matrices,i_endo_in_dr_matrices,:);
end
@ -157,7 +157,7 @@ for i_endo_in_bayestopt_smoother_varlist=bayestopt_.smoother_saved_var_list'
end
oo_.Smoother.Constant.(deblank(M_.endo_names(i_endo_declaration_order,:)))=constant_current_variable;
oo_.SmoothedVariables.(deblank(M_.endo_names(i_endo_declaration_order,:)))=atT(i_endo_in_dr,:)'+constant_current_variable;
if ~isempty(options_.nk) && options_.nk > 0 && ~((any(bayestopt_.pshape > 0) && options_.mh_replic) || (any(bayestopt_.pshape> 0) && options_.load_mh_file))
if ~isempty(options_.nk) && options_.nk > 0 % && ~((any(bayestopt_.pshape > 0) && options_.mh_replic) || (any(bayestopt_.pshape> 0) && options_.load_mh_file))
oo_.FilteredVariables.(deblank(M_.endo_names(i_endo_declaration_order,:)))=squeeze(aK(1,i_endo_in_dr,2:end-(options_.nk-1)))+constant_current_variable;
end
oo_.UpdatedVariables.(deblank(M_.endo_names(i_endo_declaration_order,:)))=updated_variables(i_endo_in_dr,:)'+constant_current_variable;
@ -191,7 +191,7 @@ if ~isempty(options_.nk) && options_.nk ~= 0
positions_in_declaration_order=oo_.dr.order_var(bayestopt_.smoother_var_list(bayestopt_.smoother_saved_var_list));
if ~(options_.selected_variables_only && ~(options_.forecast > 0)) %happens only when selected_variables_only is not used
oo_.FilteredVariablesKStepAhead(:,positions_in_declaration_order,:)=oo_.FilteredVariablesKStepAhead;
if ~isempty(PK) %get K-step ahead variances
if ~isempty(PK) && options_.filter_covariance %get K-step ahead variances
oo_.FilteredVariablesKStepAheadVariances(:,positions_in_declaration_order,positions_in_declaration_order,:)=oo_.FilteredVariablesKStepAheadVariances;
end
if ~isempty(decomp)
@ -201,7 +201,7 @@ if ~isempty(options_.nk) && options_.nk ~= 0
positions_in_declaration_order=oo_.dr.order_var(bayestopt_.smoother_var_list(bayestopt_.smoother_saved_var_list));
[junk,sorted_index_declaration_order]=sort(positions_in_declaration_order);
oo_.FilteredVariablesKStepAhead(:,sorted_index_declaration_order,:)=oo_.FilteredVariablesKStepAhead;
if ~isempty(PK) %get K-step ahead variances
if ~isempty(PK) && options_.filter_covariance %get K-step ahead variances
oo_.FilteredVariablesKStepAheadVariances(:,sorted_index_declaration_order,sorted_index_declaration_order,:)=oo_.FilteredVariablesKStepAheadVariances;
end
if ~isempty(decomp)