Michel Juillard
f77b101d7b
fixing bugs in random generator handling code
2012-08-29 22:12:50 +02:00
Michel Juillard
378413ed14
fixing typo in previous commit
2012-08-29 21:02:33 +02:00
Michel Juillard
dbdbfdd926
adding set_dynare_random_generator_state() and
...
get_dynare_random_generator_state(). Use now different seeds for
different Metropolis chains. Fixed handling of random generator state
thourghout the code.
2012-08-29 17:58:54 +02:00
Stéphane Adjemian (Charybdis)
d9f3ab5be8
Partially revert commit #69efc894c6dc9ac1250bd7450bd57443f088c242. Test for Octave/Matlab to decide how to call the print command. Added a warning stating that Octave cannot create pdf files.
2012-08-29 16:50:08 +02:00
Michel Juillard
d60202616a
fixed problem with penalty in estimation. Created a new global scalar:
...
objective_function_penalty_base. It is the only simple way that I
found to keep csminwel1.m to be able to handle general functions.
2012-08-28 12:17:07 +02:00
Michel Juillard
526d6ca76c
removing unused function
2012-08-28 11:55:16 +02:00
Marco Ratto
e519b04713
bug fix: when nograph=1, SmoothedShocks were not saved.
2012-08-27 17:45:47 +02:00
Marco Ratto
1193cab7c8
Make the function compatible with multiple file formats.
2012-08-27 16:37:14 +02:00
Stéphane Adjemian (Charybdis)
0508bf9a41
Added missing default options for stochastic simulation of nonlinear backward looking models.
...
(cherry picked from commit 4053f664ecec14ffb7269c6bbf78355ca46b6b92)
2012-08-27 12:53:11 +02:00
Marco Ratto
a7ba2b51e7
Trap error when the model does not solve for point estimation (prior mean-mode posterior mean-mode)
2012-08-24 17:09:13 +02:00
Marco Ratto
9edce5414f
bug fix for octave.
2012-08-24 16:40:24 +02:00
Marco Ratto
fda047e19c
updated penalty has to be properly passed to the objective function.
...
This relates should fix behaviour after commit 6b3bd9dd0b
.
2012-08-24 16:39:04 +02:00
Stéphane Adjemian (Charybdis)
fe98a0875d
Added the possibility to save the plots generated by the shock_decomposition command.
2012-08-24 15:03:35 +02:00
Stéphane Adjemian (Charybdis)
150256268f
Fixed mixed indices (j was used for indexing two nested loops).
2012-08-24 15:00:21 +02:00
Stéphane Adjemian (Charybdis)
69efc894c6
Fixed bug. Removed extension declaration for Encapsulated PostScript graphics file format.
2012-08-24 14:56:44 +02:00
Stéphane Adjemian (Charybdis)
b89bdb6fd6
Bug fix. Call evaluate_steady_state_routine, otherwise evaluate_smoother crashes if the steady state of the model is specified using the steady state model block (or writing a steadystate2.m routine).
2012-08-24 12:51:44 +02:00
Sébastien Villemot
56914e3d4c
Require at least Octave 3.4 (linsolve does not compile against 3.2)
2012-08-23 16:40:00 +02:00
Marco Ratto
6b91301365
Reduce memory requirements for analytic Hessian.
...
Tested with QUEST III (Ratto et al. 2009): 63 params and 59 state dimension.
2012-08-21 16:00:55 +02:00
Marco Ratto
b02303cf69
Force analytic_derivation = 1;
2012-08-21 15:53:02 +02:00
Marco Ratto
b1dd7a5137
Allow quicker evaluation of likelihood with analytic derivatives.
2012-08-21 15:46:35 +02:00
Marco Ratto
7683175e8e
Bug fix in terms for outer product gradient with analytic derivatives
2012-08-21 15:45:25 +02:00
Marco Ratto
9fd8bf954e
bug fix in formula for analytic Hessian.
2012-08-15 14:19:34 +02:00
Marco Ratto
f7aae77f16
removed useless elements + reduce loops for efficiency
...
(cherry picked from commit f2271264994af253ccdc7f01478320e4d2c2c0e7)
2012-08-13 13:04:34 +02:00
Marco Ratto
3ddafb164b
-) Added missing terms for analytic Hessian when steady state depends on estimated params;
...
-) bug fixes;
(cherry picked from commit c84f70f6630f4988716dcb4ea59315180bbb36e7)
2012-08-13 13:04:07 +02:00
Michel Juillard
70e162c736
removed globals from solve1.m
2012-08-06 23:32:40 +02:00
Sébastien Villemot
db62e3d7a0
issquare exists under Octave and returns the matrix dim if it is square
2012-08-06 18:26:16 +02:00
Sébastien Villemot
2e65a9ab96
Provide a better implementation of linsolve for Octave
...
Closes : #273
2012-08-06 18:26:16 +02:00
Michel Juillard
a400305f20
updating list of info codes returned by dynare_resolve() in DsgeVarLikelihood
2012-08-06 12:02:35 +02:00
Michel Juillard
2e9ae5df28
dealing with info == 8 (Nan in Jacobian) during estimation
2012-08-06 12:00:03 +02:00
Sébastien Villemot
336fc56518
Merge remote-tracking branch 'jpfeifer/master'
2012-08-06 11:44:15 +02:00
Michel Juillard
7e221e5d31
factoring setting of bayestop_.penalty = Inf in Metropolis
2012-08-05 15:27:14 +02:00
Michel Juillard
7959102d43
removed global and varargin in Metropolis routines
2012-08-05 15:10:21 +02:00
Johannes Pfeifer
1a968d5f22
Added check for NaN in Jacobian to stochastic_solvers and added explicit error message to print_info.m
2012-08-05 13:09:24 +02:00
Michel Juillard
b1ad31cf56
removed seeding random generators from clock in cmaes and metropolis hastings
2012-08-05 10:59:59 +02:00
Michel Juillard
6b3bd9dd0b
penalty is now passed as a field of bayestopt_ (or BayesInfo)
2012-08-02 22:23:29 +02:00
Marco Ratto
47353a276b
bug fix: the anlytic hessian was wrong with estimated stderr.
...
Manual cherry pick from Commit: 36853cb7615de41ca61ed92decf8bbe81ab40cba
2012-08-02 14:57:20 +02:00
Michel Juillard
fe8ae2c565
adding doc *.m file for k_order_perturbation MEX
2012-08-02 11:25:52 +02:00
Michel Juillard
9e6c58572e
fixing bug in pruning at order=3
2012-08-02 10:09:44 +02:00
Michel Juillard
2244af9c62
adding pruning option for 3rd order
2012-07-31 21:50:59 +02:00
Sébastien Villemot
488de200cf
Merge remote-tracking branch 'jpfeifer/master'
2012-07-30 15:50:30 +02:00
Michel Juillard
87b78f657b
dynatable: fixed bug when table contains Inf or NaN
2012-07-25 17:52:00 +02:00
Michel Juillard
24e1c49928
simulations: draw random shocks so that, given a seed, the sequence is the same, independently of the number of periods
2012-07-25 15:00:57 +02:00
Michel Juillard
d32e076b77
adding missing function linsolve for Octave (inefficient and minimal implementation
...
for current needs)
2012-07-22 22:18:32 +02:00
Michel Juillard
e7a8261b17
first order and cycle reduction: fixing bugs in previous commits
2012-07-22 19:54:12 +02:00
Michel Juillard
ba1f5eedc2
first order solver:
...
-pass along errors returned by cycle reduction algorithms;
-logarithmic reduction fails on error
-replace expensive and random condest() by call to linsolve()
-uses linsolve() to compute hx
2012-07-22 12:56:51 +02:00
Michel Juillard
2c450c79f8
cycle reduction: return error codes for over- and under-determined
...
dynamic systems
2012-07-22 12:51:55 +02:00
Michel Juillard
edf826f315
added code to optionally run cycle_reduction algorithm on GPU. Added
...
options_.gpu (no interface yet, use options_.gpu=1). Added test
./tests/first_order/fs2000_cr.mod
2012-07-20 17:06:12 +02:00
Michel Juillard
8d33a434e5
fixing bugs in dyn_first_order_solver()
2012-07-19 18:57:56 +02:00
Michel Juillard
d39bbdaa85
forcing reinitialization of persistent variables of
...
dyn_first_order_solver.m in global_initialization.m This is necessary
if noclearall option is used.
2012-07-19 17:08:42 +02:00
Johannes Pfeifer
b10b2160ed
Made error message if initial parameter values are outside prior bounds more explicit.
2012-07-19 12:30:41 +02:00