41e027aee3
added steady option stop_on_error. stop_on_error = 0 permits to
...
continue (for example, when chaining homotopy steps), even if the
steady state computation was not successful.
2012-04-16 22:40:50 +02:00
1014bb038d
made loglinear option available for purely backward and purely forward
...
models (for now only with order==1)
2012-04-15 14:06:37 +02:00
2f5d76dc8c
ms-sbvar: set A0, A+, Zeta, and Q based on output_file_tag passed to estimation
2012-04-13 16:12:24 +02:00
b96c35a1a4
ms-sbvar: remove unused code
2012-04-13 15:56:25 +02:00
a44fa426c4
Simplify logic of test since Octave 3.2 is now the minimal requirement
2012-04-04 22:16:39 +02:00
339d88220d
qz() has changed between Octave 3.2 and 3.4
2012-04-04 22:13:34 +02:00
9bcb1883bd
fixing bug related to use of QZ decomposition in Octave
2012-04-04 21:46:24 +02:00
d4723038be
1) fixed infinite loop
...
2) fixed names of sylvester functions
2012-04-04 16:43:44 +02:00
32de0fd675
fixed typos
2012-04-04 16:41:49 +02:00
f455557376
compact all sylvester routines, with generic dimension of d argument
2012-04-04 16:38:22 +02:00
76093094a6
Force precision in sylvester equation solution, particularly useful for different behaviour of the function schur under octave
2012-04-04 10:39:19 +02:00
a49e550448
typo fixed
2012-04-04 10:36:54 +02:00
fe6aa88c27
1) check if AHess is positive semi-definite;
...
2) fix memory issues related to chh
3) fix issue for the rare case of empty indok (octave)
2012-04-04 10:36:32 +02:00
3ca9485edd
bug fix when calling gsa
...
set default order=1
set analytic derivation =1
2012-04-04 10:35:18 +02:00
28d985433b
bug fix: wrong index used in test
2012-04-03 18:33:24 +02:00
08bcf39fe7
estimation: clear subsamples substructure whenever a subsamples statement is encountered
2012-04-03 18:33:23 +02:00
7e6bd80f17
estimation: remove s from parameters substructure
2012-04-03 18:33:23 +02:00
31e0bc2e1e
estimation: simplify global_initialization
2012-04-03 18:33:23 +02:00
2057365143
Require at least Octave 3.2.2 (and therefore drop rcond.m and bicgstab.m)
...
Octave 3.2 has been around for a long time, and is included in most
distributions.
2012-04-03 17:57:38 +02:00
90cb2ffb59
Under Octave >= 3.4, don't sleep for 2 seconds
2012-04-03 17:46:48 +02:00
3be6b610c8
Merge remote-tracking branch 'houtanb/master'
2012-04-02 15:18:12 +02:00
a35e0cb3e3
fixed bug in initial estimation checks when mode_compute == 5
2012-03-31 11:01:32 +02:00
9741b3b3ec
estimation: reorganize options output
2012-03-30 15:37:15 +02:00
93db8937ab
estimation: reorganize Matlab structure output
2012-03-30 15:37:14 +02:00
2be333da55
estimation: change name handling for correlation statements
2012-03-30 15:37:14 +02:00
be33621ca5
Fixed texinfo header + Cosmetic changes.
2012-03-30 12:08:13 +02:00
9bfc3f6ec6
Fixed bugs and typos.
2012-03-30 12:00:08 +02:00
2db0c98b9b
Fixed texinfo header.
2012-03-30 11:55:01 +02:00
795d0fce16
Cosmetic changes.
2012-03-30 11:48:23 +02:00
6c6b62937f
Added smooth resampling.
2012-03-30 11:47:24 +02:00
294dbdc52c
Fixed bug in the evaluation of the likelihood.
2012-03-30 11:21:24 +02:00
9cc4661942
Fixed bug in initial_estimation_checks + Cosmetic changes.
2012-03-30 11:12:59 +02:00
7feb27fe1c
estimation: rewrite options output
2012-03-29 16:45:46 +02:00
d54e4c4b17
estimation: rewrite prior output
2012-03-29 16:07:01 +02:00
46a655fbd5
estimation: rewrite subsamples statement as an actual statement
2012-03-28 18:51:52 +02:00
90e5c6e7c6
Replace str2mat (deprecated in MATLAB, removed in Octave 3.6) by char
2012-03-21 10:11:17 +01:00
a899594f3d
initializing globals ys0_ ex0_ ex_det0_
2012-03-17 20:34:58 +01:00
41db06f5ae
removing dr1.m
2012-03-15 16:50:41 +01:00
8fa6a9883a
replacing call to dr1() by call to stochastic_solvers()
2012-03-15 16:45:25 +01:00
d6f8e252e5
Merge remote-tracking branch 'origin/master' into dr1break
2012-03-15 16:18:37 +01:00
5d4d318bc7
fixing bug in estimation by maximum likelihood
2012-03-15 15:33:02 +01:00
5954cf2011
Disable more warnings under Octave
2012-03-14 16:39:34 +01:00
e62613e6cb
Fix string warnings for Octave 3.6
2012-03-14 16:35:20 +01:00
c5fa0b82a9
Merge remote-tracking branch 'marco/master'
2012-03-14 12:49:56 +01:00
ebfeae2ccf
Provisions for MATLAB 7.14 (R2012a)
2012-03-13 17:43:30 +01:00
75675d3677
Code simplification, do not pass endo_simul and exo_simul to bytecode through the global structure oo_.
2012-03-12 10:34:28 +01:00
4aedf7e9c0
Added a commented line to print the effective number of particles if needed.
2012-03-12 10:29:25 +01:00
1fcf708b80
Merge remote-tracking branch 'origin/master' into dr1break
2012-03-10 18:21:14 +01:00
4ba016cc16
Merge remote-tracking branch 'ferhat/master'
2012-03-09 20:13:22 +01:00
323e6fdf4f
Added a second gstep option used as a parameter for the routine computing the hessian matrix.
...
In some cases, for instance for the non linear filters, it helps to reduce this new gstep parameter
to get a positive definite hessian matrix. options_.gstep is now a 2*1 vector. The first element is
the old gstep parameter, the second element is the new gstep parameter. The step defined for the
computation of the hessian matrix is now:
h1=max(abs(x),sqrt(gstep(1))*ones(n,1))*eps^(1/6)*gstep(2);
2012-03-09 16:36:26 +01:00
13eb218513
Fixed bug (ub is unknown but defined as a field in bayestopt_).
2012-03-09 15:06:04 +01:00
1ac55d8be8
Fixed bug.
2012-03-09 12:57:58 +01:00
ec9b47426a
Added an option for the threshold level of neff in the generic case of the resampling routines.
2012-03-09 12:44:25 +01:00
ec2528ae9b
estimation: fixes to options statement
2012-03-09 11:47:25 +01:00
75331d7a4f
The global oo_.steady_state should also be updated in the homotopy loop
2012-03-09 11:47:13 +01:00
dd67a81e57
estimation: place parameter priors in their own substructure
2012-03-09 11:46:09 +01:00
7889ac21b2
Merge remote-tracking branch 'ferhat/master'
2012-03-09 10:30:01 +01:00
ad580b8cc6
MS-SBVAR: clean up plotting functions
2012-03-08 18:25:37 +01:00
d212772212
MS-SBVAR: change state to regime to coincide with name previous name change
2012-03-08 17:55:40 +01:00
402363bab2
MS-SBVAR: rewrite ms_variance_decomposition
2012-03-08 17:54:35 +01:00
5f60f1bd27
MS-SBVAR: rewrite ms_forecast
2012-03-08 17:54:05 +01:00
11c63e5050
MS-SBVAR: clean up plot_ms_*.m
2012-03-08 17:54:05 +01:00
4e41ad7285
MS-SBVAR: remove unused argument to plot_ms_irf.m
2012-03-08 17:54:05 +01:00
864e0b7c7e
MS-SBVAR: rewrite ms_irf
2012-03-08 17:53:58 +01:00
f12e8a3754
Force the weights to sum up to one, and not to the number of particles as in WK (Econometrica 2010).
2012-03-08 16:36:31 +01:00
8184ff3fdd
Fixed bug (wrong calling to the resample routine: missing inputs).
2012-03-08 15:53:20 +01:00
7f1ab166f5
Fixed bug.
2012-03-08 15:44:14 +01:00
dfbd6c8f2a
Adds the index of equations (M_.blocksEQU) for a block decomposed model without bytecode option and use this index in resid.m to retrieve the residuals
2012-03-08 15:29:56 +01:00
d995234003
Fixed bugs.
2012-03-08 12:49:17 +01:00
f932fc8c78
Return the whole simulated paths instead of only the contemporaneous reaction of the endogenous variables.
...
Otherwise we can not test the number of periods and we cannot use the previous solution
to initialize the following (stochastic) perfect foresight problem.
2012-03-08 12:49:04 +01:00
464dcb651e
Changed defaults related to extended path. Removed some unused options.
2012-03-08 12:45:53 +01:00
895b58cec4
Merge remote-tracking branch 'ferhat/master'
2012-03-07 09:22:42 +01:00
de07d3b1d7
Removed useless else statement.
2012-03-06 22:54:21 +01:00
93fa8a6c94
Bug fix. Definition of the penalty when the updated parameters are complex.
2012-03-06 22:54:21 +01:00
746fb8ca56
Fixed bug related to the update of the deep parameters in the steady state model.
...
When a deep parameter was initialized to NaN (this is the default) any update of this parameter
done in the steady state model block was not propagated to M_.params. The problem was caused by
the definition of updated_params_flag.
2012-03-06 22:54:20 +01:00
8ac14f8c80
Adds fixed point solvers for Sylvester and Lyapunov equations
2012-03-06 12:03:23 +01:00
a2f3a53634
Parallelization of local_state_space_iteration_2 (used in non linear filters).
2012-03-05 23:11:49 +01:00
6c091015eb
Fixed texinfo header.
2012-03-05 23:11:49 +01:00
cf19df92ef
Added an option fot the mode_check plots for defining the size of the neighbourhood around the estimated posterior mode.
2012-03-05 15:03:55 +01:00
cf9e8a6714
Bug fix. When the estimation command is used, do not issue an error message if options_.order>1. In this
...
case reset the options_.order equal to one (except if options_.particle.status==1).
2012-03-05 14:56:02 +01:00
0d58e1855d
Fixed compatibility issue with octave (use set_dynare_seed routine instead of the matlab's commands
...
to set the seed of the random number generators).
2012-03-05 14:37:55 +01:00
59d3d30c90
Fixed typo. Do not use the non linear filter automatically when order==2, otherwise the non linear filters are used if the estimation command is used after a stoch_simul command with order==2. The non linear filter is used only if options_.particle.status is explicitely set to 1 (before the estimation command). A new estimation option is needed...
2012-03-05 14:21:52 +01:00
d64817437e
now extended_path with order = 0 calls solve_perfect_foresight_model()
...
and order ~= 0 solve_stochastic_perfect_foresight_model().
checking only first period of simulation for convergence when checking horizon
2012-03-05 12:35:55 +01:00
e1a5e0de53
Objective function needs to be passed as a string to CMAES optimization routine.
2012-03-04 22:13:00 +01:00
83b536c618
Moved CMAES options in global_initialization, so that the options of this routine can be modified by
...
the user by writting in options_.cmaes.
2012-03-04 22:13:00 +01:00
cc281897e0
Added new routine for non linear filter (auxiliary particle filter).
2012-03-04 22:13:00 +01:00
8de93c21f2
Fixed bug (cost_flag->exit_flag).
2012-03-04 22:13:00 +01:00
724b99412f
Added option for kitagawa/stratified approach in traditional resampling.
2012-03-04 22:12:59 +01:00
3b471e3a99
Bug fix. Added missing input (method specification) to the resampling routine.
2012-03-04 22:12:59 +01:00
6b736cf622
Bug fix. Added missing input to the non linear likelihood routine (DynareOptions).
2012-03-04 22:12:59 +01:00
fb4def04bd
Fixed bug related to the name of the sub-structure for the non linear filters.
2012-03-04 22:12:59 +01:00
bbf63f0da9
Deleted trailing white spaces.
2012-03-04 22:12:59 +01:00
978f29f59d
Added missing field in bayestopt_ (restrict_var_list, indices for the state and observed variables).
2012-03-04 22:12:59 +01:00
5458492674
Fixed bug.
2012-03-04 22:12:59 +01:00
f3e27e82d6
Fixed typo (misspelled EstimatedParameters structure).
2012-03-04 22:12:59 +01:00
1ab6aae56e
Fixed typo (number of observed variables).
2012-03-04 22:12:59 +01:00
f798118902
Updated test on options_.order (current version can estimate dsge models with non linear filters considering a second order approximation of the model).
2012-03-04 22:12:59 +01:00
6a8095303e
Added particle filter flag.
2012-03-04 22:12:59 +01:00
2f3c8f910f
Fixed bug related to the dynare implementation of the simplex algorithm.
2012-03-04 22:12:58 +01:00
8aea0bf14c
Updated copyright year.
2012-03-04 22:12:58 +01:00
13f1e62de8
Deleted trailing white spaces.
2012-03-04 22:12:58 +01:00
859335b34f
Fixed bug in extended path routine (appearing when some innovations have zero variance).
2012-03-04 22:12:58 +01:00
e882a2ea36
Added unitary test.
2012-03-04 22:09:18 +01:00
192b5c7113
Fixed bugs in routines related to the extended path algorithm.
2012-03-04 22:09:18 +01:00
0aaa7a8a13
Fixed bug in internal tests of local_state_space_iteration_2 routine.
2012-03-04 22:07:27 +01:00
100ff94644
Fixed bug related to the path to the testes routine.
2012-03-04 22:07:27 +01:00
40329e3e29
Added routine to compute one step ahead state space iteration (mex and m). The state space
...
equations are approximated at order two around the deterministic steady state.
2012-03-04 22:07:27 +01:00
8f95711445
Fixed texinfo header. Added two internal tests.
2012-03-04 22:07:27 +01:00
f9671be3d2
Deleted trailing white space.
2012-03-04 22:07:27 +01:00
939318a380
Added a comment about the timming tests done in the third section of the internal tests.
2012-03-04 22:07:27 +01:00
0a9daa519a
Updated copyright year.
2012-03-04 22:07:27 +01:00
78fe29eeef
Cosmetic changes in texinfo headers. Deleted trailing white spaces. Changed the value of show_time in the third internal test.
2012-03-04 22:07:27 +01:00
2a88e49f9f
Cosmetic changes in texinfo header. Deleted trailing spaces.
2012-03-04 22:07:26 +01:00
ce40f0ebfc
Add CMAES (Ticket #245 ) for mode optimization
2012-03-04 22:07:26 +01:00
bad8746e77
corrected bugs in extended path. The code works now with
...
./tests/ep/linear.mod
2012-03-04 20:59:57 +01:00
2234cdc5bd
g0 has to be initialized
2012-02-17 10:41:42 +01:00
441e58cd01
Need to evaluate the recursive blocks to compute the temporary terms
2012-02-17 10:38:39 +01:00
09a3f3a084
Updates M_params only if estim_params_ has a field param_vals (not the case for a kalman_smoother)
2012-02-17 10:36:21 +01:00
e184c1f368
ghx matrix has to be stored as soon as the block contains lagged or mixed variables
2012-02-17 10:24:31 +01:00
c848b3337a
added nodisplay and graph_format options and applied to prior plots and graphs in estimstion routine. Tex tables also splitted when more than 50 params are estimated.
...
nograph: no figure is done
nodisplay: plots are done behind the scenes
graph_format: default only eps figs are saved. it accepts cell vector like {'eps','pdf','fig'}
2012-02-13 11:58:28 +01:00
f09c10e2d3
Merge remote-tracking branch 'ratto/master'
2012-02-13 11:15:24 +01:00
22e76c8ac8
correcting dimension bugs in solve_stochastic_perfect_foresight_model;
...
doesn't work yet; now Jacobian is singular
2012-02-11 15:42:45 +01:00
3564eca525
new version of solve_stochastic_perfect_foresight_model; doesn't work yet
2012-02-11 14:59:51 +01:00
d5628a277e
Fixed i_cols iterations.
2012-02-11 00:01:22 +01:00
e2b7848f6c
security checks before removing remote folders
2012-02-10 22:44:09 +01:00
1129e29ceb
bug fix for octave
2012-02-10 22:43:11 +01:00
31feb32bd3
bug fix for remote unix called by windows master
2012-02-10 22:42:29 +01:00
ff5a802f65
allow the possibility to close all slaves from a condition found in one single thread (e.g. in a while loop where all threads look for the verification of a condition in a loop)
2012-02-10 22:41:54 +01:00
77e84c5328
removed left keyboard line + cosmetics.
2012-02-10 22:38:41 +01:00
060dc74cf5
Fixed computation of the expectations (it==2 case).
2012-02-10 19:05:50 +01:00
ad2f9574e5
Copy of solve_perfect_foresight_model.m --> stochastic version (with ,gaussian quadrature).
2012-02-10 19:05:50 +01:00
6c9eeec7e5
Removed codes related to stochastic extended path.
2012-02-10 19:05:50 +01:00
ec022385c0
added info about each parameter for the sequence of univariate iterations.
...
manually cherry picked from 4.2. branch:
1a5a294c5f93d5368052095e138f852aab1d3177
2012-02-10 17:37:02 +01:00
44b03e5f19
modifying extended-path for parallel toolbox
2012-02-07 16:31:57 +01:00
bc6042fbe2
Improved display of bivariate projections when pprior=0,
...
specially useful when neighborhood_width is used
2012-02-06 16:01:09 +01:00
9dfc0d5c08
Added calls to matlab implementation of the perfect foresight solver in homotopic routine (extended path approach).
2012-02-04 18:52:03 +01:00
87f4dad51a
Changed default value of use_bytecode option. By default, the bytecode solver is not used.
...
Note that, even for small models, there is a (very) substantial gain in using the use_dll option.
2012-02-04 16:56:09 +01:00
704b0c9659
Added an option specifying if the bytecode solver has to be used first.
2012-02-04 16:26:22 +01:00
8a35ee7363
Added an option to skip the test on the number of periods over which the perfect foresight models are solved.
2012-02-04 16:19:15 +01:00
3f2d2b3497
Oups! I forgot to add this routine in commit ac54f7f528. This routine provides a matlab implementation of the perfect foresight model solver.
2012-02-03 15:50:45 +01:00
43f46f2886
Removed stability test over the last periods of the perfect foresight solution.
2012-02-03 14:06:35 +01:00
64ebd1d0d7
Changed the default value of options_.ep.fp. Test the stability of the solution only for the first period (when the value of periods is increased).
2012-02-03 14:05:06 +01:00
7978070814
Fixed bug.
2012-02-03 12:52:37 +01:00
fed2e229e5
Try first Ferhat's code and if it fails try the matlab's implementation of the the perfect foresight model solver.
2012-02-03 12:39:48 +01:00
ac54f7f528
Added comments and the possibility to use a matlab implementation of the perfect foresight model solver.
2012-02-03 11:25:03 +01:00
dd1a48c7c7
MS-SBVAR: support non 4 digit years (fix thanks to Margarita Zabelina)
2012-02-02 12:30:50 +01:00
837451c724
MS-SBVAR: support annual data (fix thanks to Margarita Zabelina)
2012-02-02 12:30:50 +01:00
e4546ba32f
change file format to unix
2012-02-02 12:30:49 +01:00
f44db9370d
fixing problem with Ramsey policy and auxiliary variables
2012-02-02 12:04:41 +01:00
4ac4fdb2f0
matlab: add missing error code
2012-01-30 18:04:24 +01:00
Michel Juillard
Houtan Bastani
Sébastien Villemot
Marco Ratto
Stéphane Adjemian (Charybdis)
Ferhat Mihoubi
Johannes Pfeifer