GitLab

We already used a variable called offset

Johannes Pfeifer authored Nov 15, 2018 and Sébastien Villemot committed Nov 16, 2018

1. Rely on actually computed standard deviations in bayestopt_ instead of potentially unset estim_params_
2. Remove arbitrary normalizations/omitted normalization in case of division by 0 in normalization
3. Distinguish between 0 identification and division by 0 due to normalization in plots

(cherry picked from commit 7341e21a)

- M and MEX files are now under +${MODELNAME}/
- bytecode, C source and JSON now under ${MODELNAME}/model/

Johannes Pfeifer authored Jan 31, 2018 and Stéphane Adjemian committed Feb 01, 2018

https://github.com/DynareTeam/dynare/commit/99dbc8c74d0ecf4c7a610c4a394fd85028cf2741 introduced a bug by disabling the resetting of the prior bounds used for the MCMC. The manual clearly states the bounds are only operational during mode-finding, but not during MCMC (and therefore prior sampling)

Functionality is only supported in Matlab from 2014a/8.3 onwards. Closes #1266

Johannes Pfeifer authored Jun 01, 2016 and Stéphane Adjemian committed Jun 15, 2016

Closes #1149
Mirrors 1ad8df46

Pass title texts to ident_bruteforce.m to make sure that previous TeX output is not overwritten by advanced option

1) Filter out measurement errors with error message that suggests to explicitly write measurement errors in model definition
2) allow identification checks with correlations, by switching to numerical derivatives

Closes #1105

Do not pass options_ structure but only the required field (prior_trunc).

This reverts commit 1ad8df46.

Closes #945

Closes #677

Closes #949

Usually happens when model has a unit root and unconditional second moments are NaN

Trap the case where only one parameter is estimated and only one non-constant row is available in Jacobian: (1) do not use vnorm; (2) do not use advanced option.
(cherry picked from commit d9856879)

2) fixed bug for the ML base (bayestopt_.pshape==0);
3) make more transparent the error when the model does not solve at prior_mode or  prior_mean;
4) when prior_mean or prior_mode fail, try 50 times with randomly generated samples from the prior before interrupting identification.

This commit fixes issues related to fabiac forum request and subsequent discussion with Johannes.

http://www.dynare.org/phpBB3/viewtopic.php?f=1&t=4402&p=10771&hilit=identification&sid=64c6f9d987a2641e79dd5722137eb483#p10771

If options_.SpectralDensity.trigger is specified, identification would
otherwise plot graphs in every MC step

-) fixed bugs when there is only one parameter estimated or when simulated moments uncertainty provides zero diagonal terms in the covariance matrix of model matrices;
-) fixed bug in computing numerical derivatives of the LRE matrices in getJJ;

1) allow to compute derivatives starting from NUMERICAL derivatives of jacobian and steady state: this has a minor cost in accuracy and allow apply without errors identification and estimation with numerical derivatives;
2) added trap in dynare_estimation_init: if steadystate changes param values, automaticly shifts to numerical derivs of jacoban and steady state +  analytic derivatives of all the rest;
3) bug fixes for 2nd order derivatives w.r.t. model parameters;