ae921f6bf1
Merge remote-tracking branch 'rattoma/shock_decomp'
2017-03-24 16:20:47 +01:00
b5262ca1e8
Merge remote-tracking branch 'rattoma/deterministic'
2017-03-24 16:20:01 +01:00
e103462050
Merge remote-tracking branch 'rattoma/gsa'
2017-03-24 16:19:44 +01:00
732aa8c30c
Merge remote-tracking branch 'rattoma/slice'
2017-03-24 16:19:04 +01:00
598422931b
reporting: submodule update
2017-03-24 16:16:48 +01:00
b3ecc8c003
preprocessor: add plot_shock_decomposition command. closes #1406
2017-03-24 12:11:36 +01:00
b8b6228d67
graph_decomp_detail: fix bug, extra argument not referenced
2017-03-24 11:24:14 +01:00
93130ee922
fix bug in WriteShockDecomp2Excel
2017-03-24 10:34:59 +01:00
c467c92d29
shock_decomposition: move fig_names to fig_name (fixing typo)
2017-03-24 10:34:59 +01:00
e5f1d20221
change dyn_figure to take a nodisplay and graph_format arguments as opposed to options_. closes #1415
2017-03-23 17:59:05 +01:00
2aa85f978a
change dyn_figure to take a nodisplay argument as opposed to options_. #1415
2017-03-23 14:30:16 +01:00
afcb8eab35
catch M_ output argument from evaluate_smoother;
...
remove unused variable lead_lag_incidence
2017-03-23 08:25:14 +01:00
234aaacf92
factorized display of critical values.
2017-03-23 00:08:05 +01:00
c60379ae96
upgrade call to evaluate_smoother
2017-03-22 23:32:19 +01:00
1db86865f7
better trap missing input arguments
2017-03-22 23:24:41 +01:00
99dbc8c74d
bug fix with redundant call to prior_bounds
2017-03-22 23:08:18 +01:00
b77791c541
ensure invhess is empty with slice for proper prior initialization of chains
2017-03-22 23:06:19 +01:00
780c1f5216
provide userdata with info on actual varlist used
2017-03-22 22:37:59 +01:00
a9dd046db7
fill in end of sample conditional shock decompositions.
2017-03-22 22:30:14 +01:00
76097a799f
Added missing struct2array routine.
2017-03-22 15:20:13 +01:00
ee2c30fd97
Partialm cherry-pick of 28e95d817a181f758dab595d6a14de0ff348bed5
...
Give more informations in case of NaN, Inf or non real variables.
Closes PR #1385 .
2017-03-22 14:04:35 +01:00
85fe69b265
Cosmetic change (use logicals).
2017-03-20 11:39:22 +01:00
9abac8ccdb
Store indicator whether smoother was used with loglinear
...
Required to keep track of whether estimates were logged. Related to #1407
2017-03-20 11:36:10 +01:00
5424fde74e
metropolis_draw.m: Replace round by ceil to avoid crashes when 0 draws would be used
...
(cherry picked from commit 2dc3ad22fdb8fedee7fdd2ed3577978bd8f37324)
2017-03-20 10:19:14 +01:00
94ef60556e
Account for loglinear option when reading in M_.endo_histval
...
(cherry picked from commit d088420c21eba397459015b6a957c7eff7eafc36)
2017-03-20 09:37:58 +01:00
6e268c34d4
Cosmetic fixes to dynare_estimation_init.m
...
(cherry picked from commit 7ea24165c0a26393753c58170f74fab3f7076b3e)
2017-03-20 09:27:04 +01:00
9427b58bfd
add parameter names to mean file, to ensure same behavior of standard mode files.
2017-03-19 12:19:24 +01:00
389123d0b5
bug fixes.
2017-03-19 12:19:24 +01:00
121333d913
currently, the mode_file specified by the user is re-set to empty. This re-sets mode_file to the user defined value.
2017-03-19 12:19:24 +01:00
b916d788f4
trap cases where both point and posterior estimates are available
2017-03-19 12:19:24 +01:00
a869f5a2c4
Updated m-unit-tests submodule.
2017-03-19 12:15:22 +01:00
2b28ba70d9
Make remote parallel computation compatible with new steadystate and auxiliary variable files
2017-03-18 17:28:44 +01:00
6c21d44968
Fixed encoding.
...
For a non identified an enconding issue was leading scatter_plots function to crash...
2017-03-18 16:28:30 +01:00
d1550293bf
bug fix for the rare case where r2 does not change for some observable
2017-03-18 10:27:36 +01:00
576b2d0f85
end-of-line fix
2017-03-18 10:27:36 +01:00
0845aa8430
Fixed header and license of GSA.
2017-03-18 10:27:36 +01:00
a40a854a59
Set copyright for pick.m
2017-03-18 10:27:36 +01:00
0f07d17d8c
Introduced scatter plot analysis of output results, borrowing from by GLUWIN toolbox.
2017-03-18 10:27:36 +01:00
a47f6fc2ca
- save also info on selected endo names
...
- bug fix for one single endo variable
- when nograph, no older plot is deleted
2017-03-18 10:27:36 +01:00
0752d6a8e6
Trap incompatible user define options in a more transparent way. Fixes #1016 .
2017-03-18 10:27:36 +01:00
a5a086040b
Bug fixes for latex and to new smooth* files saved by posterior sampler
2017-03-18 10:27:36 +01:00
e71ed1bad6
Partially revert f70f45d1271e716498a8761a8e8476f54da3607d. Do not discard repetitions.
2017-03-18 00:06:15 +01:00
c02de3c652
fixed text to error message
2017-03-18 00:06:15 +01:00
e180e2f075
1) fixed while for b index
...
2) ensure there are no repeated draws in sub-draws;
3) speed-up subdraw sampling when all retained draws are used.
2017-03-18 00:06:15 +01:00
20b0fc8bac
update mh_mode with slice and load_mh_file
2017-03-18 00:06:15 +01:00
1e3e2f70a9
set posterior sampler options when load_mh_file
2017-03-18 00:06:15 +01:00
49b45a3ba7
sampler options to be set also when load_mh_file = 1
2017-03-18 00:06:15 +01:00
fd8073720e
For very short chains, maximize the max number of sub_draws across chains
2017-03-18 00:06:15 +01:00
bc6c6fc486
Removed remaining calls or references to cputime.
2017-03-17 22:49:01 +01:00
a68e958cf7
As agreed in #1375 , moved treatment of filtered_vars in dynare_estimation_init.m
2017-03-17 15:43:00 +01:00
4f93c3df15
proper use of varlist in smoothed variables + sort in alphabetic order to find more easily plots for large numbers of variables.
2017-03-17 15:43:00 +01:00
4b2d139c7a
- 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.
2017-03-17 15:43:00 +01:00
510bbe200e
Bug fix (wrong number of output arguments when calling evaluate_smoother).
2017-03-17 15:09:05 +01:00
0adb065b7e
Bug fix (wrong count of the number of output arguments).
2017-03-17 14:21:53 +01:00
5d8d5e90d8
evaluate_smoother.m: pass updated M_.params and oo_ back to base workspace
...
Closes #1373
2017-03-17 13:15:39 +01:00
b6a8887da3
Bug fix (missing declaration of freq if initial_date is empty).
2017-03-17 12:42:00 +01:00
6b64c47ec2
Bug fix with stored shock_decomp options
2017-03-17 12:41:59 +01:00
ab21c383a6
Added check for change in options_.nobs, for rolling calls to smoother
...
Use select_qz_criterium_value
to fix all cases where the default
options_.qz_criterium = [];
is passed to evaluate_smoother and dataset_ is already set.
qz_criterium also set to fix evaluate_likelihood.m for similar issues
2017-03-17 12:41:59 +01:00
19cc38b39e
allow to start from quarterly growth rate variables
2017-03-17 12:41:59 +01:00
ce6f76713d
cosmetic change
2017-03-17 12:41:59 +01:00
7a6b4f3464
extra option about choosing the time periods for which to plot decompositions
2017-03-17 12:41:59 +01:00
1920fb0bfe
fixed timing bugs in annualized shock decomposition
2017-03-17 12:41:59 +01:00
d7723877cf
bug fixes for annualized nominal variables
2017-03-17 12:41:59 +01:00
ac43beb801
bug fix with 'others' in conditional decomposition
2017-03-17 12:41:59 +01:00
9005054c6f
optional output argument for plot_shock_decomposition.m + several fixes to annualized computations and plots.
2017-03-17 12:41:59 +01:00
39c3cf7718
bug fix with realtime conditional shock decomp
2017-03-17 12:41:59 +01:00
a8107e154a
added new annual aggregation types, for consistent quantity/price aggregations, if required.
2017-03-17 12:41:59 +01:00
cb3362b89d
missing utility for xls export of shock decomps.
2017-03-17 12:41:59 +01:00
5a51e9cc78
fixed zero line in all cases
2017-03-17 12:41:59 +01:00
a50a4f458a
Bug fix: all shock decomp options need to be stored in mydata for uimenu's.
2017-03-17 12:41:59 +01:00
e21c887775
bug fix with steady state option
2017-03-17 12:41:59 +01:00
61d55127ed
Fixed bug with naming of screen option
2017-03-17 12:41:59 +01:00
e7442bcf78
added geometric average.
...
Set neutral type names, replacing stock/flow/deflator.
2017-03-17 12:41:59 +01:00
46d6e59c9a
small fix to file names
2017-03-17 12:41:59 +01:00
bfe6d04bf2
- Added the possibility to compute annualized variables and associated decompositions out of quarterly ones, without blowing up variable definitions
...
- Fixed call to plot realtime conditional shock decomposition
2017-03-17 12:41:59 +01:00
bb4ecbf5a8
Realtime conditional forecast is better computed as residual from the realtime shock decomposition and the model unconditional forecast
2017-03-17 12:41:59 +01:00
a7ae355e8f
Fixed dating issues in some plot shock decomps
2017-03-17 12:41:59 +01:00
441d34a6ee
Fixed bug with realtime conditional shock decompo
2017-03-17 12:41:59 +01:00
f87e240b30
Some better diversification of fig names and captions
2017-03-17 12:41:59 +01:00
92f79a14f4
Added realtime conditional forecast shock decomposition
2017-03-17 12:41:59 +01:00
3bebeb1832
Legend for detailed plots
2017-03-17 12:41:59 +01:00
3fdf7be6e8
- removed useless internal option fig_mode
...
- better captions/figure nsmes
- moved plots to graphs subdir
- fixed latex issuea [label and caption].
2017-03-17 12:41:59 +01:00
10773e2cbb
results of realtime_shock_decomposition.m should be plotted using only plot_shock_decomposition.m, so avoid unexpectable behavior.
2017-03-17 12:41:59 +01:00
d304fd26c1
shock_decomposition.m has to rely on plot_shock_decomposition.m, to make behaviour robust.
2017-03-17 12:41:59 +01:00
c2d2c46edf
fixed figure name
2017-03-17 12:41:59 +01:00
02b3e20377
allow uimenu's also for colored graphs + better organization of userdata info
2017-03-17 12:41:59 +01:00
4ff586247e
bug fix in Others group label
2017-03-17 12:41:59 +01:00
2c54d4432a
- Bug fix with screen_shocks option
...
- fix to figure title
2017-03-17 12:41:59 +01:00
ad55e1b085
initialize screen_shocks option of shock decomp
2017-03-17 12:41:59 +01:00
9e73a9e86b
add steady state line
2017-03-17 12:41:59 +01:00
94297b71cc
bug fixes around plot_shock_decomposition.m utilities
2017-03-17 12:41:59 +01:00
0bf30889f9
Fixed header + bug fix for default options + set only local nograph
2017-03-17 12:41:59 +01:00
e3c825f01a
fixed header
2017-03-17 12:41:59 +01:00
73e5650dcc
Revert "shock_decomposition should also follow global options_.nograph"
...
This reverts commit 01971dbeb032642421f60b468517fbbe0995a222.
2017-03-17 12:41:59 +01:00
ba7802204f
end of lines
2017-03-17 12:41:59 +01:00
d2d6e19d87
fixes around default options in shock decompositions
2017-03-17 12:41:59 +01:00
7733484912
New utility plot_shock_decomposition.m
...
new graph_decomp_detail.m for detailed plots of shock contributions
new expand_group.m for optional uimenu operation in case of grouped shocks.
2017-03-17 12:41:59 +01:00
3ed94b2e46
New realtime_shock_decomposition.m routine + provisions in global initialization
2017-03-17 12:41:59 +01:00
4a314f7f8a
shock_decomposition should also follow global options_.nograph
2017-03-17 12:41:59 +01:00
a092cc3b73
Set default graph options and extend graph_decomp.m accordingly.
2017-03-17 12:41:59 +01:00
e174d04579
- New option init_state [options_.shock_decomp.init_state] to compute the shock decomposition conditional to the state variables in period 1;
...
- force plot_priors=0
- when grouping shocks, include the residual group Others, since user defined groups may not add up to the smoothed variable.
2017-03-17 12:41:59 +01:00
f27f210654
New utility for changing M_.params only locally
2017-03-17 12:41:59 +01:00
a6f7e02d17
added utilities for optimal distinguishable colors.
2017-03-17 12:41:59 +01:00
06326d4477
Fixed bug introduced in cad7adc01f.
2017-03-16 18:14:23 +01:00
dbffe2a841
Do not check boundaries if the model is not estimated.
2017-03-16 15:50:44 +01:00
cc34e16458
Set temporarily mode_compute and mh_replic to zero if estim_params_ is empty.
2017-03-16 15:45:08 +01:00
cad7adc01f
Added routine for testing if we are currently estimating a model.
2017-03-16 15:03:46 +01:00
5f5c093b2e
Renamed isestimation function + Cosmetic changes.
2017-03-16 15:00:02 +01:00
d0ab90275b
Merge pull request #1402 from JohannesPfeifer/prior_asymptote
...
Add warning if prior has an asymptote
2017-03-15 16:33:10 +01:00
1ff816d7bf
Updates particles submodule.
2017-03-15 16:30:29 +01:00
4f3b35acba
Account for Matlab 2017a.
...
(cherry picked from commit f3a1619040fbb831830791424015d3d3bf8bdbae)
2017-03-13 23:42:48 +01:00
9f7b084970
Merge branch 'master' into json
2017-03-10 17:20:15 +01:00
87fc09863d
Add warning if prior has an asymptote
...
Can happen with beta and gamma distribution. Closes #1401
2017-03-10 15:16:33 +01:00
56edaf5cb0
Fix check for existence of dname variable
2017-03-07 13:55:29 +01:00
5063020c5d
Merge branch 'master' into json
2017-03-03 14:42:25 +01:00
ed3e74994f
preprocessor: add onlyjson option to allow exit upon writing of JSON output. #1387
2017-02-28 14:34:22 +01:00
927b813ed8
Removed printing of debugging info.
...
Introduced in ffebf6b262
2017-02-15 13:35:33 +01:00
ffebf6b262
Check positive steady state for loglinear option...
...
1e-9 is too far from zero. Set 1e-16 as a threshold value for the positivity test.
2017-02-15 12:41:56 +01:00
9fbef0c107
Removed penalty_hessian routine.
...
+ Code factorization.
+ Added an option for using the penalized objective when computing numerically
the hessian at the mode.
Previous behaviour (introduced with penalty_hessian routine) was to compute the
hessian matrix at the mode with the penalized objective function (instead of
the original objective function). This behaviour hides problematic situations,
where the computed hessian (using the original objective) would not be full
rank. For instance, if the estimation ends up with a parameter on (or very
close to) the bounds of its possible values (which is often not a desirable
outcome), the estimated posterior variance would be zero for this
parameter (with the original objective) because the hessian is not finite in
this direction, while the posterior variance would be positive if the penalized
objective is used instead. But this estimate would not be reliable by
construction of the penalty which is quite ad-hoc (more fundamentally I do not
think that there exists any rational for approximating the covariance matrix
with the inverse of the hessian matrix if the mode is on the border of the set
of possible values).
This commit restore the behaviour previous to 2446ab02ba
2017-02-08 13:07:25 +00:00
5b73d7d59c
Added new optimization routine (particleswarm).
...
Only available under Matlab if the Global Optimization Toolbox is installed.
2017-02-08 13:07:25 +00:00
35078e13ee
Completed header + cosmetic changes.
2017-02-08 13:07:25 +00:00
8d4abe4a9f
Cosmetic changes.
2017-02-08 13:07:24 +00:00
e71e89bb38
Added check on Octave version.
2017-02-08 13:07:24 +00:00
d05dd34d30
preprocessor: write JSON output. #1387
2017-02-06 17:41:18 +01:00
62c954ff12
Fixes #1287 .
2017-01-25 22:13:36 +01:00
0b1b873160
Updated dseries submodule.
...
Fixes remaining Octave failing unit tests.
2017-01-25 09:18:28 +01:00
875497bc71
Unit test lpdfgweibull.m: adjust upper integration bound under Octave
...
Does not correctly integrate otherwise as more than 99% of support is 0
2017-01-18 11:01:13 +01:00
ce4ab7f7d8
Unit test wblinv.m under Octave: adjust tolerance
2017-01-18 10:59:53 +01:00
12b8a915e9
submodule update: reporting
2017-01-09 16:20:15 +01:00
ba9af24a67
Merge pull request #1370 from JohannesPfeifer/updated_variables
...
Output UpdatedVariables in oo_ without filtered_variables
2017-01-05 11:50:06 +01:00
4dfec0d7fb
Merge pull request #1369 from JohannesPfeifer/disp_dr_aux_var
...
disp_dr: Add case for aux_vars.type==0
2017-01-05 10:44:44 +01:00
77a93d6ff9
Merge pull request #1362 from JohannesPfeifer/MCMC_jumping_covariance
...
Moving loading of MCMC_jumping_covariance to after display of standar…
2017-01-05 09:02:59 +01:00
b464226eaa
Merge pull request #1358 from JohannesPfeifer/nograph
...
Implement posterior_nograph option
2017-01-05 06:52:08 +01:00
90982d5fa0
Merge pull request #1357 from rattoma/parallel
...
Provisions for new node option NumberOfThreadsPerJob. Document option
2017-01-04 23:29:24 +01:00
c893c0620c
Merge branch 'pr#1354'
2017-01-04 15:12:32 +01:00
2c5b1fed2d
Use master function lyapunov_solver.m to call individual solvers
2017-01-04 11:25:44 +01:00
b946d3c082
Merge pull request #1353 from JohannesPfeifer/Plot_posterior
...
PlotPosteriorDistributions.m: remove eval commands and fix typo in "m…
2017-01-04 11:20:06 +01:00
e815c2b4db
Output UpdatedVariables in oo_ without filtered_variables
...
Makes it consistent with manual and closes #1366
2017-01-04 10:57:00 +01:00
515e080f49
Do not compute static residuals if steady state routine failed.
...
Avoids cryptic model dependent error message.
2017-01-04 10:26:22 +01:00
65c402469c
disp_dr: Add case for aux_vars.type==0
...
Closes #1367
2017-01-04 09:14:01 +01:00
3c892170ef
Fixed bug with getting CPU number on mac [see #1272 #838 ]
2017-01-03 19:24:18 +01:00
ee7611c8ad
added note to commented text, as an alternate option, should bugs/problems be reported with line 60
2017-01-03 19:15:53 +01:00
7b4fc9ec4a
Moving loading of MCMC_jumping_covariance to after display of standard errors and computation of Laplace approximation
...
Closes #1255
2016-12-27 12:07:22 +01:00
d8b0495c77
Make PosteriorIRF.m and pm3.m honor nograph and posterior_nograph option
2016-12-26 13:11:01 +01:00
452d3cd986
Implement preprocessor interface for posterior_nograph option
2016-12-26 13:07:46 +01:00
28d4ed1918
Finish implementation and documentation of nograph option of shock_decomposition.m
2016-12-26 12:52:28 +01:00
4ecf1d9f5d
Workaround to avoid popup figures with nodisplay option saved as fig in MATLAB.
2016-12-20 09:14:48 +01:00
a55cbdd5fe
Provisions for new node option NumberOfThreadsPerJob. Document option
2016-12-19 23:28:46 +01:00
352786c2b4
load_mh_file: recompute results by default and load on request
...
Prevents general overwriting of results in oo_ and having stale results
2016-12-19 19:37:34 +01:00
5707b4d965
Make evaluate_smoother.m output former global variables for correct updating of these objects
2016-12-18 12:43:29 +01:00
Houtan Bastani
Marco Ratto
Stéphane Adjemian (Charybdis)
Johannes Pfeifer