Marco Ratto
5fa6265944
Implement new option smoother_redux, to allow fast smoother for very large models. It runs smoother only for state variables and computes the remaining ones ex-post
...
Contains improvements, in order to recover as much as possible static unobserved (filtered, smoothed, updated, k-step ahead), Variance, State_uncertainty, k-step ahead variances trying to map lagged states onto current ones using pinv(T). This has exceptions (namely lagged shocks which are ONLY used to recover static NON observed variables). this exception is also trapped.
For such extensions we can only recover smoothed variables starting from d+1. Variances CANNOT be recovered for such variables (the smoother gives ZERO.)
2021-05-27 16:59:15 +02:00
Johannes Pfeifer
7541b1b116
prior_draw.m: cosmetic fix
2021-05-27 16:46:40 +02:00
Johannes Pfeifer
5d1d5b8a85
prior_posterior_statistics_core.m: filter out cases where model does not solve
...
Relax qz_criterium by factor of 10, when taking posterior subdraws. Also makes sure `prior` option works as intended.
2021-05-27 16:46:28 +02:00
Sébastien Villemot
ae574fff78
Merge branch 'diagnostics' into 'master'
...
model_diagnostics.m: only provide message if imaginary component reaches tolerance level
See merge request Dynare/dynare!1855
2021-05-27 14:23:55 +00:00
Stéphane Adjemian (Charybdis)
3e7cd877d3
Fix error message consistently with 2743ca8fd
.
...
[skip ci]
2021-05-26 19:11:01 +02:00
Stéphane Adjemian (Charybdis)
2743ca8fdb
Fix workspace if dynare is called from a function.
2021-05-26 19:07:44 +02:00
Marcoo Ratto
1645f38269
Implement heteroskedastic filter and smoother
2021-05-26 18:45:16 +02:00
Sébastien Villemot
0d549bd87a
Preprocessor: streamline use_dll compilation under MATLAB
2021-05-25 18:57:40 +02:00
Sébastien Villemot
3bdb07d11e
Build system: remove -lstdc++ when linking MEX for MATLAB under macOS
...
It is unneeded.
2021-05-25 18:57:40 +02:00
Sébastien Villemot
ba3522a9ba
Merge branch 'compute_decision_rules' into 'master'
...
Use of the compute_decision_rules function to factorize code
See merge request Dynare/dynare!1856
2021-05-25 16:18:43 +00:00
Sébastien Villemot
c78e37290a
Merge branch 'ident_opt' into 'master'
...
identification: support optimal policy
See merge request Dynare/dynare!1852
2021-05-25 13:07:06 +00:00
Selma Malmberg
9c6cc04b0d
Use of the compute_decision_rules function to factorize code (Ref.#1173)
2021-05-20 11:36:29 +02:00
Johannes Pfeifer
7e72b52677
model_diagnostics.m: only provide message if imaginary component reaches tolerance level
2021-05-20 11:23:48 +02:00
Johannes Pfeifer
456dc86ad1
identification: support optimal policy
...
Related to https://git.dynare.org/Dynare/dynare/-/merge_requests/1837
2021-05-14 13:38:42 +02:00
Johannes Pfeifer
7d4b01883d
get_identification_jacobians.m: remove unnecessary reordering of equations
2021-05-14 13:35:01 +02:00
Sébastien Villemot
5e36b688ee
Merge branch 'forcst_bug' into 'master'
...
forcst.m: fix case with measurement error
See merge request Dynare/dynare!1850
2021-05-12 17:08:58 +00:00
Johannes Pfeifer
c8bc12ab6c
forcst.m: fix case with measurement error
2021-05-12 17:44:35 +02:00
Sébastien Villemot
fb31842856
Build system: remove -lstdc++ when linking MEX for MATLAB under GNU/Linux
...
It is not needed, and it prevents the -static-libstdc++ flag from work
correctly if supplied by the user.
2021-05-11 22:57:16 +02:00
Sébastien Villemot
cf37b5151a
Testsuite: indicate that glibc bug has been fixed in version 2.34
...
[skip ci]
2021-05-11 18:54:07 +02:00
Sébastien Villemot
0d6bc47158
Merge branch 'rattoma/dynare-mh_initialize_from_previous_mcmc'
...
Ref. !1841
2021-05-11 18:51:57 +02:00
Sébastien Villemot
306e963efa
Preprocessor: add mh_initialize_from_previous_mcmc option family to estimation command
...
Ref. !1841
2021-05-11 18:51:29 +02:00
Marco Ratto
a0f2b2399b
New option mh_initialize_from_previous_mcmc that allows to pick initial values for a new MCMC from a previous one
...
If an additional parameter is estimated, it is initialized using prior_draw.
2021-05-11 18:48:43 +02:00
Sébastien Villemot
d8fc564427
Merge branch 'JohannesPfeifer/dynare-lagged_exo'
...
Ref. !1848
2021-05-11 17:47:35 +02:00
Johannes Pfeifer
cc045caa6f
loglinear: Take care of non-logged lagged/leaded exogenous variables
...
Closes : #1783
2021-05-11 17:47:25 +02:00
Johannes Pfeifer
5a2fbe2f89
dynare.m: remove non-working line feed
2021-05-11 14:24:02 +02:00
Sébastien Villemot
702e22bc51
Preprocessor: the fast_realtime option now accepts a vector of integers
...
Ref. !1842
2021-05-10 15:55:56 +02:00
Sébastien Villemot
bdb743eaa7
Merge branch 'rattoma/dynare-fast_realtime_shock_decomp'
...
Closes !1842
2021-05-10 15:55:28 +02:00
Marco Ratto
2d39991a0a
allow for a more accurate (and more expensive) fast realtime modality, triggered by fast_realtime=INTEGER_VECTOR and where vector fast_realtime controls periods for which the smoother is run. If max(fast_realtime)<nobs, the smoother is also run for the full sample. Manual updated accordingly.
2021-05-10 15:55:03 +02:00
Sébastien Villemot
8475dfe26d
Build system: update m4 macros from autoconf-archive
2021-05-10 14:55:34 +02:00
Sébastien Villemot
0741963ca5
Build system: streamline handling of compilation flags for MEX
...
— allow the user to override compilation flags for MATLAB MEX (it was already
working for the preprocessor, the MEX for Octave and Dynare++)
— increase the symmetry of MEX build infrastructure between MATLAB and Octave
— when linking MEX for Octave, do not add the output of “mkoctfile -p FLIBS”.
It is unneeded, and it can create a conflict between the system compiler and
a user-supplied compiler
By the way:
— restore optimization on macOS for C/C++ MEX (it had been removed in
5df2392a09
)
— remove -fno-omit-frame-pointer on MATLAB/Linux, since it would be cancelled
by subsequent -O2 and should not be needed anyways
— remove FFLAGS under Octave, unused
2021-05-10 14:39:41 +02:00
Sébastien Villemot
bd287ca5fa
Build system: remove 32-bit MATLAB stuff, since we no longer support that configuration
2021-05-09 17:51:38 +02:00
Sébastien Villemot
3c7865a950
Merge branch 'rattoma/dynare-initial_estimation_checks_errors'
...
Ref. !1839
2021-05-07 17:52:45 +02:00
Sébastien Villemot
f4f4047588
Preprocessor: add no_init_estimation_check_first_obs option to estimation
...
Ref. !1839
2021-05-07 17:52:24 +02:00
Marco Ratto
74121c2aee
New no_init_estimation_check_first_obs option that allows to skip check of singularity in first observation
...
(like deficient # of shocks in the presence of stockflow observed)
2021-05-07 17:51:03 +02:00
Sébastien Villemot
088fd64114
Merge branch 'rattoma/dynare-smoother_fixes'
...
Ref. !1847
2021-05-07 16:35:36 +02:00
Marco Ratto
551917581f
Bug fixes in fitered variances of smoother
...
- kalman_algo=1: kstep-ahead variances were WRONG, since Pf was initialized using P in PREVIOUS period
- kalman_algo=2: output argument for filtered varainces should be P1, not P (P are UPDATED variances, there).
For kalman_algo=2, also make a small factorization fix (compute P(:,:,t+1) before defining Pf, so to compute 1-step ahead variance only once)
2021-05-07 16:34:53 +02:00
Marco Ratto
d43a057af3
store_smoother_results: fix header explanation of output arguments
...
Variances are now in declaration order.
2021-05-07 16:34:35 +02:00
Sébastien Villemot
ca43df16e0
Merge branch 'dot_error' into 'master'
...
Filter out "Dot indexing is not supported for variables of this type." error...
See merge request Dynare/dynare!1846
2021-05-06 14:49:39 +00:00
Sébastien Villemot
29c7d1c42e
Preprocessor: new parallel_use_psexec command-line option
...
Ref. !1843
2021-05-06 16:31:54 +02:00
Sébastien Villemot
58db4a4c21
Merge branch 'rattoma/dynare-parallel_without_psexec'
...
Ref. !1843
2021-05-06 16:31:15 +02:00
Marco Ratto
9ed9aa5963
For local clusters only, allow using start instead of psexec with windows, to properly allocate affinity when there are more than 32 cores. Manual updated accordingly.
2021-05-06 16:31:06 +02:00
Johannes Pfeifer
f43ee91ed8
Filter out "Dot indexing is not supported for variables of this type." error if variable conflicts with namespace
2021-05-06 10:59:19 +02:00
Sébastien Villemot
f9fb3bb386
Octave compatibility fix: xlswrite() from the io package does not return a 2nd output
2021-05-05 15:47:14 +02:00
Sébastien Villemot
749fc219e5
Testsuite: increase test tolerance under Octave for ecb/SURGibbs/fulton_fish.mod
2021-05-05 15:43:19 +02:00
Sébastien Villemot
96ad0fbec5
SUR / pooled OLS: workaround for Octave bug in dbstack(N)
...
See https://savannah.gnu.org/bugs/?60531
By the way, also add a safety check in pooled OLS for the case where the
function is directly called from the top-level.
2021-05-05 15:27:15 +02:00
Sébastien Villemot
62e42e08a8
Testsuite: more fixes related to stochastic block decomposition + mfs > 0
...
Ref. #1726
2021-05-05 10:15:52 +02:00
Sébastien Villemot
04ba15e5cc
Preprocessor: forbid check and stoch_simul in block-decomposed mode with mfs > 0
...
Closes : #1726
2021-05-04 17:18:20 +02:00
Sébastien Villemot
62aa799d3e
Manual: matrices ghx and ghu are ordered differently if block option is given
...
By the way, also specify that columns of ghx are ordered according to
oo_.dr.state_var in all cases.
Ref. #1726
2021-05-04 17:17:55 +02:00
Sébastien Villemot
0a7d60e7a6
Stochastic solution under block decomposition: minor simplification
2021-05-04 16:33:15 +02:00
Sébastien Villemot
58b2bc78c3
Testsuite: check that stochastic solution in block decomposition mode is correct
...
Only test with mfs=0, since the result is wrong for other mfs values.
Ref. #1726
2021-05-04 16:31:44 +02:00