Commit Graph

4206 Commits (a9292c5086d04ec0b681f88e665421ad09f266b5)

Author SHA1 Message Date
Johannes Pfeifer a9292c5086 Add warning if prior allows for negative variances and if negative estimated variances are encountered. Closes #522
Due to the use of variances for Sigma_e and subsequently backing out the standard deviation from these variances, the sign of the standard deviation does not matter and no bound needs to be imposed.
2014-02-04 18:59:28 +01:00
Sébastien Villemot a5be9dd845 Add trust-region nonlinear solver.
Closes #260
2014-02-04 17:56:27 +01:00
Sébastien Villemot 3fe5a728c6 Remove bad_cond_flag argument from solve1.
This flag was not grounded on any solid theoretical foundation.

This commit actually makes solve_algo=2 to be exactly the same than solve_algo=4.
2014-02-04 17:56:27 +01:00
Sébastien Villemot 0bcc628ba3 Add missing semicolon. 2014-02-04 17:56:27 +01:00
Houtan Bastani 150e265cd2 reporting: check passed values 2014-02-04 17:03:25 +01:00
Sébastien Villemot 3348481104 Add new options consider_all_endogenous and consider_only_observed.
Closes #336
2014-02-04 14:49:06 +01:00
Sébastien Villemot 0df2ba7d12 Fix typo. 2014-02-04 14:38:11 +01:00
Sébastien Villemot 5862099416 Merge pull request #607 from JohannesPfeifer/unique_varlist
Unique varlist and choosing endogenous variables for estimation
2014-02-04 05:14:05 -08:00
Houtan Bastani 4dd462e2bb fix bug introduced in 4caa36232b 2014-02-03 15:53:16 +01:00
Sébastien Villemot 5308da2cad Upgrade the minimum requirement to MATLAB 7.5 (R2007b). 2014-02-03 15:52:54 +01:00
Sébastien Villemot 6d1ee23437 Fix test for availability of xlswrite. 2014-02-03 15:52:54 +01:00
Houtan Bastani 4caa36232b reporting: fix assertion 2014-02-03 15:49:29 +01:00
Stéphane Adjemian (Scylla) 21e9b22d7c Merge branch 'get-prior-info-fixes' 2014-02-03 14:25:53 +01:00
Johannes Pfeifer 6962bc8171 Convert use of unique to be compatible with older Matlab version 2014-02-03 13:19:11 +01:00
Johannes Pfeifer 85815ca651 Adds new option to select which variables to consider for estimation
First step for ticket #336
2014-02-03 13:10:31 +01:00
Stéphane Adjemian (Scylla) 2ee11fa860 Changed the handling of optimization options in dynare_estimation_1.m. Removed calls to strsplit. Closes #605. 2014-02-03 12:25:29 +01:00
Stéphane Adjemian (Scylla) 4b4de4102a Added new routine that converts a string of Key-Value pairs in a cell. 2014-02-03 12:25:29 +01:00
Johannes Pfeifer c6cd5b40d9 Consider only unique members of varlist in estimation 2014-02-03 10:03:17 +01:00
Johannes Pfeifer e0ed06e608 Consider only unique variables in stoch_simul 2014-02-03 09:54:27 +01:00
Michel Juillard c1e0d68351 removed useless test with exist() that is very expensive in
Octave (but apparently not in Matlab)
2014-02-01 16:26:18 +01:00
Stéphane Adjemian (Scylla) 63c289adba If us and ts are dseries objects, chain(ts,us) and ts.chain(us) return the same dseries object. 2014-02-01 11:38:10 +01:00
Stéphane Adjemian (Scylla) 90b47d2704 Added chain method in dseries class. 2014-02-01 11:38:10 +01:00
Sébastien Villemot 9d36a326f1 Merge pull request #603 from rattoma/master
HP-filtered moments: Fix bug when unit root models provide NaN's or Inf's in g_omega
2014-01-31 09:09:29 -08:00
Stéphane Adjemian (Scylla) 7864e0ec82 Rewrote the error message issued when the prefilter option in used while the steady state of the observed variables is non zero (introduced in commit d82252e805). 2014-01-30 17:57:36 +01:00
Stéphane Adjemian (Scylla) f1e4ca2f48 Test if the data are positive before applying the log. 2014-01-30 17:44:47 +01:00
Stéphane Adjemian (Scylla) 8829baa3aa Merge branch 'johannes-documentation' 2014-01-30 15:33:28 +01:00
Stéphane Adjemian (Scylla) 00cfec8fb6 Issue an error message if loglinear option is equal to 1 in non linear likelihood routine. 2014-01-30 14:39:06 +01:00
Stéphane Adjemian (Scylla) 2e1ad9c51b Merge branch 'loglinear' 2014-01-30 13:06:43 +01:00
Stéphane Adjemian (Scylla) 62aa9b8fa5 Check if the steady state is strictly positive when simulating/estimating a model with loglinear option. 2014-01-30 13:03:03 +01:00
Stéphane Adjemian (Scylla) 8359b90cdf Cosmetic changes. 2014-01-30 13:00:25 +01:00
Stéphane Adjemian (Scylla) 8309ed5c96 Added a routine to test if we are currently estimating a model. 2014-01-30 12:48:48 +01:00
Marco Ratto b33da9a40d Fix bug when unit root models provide NaN's or Inf's in g_omega 2014-01-30 10:38:46 +01:00
Johannes Pfeifer d82252e805 Filter out inconsistent specification of prefiltering when observation equation is not mean zero 2014-01-30 10:07:41 +01:00
Johannes Pfeifer 8ef4e45eaf More explicit commenting of stoch_simul.m related to loglinear option 2014-01-30 09:09:25 +01:00
Johannes Pfeifer 4bd7f4952a Make sure that estimation works after loglinear option in stoch_simul with initval-block
stoch_simul logs the steady state. Here, it is transformed back
2014-01-30 08:53:58 +01:00
Johannes Pfeifer 15f0353d02 Make sure repeated runs of stoch_simul.m with loglinear option do not crash due to logged steady state 2014-01-29 19:52:20 +01:00
Johannes Pfeifer 3da8e92aa5 Fix option loglinear for stoch_simul.m by also logging steady state.
Must be done to assure correct simulations and output of moments. Cannot be done in the solver itself as estimation performs logging of steady state also outside of solver
2014-01-29 19:39:00 +01:00
Johannes Pfeifer dfd9e3819b Filter out use of loglinear with simulated moments
Does not work because steady state is not logged
2014-01-29 19:20:52 +01:00
Stéphane Adjemian (Scylla) fa6e97a929 Fixed copyright year. 2014-01-29 17:22:44 +01:00
Stéphane Adjemian (Scylla) b93eab8bf2 Added overloaded cumprod function (dseries class). 2014-01-29 17:22:10 +01:00
Stéphane Adjemian (Scylla) 4254da0a00 Fixed bug. Closes #590. 2014-01-29 13:51:34 +01:00
Stéphane Adjemian (Scylla) 5f46f983c1 Fixed bug. Closes #591. 2014-01-29 12:57:55 +01:00
Stéphane Adjemian (Scylla) b9c4f7d8d1 Deactivate the rescaling of the perturbation in numgrad3_ and numgrad5_ routines. 2014-01-29 12:36:34 +01:00
Johannes Pfeifer 9d4adbcf72 Merge branch 'master' of https://github.com/DynareTeam/dynare 2014-01-28 18:46:27 +01:00
Johannes Pfeifer 16b7a9ed98 Change get_prior_info.m to reflect user-specified bounds. Also corrects LaTeX-code 2014-01-28 18:45:40 +01:00
Johannes Pfeifer c2345fd23b Change minus_logged_prior_density.m to reflect new treatment of estimated and calibrated covariances 2014-01-28 18:42:47 +01:00
Johannes Pfeifer d5446e734a Add comments to make clear distinction between generalized distribution and user-imposed bound 2014-01-28 18:41:07 +01:00
Johannes Pfeifer 0ec62c6360 Fix Latex Bug in display_estimation_results_table.m
Missing line break resulted in compilation errors
2014-01-28 18:39:51 +01:00
Sébastien Villemot 66c1aa78f0 Move add_auxiliary_variables_to_steadystate to partial_information directory.
This function is obsolete, but the PI solver still uses it, so make that
relationship explicit.
2014-01-28 17:48:08 +01:00
Johannes Pfeifer 9e99eed6da Fix sample on which Geweke convergence diagnostics is computed
Was only correct for mh_drop of 0.5
2014-01-27 10:34:13 +01:00