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

Add headers to identification routines

time-shift
Johannes Pfeifer 2016-03-29 15:58:32 +02:00
parent f37d3afb14
commit bf5a3ab6da
3 changed files with 74 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

34
matlab/getH.m
View File

@ -1,8 +1,38 @@
function [H, dA, dOm, Hss, gp, d2A, d2Om, H2ss] = getH(A, B, M_,oo_,options_,kronflag,indx,indexo,iv)
% function [H, dA, dOm, Hss, gp, d2A, d2Om, H2ss] = getH(A, B, M_,oo_,options_,kronflag,indx,indexo,iv)
% computes derivative of reduced form linear model w.r.t. deep params
%
% Copyright (C) 2010-2012 Dynare Team
% Inputs:
% A: Transition matrix of lagged states from Kalman filter
% B: Matrix in state transition equation mapping shocks today to
% states today
% M_: structure storing the model information
% oo_: structure storing the results
% options_: structure storing the options
% kronflag: Indicator whether to rely on Kronecker products (1) or
% not (-1 or -2)
% indx: Index of estimated parameters in M_.params
% indexo: Index of estimated standard deviations in M_.exo_names
% iv: Index of considered variables
%
% Outputs:
% H: dTAU/dTHETA: Jacobian of TAU, vectorized form of
% linearized reduced form state space model, given ys [steady state],
% A [transition matrix], B [matrix of shocks], Sigma [covariance of shocks]
% TAU = [ys; vec(A); dyn_vech(B*Sigma*B')].
% dA: [endo_nbr by endo_nbr by (indx+indexo)] Jacobian of transition matrix A
% dOm: [endo_nbr by endo_nbr by (indx+indexo)] Jacobian of Omega = (B*Sigma*B')
% Hss: [endo_nbr by (indx)] Jacobian of steady state with respect to estimated
% structural parameters only (indx)
% gp: Jacobian of linear rational expectation matrices [i.e.
% Jacobian of dynamic model] with respect to estimated
% structural parameters only (indx)
% d2A: Hessian of transition matrix A
% d2Om: Hessian of Omega
% H2s: Hessian of steady state with respect to estimated
% structural parameters only (indx)
% Copyright (C) 2010-2016 Dynare Team
%
% This file is part of Dynare.
%

39
matlab/getJJ.m
View File

@ -1,6 +1,43 @@
function [JJ, H, gam, gp, dA, dOm, dYss] = getJJ(A, B, M_,oo_,options_,kronflag,indx,indexo,mf,nlags,useautocorr)
% function [JJ, H, gam, gp, dA, dOm, dYss] = getJJ(A, B, M_,oo_,options_,kronflag,indx,indexo,mf,nlags,useautocorr)
% computes derivatives of 1st and 2nd order moments of observables with
% respect to estimated parameters
%
% Inputs:
% A: Transition matrix of lagged states from Kalman filter
% B: Matrix in state transition equation mapping shocks today to
% states today
% M_: structure storing the model information
% oo_: structure storing the results
% options_: structure storing the options
% kronflag: Indicator whether to rely on Kronecker products (1) or
% not (-1 or -2)
% indx: Index of estimated parameters in M_.params
% indexo: Index of estimated standard deviations in M_.exo_names
% mf: Index of observed variables
% nlags: Number of lags to consider for covariances and
% correlations
% useautocorr: Indicator on whether to use correlations (1) instead of
% covariances (0)
%
% Outputs:
% JJ: Jacobian of 1st and 2nd order moments of observables, i.e. dgam/dTHETA
% (see below for definition of gam)
% H: dTAU/dTHETA: Jacobian of TAU, vectorized form of
% linearized reduced form state space model, given ys [steady state],
% A [transition matrix], B [matrix of shocks], Sigma [covariance of shocks]
% TAU = [ys; vec(A); dyn_vech(B*Sigma*B')].
% gam: vector of theoretical moments of observed variables mf [JJ is the Jacobian of gam].
% gam = [ys(mf); dyn_vech(GAM{1}); vec(GAM{j+1})]; for j=1:ar and where
% GAM is the first output of th_autocovariances
% gp: Jacobian of linear rational expectation matrices [i.e.
% Jacobian of dynamic model] with respect to estimated
% structural parameters only (indx)
% dA: [endo_nbr by endo_nbr by (indx+indexo)] Jacobian of transition matrix A
% dOm: Jacobian of Omega = (B*Sigma*B')
% dYss Jacobian of steady state with respect to estimated structural parameters only (indx)
% Copyright (C) 2010-2012 Dynare Team
% Copyright (C) 2010-2016 Dynare Team
%
% This file is part of Dynare.
%

7
matlab/identification_analysis.m
View File

@ -7,11 +7,12 @@ function [ide_hess, ide_moments, ide_model, ide_lre, derivatives_info, info] = i
% o indx [array] index of estimated parameters
% o indexo [array] index of estimated shocks
% o options_ident [structure] identification options
% o data_info [structure] data info for Kalman Filter
% o dataset_ [structure] the dataset after required transformation
% o dataset_info [structure] Various informations about the dataset (descriptive statistics and missing observations) info for Kalman Filter
% o prior_exist [integer]
% =1 when prior exists and indentification is checked only for estimated params and shocks
% =0 when prior is not defined and indentification is checked for all params and shocks
% o nem_tex [char] list of tex names
% o name_tex [char] list of tex names
% o init [integer] flag for initialization of persistent vars
%
% OUTPUTS
@ -25,7 +26,7 @@ function [ide_hess, ide_moments, ide_model, ide_lre, derivatives_info, info] = i
% SPECIAL REQUIREMENTS
% None
% Copyright (C) 2008-2013 Dynare Team
% Copyright (C) 2008-2016 Dynare Team
%
% This file is part of Dynare.
%