Commit Graph

74 Commits (b8180f3328e1b47a0b90bd21e11b87feadc60f16)

Author SHA1 Message Date
Stéphane Adjemian (Charybdis) b8180f3328 Allow k order approximation in SIP filter. 2019-12-20 11:21:13 +01:00
Stéphane Adjemian (Charybdis) bc1c173988 Updated online filter routines.
- Rewrote doc headers
 - Changed function signatures
 - Removed persistent variables
 - Compute the mode of the particle weights
 - Return the covariance matrix of the particles in the last period
 - Various cosmetic changes
2019-07-11 12:45:26 +02:00
Frédéric Karamé 2aed30d686 New SMC sampler (Herbst and Schorfheide). 2019-02-11 10:59:51 +01:00
Sébastien Villemot f9e85d3c80 Update URLs for the move to GitLab 2018-09-10 12:18:49 +02:00
Frédéric Karamé 713e3ffcfb include modifications for parallelisation. 2018-06-28 09:22:34 +02:00
Frédéric Karamé a299c04d7e include parallelisation, correction of the likelihood calculation for the Amisano and Tristani option and the possibility of monte carlo draws for the proposal. 2018-06-28 09:21:50 +02:00
Frédéric Karamé 6e89a1ff9e Replace the likelihood of the mean by the mean of the likelihoods. 2018-06-28 09:20:19 +02:00
Frédéric Karamé 4de28e10c0 Initial particles are drawn in the prior distribution + bug fixes. 2018-06-13 18:43:04 +02:00
Frédéric Karamé 801d175890 Add flags to deal with errors on Cholesky matrices in CPF filter. 2018-06-11 09:46:17 +02:00
Frédéric Karamé b66d78d434 Correct the calculation of the incremental weights. 2018-06-09 14:59:20 +02:00
Frédéric Karamé cecc11c73f Add the option of Monte Carlo nonlinear Kalman filter. 2018-06-08 20:58:48 +02:00
Frédéric Karamé b7597b5060 Add the possibility of proposal approximated with Monte Carlo. 2018-06-08 20:57:38 +02:00
Frédéric Karamé be31310838 Correct the calculation of the incremental weights. 2018-06-08 20:56:24 +02:00
Frédéric Karamé d3d6a037b1 Added routines for Dynamic Striated Metropolis Hastings. 2018-06-01 11:16:29 +02:00
Stéphane Adjemian (Charybdis) 1eee8eb3f2 Fixed copyright notices. 2017-05-18 23:59:49 +02:00
Stéphane Adjemian (Charybdis) a812a00b17 Fixed indentation. 2017-05-18 23:59:10 +02:00
Houtan Bastani 215245758d update call to dyn_saveas 2017-03-23 17:57:17 +01:00
Houtan Bastani ce79dbe420 update call to dyn_figure 2017-03-22 17:08:02 +01:00
Stéphane Adjemian 0c28a29392 Merge pull request #8 from JohannesPfeifer/patch-3
Properly terminate nonlinear_kalman_filter if non-positive definite m…
2017-03-15 16:27:38 +01:00
Johannes Pfeifer bd8e804435 Properly terminate nonlinear_kalman_filter if non-positive definite matrices are encountered
Otherwise, crashes can happen during mode-finding
2017-01-23 12:20:24 +01:00
Frédéric Karamé 814cb693f6 Fix a bug in the predictive density calculation. 2016-06-30 12:12:58 +02:00
Frédéric Karamé bdd519a5f5 Modifications to facilitate Cholesky of states variances. 2016-06-30 11:42:45 +02:00
Houtan Bastani 4644a31291 fix latex file ending for linux compilation 2016-06-14 11:48:35 +02:00
Frédéric Karamé e6d081d58a Fixed bug on a structure name. 2016-06-01 18:24:27 +02:00
Frédéric Karamé 8221d5c780 Modification of a structure name when calling incremental_weights. 2016-06-01 18:19:26 +02:00
Frédéric Karamé 0141b34310 Fixed bug on structure name. 2016-06-01 17:11:28 +02:00
Stéphane Adjemian (Charybdis) 005b0907c0 Changed logic for setting weights method in CPF. 2016-06-01 16:07:07 +02:00
Stéphane Adjemian (Charybdis) 8f4fe9b341 Renamed matlab routine. 2016-06-01 14:12:02 +02:00
Frédéric Karamé 75b17ce780 Add the possibility of Gaussian-Mixture Particle Filter without resampling. 2015-12-04 16:00:04 +01:00
Frédéric Karamé 394df6f537 Make the distinction between original Amisano and Tristani's way to calculate particles weights and the likelihood (CPF1) and Murray's way (CPF2). 2015-12-04 15:29:25 +01:00
Frédéric Karamé 9a366d6576 Re-activate the option of resampling in Gaussian Particle Filter. 2015-12-04 14:48:26 +01:00
Frédéric Karamé df842aa542 Correct the auxiliary filter part. 2015-12-04 12:42:42 +01:00
Frédéric Karamé 3c18bf1f6b Fix a bug in the gaussian density calculation. 2015-12-04 11:44:45 +01:00
Frédéric Karamé e6d627396a Fix the calculation of the likelihood on the APF. 2015-10-20 15:12:56 +02:00
Frédéric Karamé 7816397856 Add the possibility to use an sparse-grid Kalman filter with approximation at order 2. 2015-10-09 14:42:53 +02:00
Frédéric Karamé 46f367b299 Minor modifications in the inputs of the function. 2015-10-07 15:25:16 +02:00
Frédéric Karamé 8b79fb562b Correction on a bug and extension of the original paper for the calculation of incremental weights. 2015-10-07 15:23:42 +02:00
Frédéric Karamé bf956f9bb0 Modify the inputs of gaussian_density. 2015-10-06 17:21:02 +02:00
Frédéric Karamé ce9dace8ef fix a bug on format matrix. 2015-10-06 17:18:09 +02:00
Frédéric Karamé 57f4f5b157 Fix a bug in likelihood calculation. 2015-10-02 16:05:26 +02:00
Frédéric Karamé a64073ec3c Modify the weights to prevent crash in the cholesky of the covariance. 2015-10-02 16:02:49 +02:00
Stéphane Adjemian c4b465a2c4 Merge pull request #6 from JohannesPfeifer/patch-1
Reflect new input argument of lyapunov_symm.m of Dynare
2015-08-12 11:45:04 +02:00
Johannes Pfeifer d5ef4962bd Reflect new input argument of lyapunov_symm.m of Dynare
Related to https://github.com/DynareTeam/dynare/pull/983
2015-07-24 12:57:17 +02:00
Stéphane Adjemian (Hermes) 042303da68 Cosmetic change (display of online approach progress). 2015-05-28 15:25:18 +02:00
Frédéric Karamé 47bfcece40 update the function to the recent modifications in dynare_estimation. 2015-05-27 15:26:59 +02:00
Frédéric Karamé 6ab142a75b simplification of gaussian mixtures 2015-05-27 11:48:06 +02:00
Frédéric Karamé 34cf8ca284 adapt the code to dynare evolutions 2015-04-15 10:40:27 +02:00
Frédéric Karamé 19858696bf fix a bug in the indices of the proposal calculations. 2015-01-28 15:23:09 +01:00
Frédéric Karamé 49d25782bb factorize some codes across options and modify the definition of mixtures. 2015-01-23 14:40:55 +01:00
Frédéric Karamé 2b0a83ca91 factorize some codes across options. 2015-01-23 10:05:16 +01:00