Skip to content
Snippets Groups Projects
Commit 19b7d202 authored by Willi Mutschler's avatar Willi Mutschler Committed by Johannes Pfeifer
Browse files

testsuite: Add test cases

parent b273a279
Branches
Tags
No related merge requests found
Show whitespace changes
Inline Side-by-side
Showing
with 231 additions and 17 deletions
tests/.gitignore
+ 2
− 0
View file @ 19b7d202
......@@ -50,6 +50,8 @@ wsOct
!/ep/mean_preserving_spread.m
!/ep/rbcii_steady_state.m
!/estimation/fsdat_simul.m
!/estimation/method_of_moments/RBC_MoM_steady_helper.m
!/estimation/method_of_moments/RBC_Andreasen_Data_2.mat
!/expectations/expectation_ss_old_steadystate.m
!/external_function/extFunDeriv.m
!/external_function/extFunNoDerivs.m
......
tests/Makefile.am
+ 3
− 3
View file @ 19b7d202
......@@ -25,8 +25,6 @@ MODFILES = \
measurement_errors/fs2000_corr_me_ml_mcmc/fs2000_corr_ME.mod \
TeX/fs2000_corr_ME.mod \
estimation/MH_recover/fs2000_recover_tarb.mod \
estimation/method_of_moments/RBC_MoM.mod \
estimation/method_of_moments/RBC_MoM_SMM_ME.mod \
estimation/fs2000.mod \
gsa/ls2003a.mod \
optimizers/fs2000_8.mod \
......@@ -50,7 +48,9 @@ MODFILES = \
estimation/t_proposal/fs2000_student.mod \
estimation/tune_mh_jscale/fs2000.mod \
estimation/method_of_moments/AnScho_MoM.mod \
estimation/method_of_moments/RBCmodel_MoM.mod \
estimation/method_of_moments/RBC_MoM_Andreasen.mod \
estimation/method_of_moments/RBC_MoM_SMM_ME.mod \
estimation/method_of_moments/RBC_MoM_prefilter.mod \
moments/example1_var_decomp.mod \
moments/example1_bp_test.mod \
moments/test_AR1_spectral_density.mod \
......
tests/estimation/method_of_moments/AnScho_MoM.mod
+ 1
− 1
View file @ 19b7d202
......@@ -194,7 +194,7 @@ matched_moments_ = {
% , penalized_estimator % use penalized optimization
, pruning % use pruned state space system at higher-order
% , verbose % display and store intermediate estimation results
, weighting_matrix = OPTIMAL % weighting matrix in moments distance objective function; possible values: OPTIMAL|IDENTITY_MATRIX|DIAGONAL|filename
, weighting_matrix = ['optimal'] % weighting matrix in moments distance objective function; possible values: OPTIMAL|IDENTITY_MATRIX|DIAGONAL|filename
, additional_optimizer_steps = [4] % vector of numbers for the iterations in the 2-step feasible method of moments
% , prefilter=0 % demean each data series by its empirical mean and use centered moments
%
......
tests/estimation/method_of_moments/RBC_Andreasen_Data_2.mat 0 → 100644
+ 0
− 0
View file @ 19b7d202
File added
tests/estimation/method_of_moments/RBC_MoM.mod → tests/estimation/method_of_moments/RBC_MoM_Andreasen.mod
+ 202
− 0
View file @ 19b7d202
......@@ -19,7 +19,7 @@
% =========================================================================
% Define testscenario
@#define orderApp = 1
@#define orderApp = 2
@#define estimParams = 1
% Note that we will set the numerical optimization tolerance levels very large to speed up the testsuite
......@@ -32,20 +32,18 @@ shocks;
var u_a; stderr 0.0072;
end;
varobs n c iv;
varobs c iv n;
@#if estimParams == 0
estimated_params;
DELTA, 0.025;
BETTA, 0.98;
B, 0.45;
%ETAl, 1;
ETAc, 1.8;
ALFA, 0.65;
RHOA, 0.95;
stderr u_a, 0.01;
%THETA, 3.48;
BETTA, 0.984;
B, 0.5;
ETAc, 2;
ALFA, 0.667;
RHOA, 0.979;
stderr u_a, 0.0072;
end;
@#endif
......@@ -54,12 +52,10 @@ estimated_params;
DELTA, , 0, 1;
BETTA, , 0, 1;
B, , 0, 1;
%ETAl, 1, 0, 10;
ETAc, , 0, 10;
ALFA, , 0, 1;
RHOA, , 0, 1;
stderr u_a, , 0, 1;
%THETA, 3.48, 0, 10;
end;
@#endif
......@@ -78,9 +74,9 @@ end;
@#endif
% Simulate data
stoch_simul(order=@{orderApp},pruning,nodisplay,nomoments,periods=500);
save('RBC_MoM_data_@{orderApp}.mat', options_.varobs{:} );
pause(1);
%stoch_simul(order=@{orderApp},pruning,nodisplay,nomoments,periods=500);
%save('RBC_MoM_data_@{orderApp}.mat', options_.varobs{:} );
%pause(1);
estimated_params_init(use_calibration);
......@@ -103,38 +99,47 @@ matched_moments_ = {
[ic ] [0 ], [1 ];
[in ] [0 ], [1 ];
[iiv ] [0 ], [1 ];
[ic ic ] [0 0], [1 1];
[ic iiv] [0 0], [1 1];
[ic in ] [0 0], [1 1];
[iiv ic ] [0 0], [1 1];
[iiv iiv] [0 0], [1 1];
%[ic in ] [0 0], [1 1];
%[iiv ic ] [0 0], [1 1];
[iiv in ] [0 0], [1 1];
% [in ic ] [0 0], [1 1];
[iiv iiv] [0 0], [1 1];
[in ic ] [0 0], [1 1];
%[in iiv] [0 0], [1 1];
[in in ] [0 0], [1 1];
[ic ic ] [0 -1], [1 1];
[in in ] [0 -1], [1 1];
[iiv iiv] [0 -1], [1 1];
% [iiv iiv] [0 -1], [1 1];
[ic ic ] [0 -3], [1 1];
[in in ] [0 -3], [1 1];
[iiv iiv] [0 -3], [1 1];
[ic ic ] [0 -5], [1 1];
[in in ] [0 -5], [1 1];
[iiv iiv] [0 -5], [1 1];
};
@#for mommethod in ["GMM", "SMM"]
method_of_moments(
% Necessery options
mom_method = @{mommethod} % method of moments method; possible values: GMM|SMM
, datafile = 'RBC_MoM_data_@{orderApp}.mat' % name of filename with data
mom_method = GMM % method of moments method; possible values: GMM|SMM
, datafile = 'RBC_Andreasen_Data_2.mat' % name of filename with data
% Options for both GMM and SMM
%, bartlett_kernel_lag = 20 % bandwith in optimal weighting matrix
, order = @{orderApp} % order of Taylor approximation in perturbation
, order = 2 % order of Taylor approximation in perturbation
%, penalized_estimator % use penalized optimization
, pruning % use pruned state space system at higher-order
%, pruning % use pruned state space system at higher-order
%, verbose % display and store intermediate estimation results
, weighting_matrix = ['optimal','optimal'] % weighting matrix in moments distance objective function; possible values: OPTIMAL|IDENTITY_MATRIX|DIAGONAL|filename
, weighting_matrix_scaling_factor=1
%, additional_optimizer_steps = [4] % vector of additional mode-finders run after mode_compute
, weighting_matrix = ['DIAGONAL','OPTIMAL'] % weighting matrix in moments distance objective function; possible values: OPTIMAL|IDENTITY_MATRIX|DIAGONAL|filename
%, weighting_matrix_scaling_factor=1
, additional_optimizer_steps = [13] % vector of additional mode-finders run after mode_compute
%, prefilter=0 % demean each data series by its empirical mean and use centered moments
%
% Options for SMM
......@@ -142,6 +147,7 @@ matched_moments_ = {
%, drop = 500 % number of periods dropped at beginning of simulation
%, seed = 24051986 % seed used in simulations
%, simulation_multiple = 5 % multiple of the data length used for simulation
%, burnin = 200
%
% General options
%, dirname = 'MM' % directory in which to store estimation output
......@@ -151,30 +157,29 @@ matched_moments_ = {
%, noprint % do not print stuff to console
%, plot_priors = 1 % control plotting of priors
%, prior_trunc = 1e-10 % probability of extreme values of the prior density that is ignored when computing bounds for the parameters
% , TeX % print TeX tables and graphics
, TeX % print TeX tables and graphics
%
% Data and model options
%, first_obs = 501 % number of first observation
%, logdata % if loglinear is set, this option is necessary if the user provides data already in logs, otherwise the log transformation will be applied twice (this may result in complex data)
%, loglinear % computes a log-linear approximation of the model instead of a linear approximation
%, nobs = 500 % number of observations
%, nobs = 50 % number of observations
% , xls_sheet = willi % name of sheet with data in Excel
% , xls_range = B2:D200 % range of data in Excel sheet
%
% Optimization options that can be set by the user in the mod file, otherwise default values are provided
%, analytic_derivation % uses analytic derivatives to compute standard errors for GMM
%, huge_number=1D10 % value for replacing the infinite bounds on parameters by finite numbers. Used by some optimizers for numerical reasons
, mode_compute = @{optimizer} % specifies the optimizer for minimization of moments distance, note that by default there is a new optimizer
%, optim = ('TolFun', 1e-3
% ,'TolX', 1e-5
% ) % a list of NAME and VALUE pairs to set options for the optimization routines. Available options depend on mode_compute
, mode_compute = 13 % specifies the optimizer for minimization of moments distance, note that by default there is a new optimizer
, optim = ('TolFun', 1D-6
,'TolX', 1D-6
) % a list of NAME and VALUE pairs to set options for the optimization routines. Available options depend on mode_compute
%, silent_optimizer % run minimization of moments distance silently without displaying results or saving files in between
% , tolf = 1e-5 % convergence criterion on function value for numerical differentiation
% , tolx = 1e-6 % convergence criterion on funciton input for numerical differentiation
, se_tolx = 1e-6 % convergence criterion on funciton input for numerical differentiation
%
% % Numerical algorithms options
%, aim_solver % Use AIM algorithm to compute perturbation approximation
% , dr=default % method used to compute the decision rule; possible values are DEFAULT, CYCLE_REDUCTION, LOGARITHMIC_REDUCTION
%, dr=DEFAULT % method used to compute the decision rule; possible values are DEFAULT, CYCLE_REDUCTION, LOGARITHMIC_REDUCTION
%, dr_cycle_reduction_tol = 1e-7 % convergence criterion used in the cycle reduction algorithm
%, dr_logarithmic_reduction_maxiter = 100 % maximum number of iterations used in the logarithmic reduction algorithm
%, dr_logarithmic_reduction_tol = 1e-12 % convergence criterion used in the cycle reduction algorithm
......@@ -187,8 +192,11 @@ matched_moments_ = {
%, sylvester_fixed_point_tol = 1e-12 % convergence criterion used in the fixed point Sylvester solver
%, qz_criterium = 0.999999 % value used to split stable from unstable eigenvalues in reordering the Generalized Schur decomposition used for solving first order problems [IS THIS CORRET @wmutschl]
%, qz_zero_threshold = 1e-6 % value used to test if a generalized eigenvalue is 0/0 in the generalized Schur decomposition
, mode_check
%, mode_check_neighbourhood_size=0.5
%, mode_check_symmetric_plots=0
%, mode_check_number_of_points=25
);
@#endfor
tests/estimation/method_of_moments/RBC_MoM_SMM_ME.mod
+ 14
− 12
View file @ 19b7d202
......@@ -16,11 +16,11 @@
% =========================================================================
% Define testscenario
@#define orderApp = 2
@#define orderApp = 1
@#define estimParams = 0
% Note that we will set the numerical optimization tolerance levels very large to speed up the testsuite
@#define optimizer = 13
@#define optimizer = 5
@#include "RBC_MoM_common.inc"
......@@ -33,14 +33,14 @@ varobs n c iv;
@#if estimParams == 0
estimated_params;
DELTA, 0.02;
BETTA, 0.9;
B, 0.4;
DELTA, 0.025;
BETTA, 0.984;
B, 0.5;
%ETAl, 1;
ETAc, 1.5;
ALFA, 0.6;
RHOA, 0.9;
stderr u_a, 0.010;
ETAc, 1;
ALFA, 0.667;
RHOA, 0.979;
stderr u_a, 0.0072;
%THETA, 3.48;
stderr n, 0.01;
......@@ -76,7 +76,7 @@ end;
@#endif
% Simulate data
stoch_simul(order=@{orderApp},pruning,nodisplay,nomoments,periods=750,drop=500);
stoch_simul(order=@{orderApp},pruning,nodisplay,nomoments,periods=250);
save('RBC_MoM_data_@{orderApp}.mat', options_.varobs{:} );
pause(1);
......@@ -128,8 +128,10 @@ matched_moments_ = {
% , penalized_estimator % use penalized optimization
, pruning % use pruned state space system at higher-order
% , verbose % display and store intermediate estimation results
, weighting_matrix = OPTIMAL % weighting matrix in moments distance objective function; possible values: OPTIMAL|IDENTITY_MATRIX|DIAGONAL|filename
, additional_optimizer_steps = [4] % vector of additional mode-finders run after mode_compute
, weighting_matrix = ['identity_matrix'] % weighting matrix in moments distance objective function; possible values: OPTIMAL|IDENTITY_MATRIX|DIAGONAL|filename
, weighting_matrix_scaling_factor = 10
, burnin=250
%, additional_optimizer_steps = [4] % vector of additional mode-finders run after mode_compute
% , prefilter=0 % demean each data series by its empirical mean and use centered moments
%
% Options for SMM
......
tests/estimation/method_of_moments/RBC_MoM_prefilter.mod
+ 1
− 1
View file @ 19b7d202
......@@ -87,7 +87,7 @@ matched_moments_ = {
% [iiv iiv] [0 -1], [1 1];
};
weighting_matrix=diag([1000;ones(6,1)]);
weighting_matrix=diag([1000;ones(8,1)]);
save('test_matrix.mat','weighting_matrix')
@#for mommethod in ["GMM", "SMM"]
......
tests/estimation/method_of_moments/RBC_MoM_steady_helper.m 0 → 100644
+ 8
− 0
View file @ 19b7d202
function N = RBC_MoM_steady_helper(THETA,ETAl,ETAc,BETTA,B,C_O_N,W)
if ETAc == 1 && ETAl == 1
N = (1-BETTA*B)*(C_O_N*(1-B))^-1*W/THETA/(1+(1-BETTA*B)*(C_O_N*(1-B))^-1*W/THETA);
else
% No closed-form solution use a fixed-point algorithm
N0 = 1/3;
N = fsolve(@(N) THETA*(1-N)^(-ETAl)*N^ETAc - (1-BETTA*B)*(C_O_N*(1-B))^(-ETAc)*W, N0,optimset('Display','off','TolX',1e-12,'TolFun',1e-12));
end
\ No newline at end of file
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Please register or to comment