369d8dd7a4
Allow prefilter_decomposition as option if calib_smoother
2016-03-23 10:31:10 +01:00
fb20b464d4
Remove bayestopt_.mean_varobs and use dataset_info instead
...
Closes #255
2016-03-23 10:31:09 +01:00
6a3c2b7b44
Cosmetic changes to forcst2.m and forcst2a.m
...
Cleans and documents files
2016-03-23 10:19:42 +01:00
dcf0d75d55
Add output of trend to DsgeSmoother.m and use it to reconstruct smoothed and filtered values
2016-03-23 10:19:41 +01:00
1c816aef24
Various interrelated bugfixes dealing with detrending
2016-03-23 10:19:40 +01:00
d1bbe8f852
Restore backward compatibility with old data interface
2016-03-23 10:10:42 +01:00
f64b76f0d3
Move saving of data mean to data initialization instead of posterior computations
...
Makes it accessible for other functions that also need it
2016-03-23 10:10:41 +01:00
b235040ae1
Fix setting of no-constant option in dynare_estimation_init.m when loglinear option is used
...
Setting was based on unlogged steady states
2016-03-23 10:09:17 +01:00
891b6da31c
Clarify hierarchy of loglinear option's logging of data and prefiltering's demeaning in manual
2016-03-23 10:09:16 +01:00
e37448faaf
Fix exceptions when prefiltering was requested with loglinear option
...
The check for an existing mean was based on the unlogged steady state, resulting in wrong error messages
2016-03-23 10:09:15 +01:00
b8377b8de5
Fix constant part of forecasts with loglinear option
...
Erroneously, the unlogged steady state was added
2016-03-23 10:09:14 +01:00
452dfa6a7b
Fix wrong initialization of trend_coeff in preprocessor/ComputingTasks.cc
2016-03-23 10:06:19 +01:00
a55f00073b
Manual Merge of branch 'mh_recover' (Johannes)
2016-03-22 18:02:45 +01:00
692acc66a7
Merge pull request #1137 from JohannesPfeifer/deterministic_estimation
...
Reset seed before drawing starting points of MCMC chains
2016-03-22 16:48:27 +01:00
01e5b0943b
Merge pull request #1129 from JohannesPfeifer/debug_ramsey
...
Fix number of multipliers added by Ramsey used in evaluate_steady_sta…
2016-03-22 16:33:15 +01:00
61e6e80b5a
Allow irf_shocks in DSGE-VAR models.
2016-03-22 16:04:33 +01:00
a3c7cad354
Merge pull request #1131 from JohannesPfeifer/make_
...
Replace globals by inputs in make_ex_.m and make_y_.m
2016-03-18 15:10:06 +01:00
0ba7dc6d16
Merge pull request #1130 from JohannesPfeifer/dataset_info
...
Save dataset_info in results.mat if it exists
2016-03-18 14:30:09 +01:00
cf9ecff321
Merge pull request #1138 from JohannesPfeifer/ep_varexo_det
...
Provide informative error message if EP is used with varexo_det
2016-03-18 14:28:13 +01:00
8e3cb65258
Provide informative error message if EP is used with varexo_det
...
Closes #875
2016-03-17 10:30:14 +01:00
26c0825b6f
Reset seed before drawing starting points of MCMC chains
...
Removes stochastics between different runs while not affecting the subsequent MCMC computations for which the seed is also reset and saved
Closes #1048
2016-03-16 20:09:24 +01:00
4ca5c21858
Clarify manual on mh_recover
2016-03-16 20:01:07 +01:00
8c8587d89b
Add factored unit tests for mh_recover
2016-03-16 19:04:03 +01:00
a28d405f74
Make mh_recover continue at last existing MC file with correct seed
2016-03-16 19:01:59 +01:00
939ac02bad
fix copyright dates
2016-03-15 12:45:57 +01:00
5a91b601a2
Merge pull request #1134 from JohannesPfeifer/bvar_mdd
...
Save marginal data density from BVAR in oo_
2016-03-15 12:39:10 +01:00
6b31363d10
Save marginal data density from BVAR in oo_
...
Closes #1090
2016-03-15 11:49:25 +01:00
2643bfaa40
Fix compatibility with Flex 2.6.
...
In Flex 2.6, the C++ scanner changed. yyout is now a reference (and no
longer a pointer).
2016-03-14 16:28:24 +01:00
ce6dff1fcc
make gsa consistent with (old) change in use of options_.simul
2016-03-13 10:35:48 +01:00
bb08853784
Replace globals by inputs in make_ex_.m and make_y_.m
2016-03-11 16:28:17 +01:00
f60945facc
fix copyright dates
2016-03-11 16:22:42 +01:00
ff0f360938
doc: clarify/simplify doc for planner_discount
2016-03-11 16:02:01 +01:00
f86ac5ae66
preprocessor: allow for changing planner_discount in .mod file. closes #1092
2016-03-11 15:56:34 +01:00
1b203a3755
Save dataset_info in results.mat if it exists
2016-03-11 15:29:12 +01:00
af5368e83c
Fix number of multipliers added by Ramsey used in evaluate_steady_state.m
2016-03-11 14:55:11 +01:00
127637ffd6
Merge branch 'temporary_terms'
...
Do not evaluate unnecessary temporary variables (for instance those
used in the evaluation of the jacobian matrix when we only compute the
residuals of the dynamic equations).
2016-03-09 17:53:58 +01:00
38b21feb3e
build system: cosmetic typo fix
2016-03-09 16:53:57 +01:00
04989822ac
build system: fix lexinc path by replacing the last instance of 'flex'
2016-03-09 14:03:29 +01:00
e4294ee4fc
Merge pull request #1126 from JohannesPfeifer/mh_nblocks_check
...
Check whether at least one MCM chain is requested when using Bayesian…
2016-03-06 21:56:41 +01:00
ab6063f2b4
Check whether at least one MCM chain is requested when using Bayesian estimation
2016-03-06 21:19:17 +01:00
42ecfa382f
fxing bug in diffuse smoother with missing values
2016-03-06 21:07:50 +01:00
f471105a21
Fixed bug (wrong first input argument when calling create_TeX_loader).
...
Fixes the error reported by the testsuite on ls2003a.mod.
2016-03-05 21:28:16 +01:00
d4195ae624
build system: fix typo
2016-03-08 18:06:45 +01:00
d331cf5a7a
Fixed extended path.
...
- Removed call to make_ex_,
- Fill oo_.exo_simul in extended path routine,
- Do not update oo_.exo_simul after the call to the extended path routine,
- Cosmetic change.
(cherry picked from commit 4791649524cc7876fc25d04a925f58a546a3a67d)
2016-03-05 16:24:14 +01:00
066c79f46a
test case for previous commit
2016-03-07 21:51:45 +01:00
34e35555d6
fixed bugs when there is a single equation in the model
2016-03-07 21:51:45 +01:00
08375e7ba6
Fixed wrong ordering of the hyperparameters when walling the routine for drawing random deviates from the Weibull distribution.
2016-03-04 16:45:06 +01:00
472f6c4a1c
Fixed typo.
2016-02-29 15:46:40 +01:00
837c812c17
Efficiency changes.
...
+ Use scalar product when possible, ie replace sum(x.*x) by x'*x,
+ Add parenthesis around scalar products, ie replace a*x'*x by a*(x'*x),
+ Removed global options_ in lnsrch1 (only used for solve_tolx).
Note that the scalar products could be replaced by a loop for small
problems. If the number of unknow is smaller than 70 the loop is
faster (10 times faster if the number of unknowns is less than 10).
2016-02-26 22:06:49 +01:00
87955c61d0
Fixed bug in the routine simulating backward looking non linear stochastic models.
2016-02-26 21:31:50 +01:00
Johannes Pfeifer
Stéphane Adjemian(Charybdis)
Johannes Pfeifer
Houtan Bastani
Sébastien Villemot
Michel Juillard