1875 commits

This branch

This branch

All branches

Author	SHA1	Message	Date
mu	4c2cd58b83	t	2014-05-21 13:00:09 +01:00
mu	cd0e4b15b6	t	2014-05-21 12:59:09 +01:00
mu	c1a933cab4	t	2014-05-21 12:58:57 +01:00
mu	af5492c94e	st	2014-05-21 12:36:30 +01:00
mu	a471083c7b	old	2014-05-11 10:34:57 +01:00
mu	4a94933c19	Testing	2014-02-24 18:44:39 +00:00
mu	27488917a9	old	2014-02-24 15:05:14 +00:00
mu	ad6ea205b7	old	2014-02-24 15:04:25 +00:00
mu	8b9170ad96	xt	2014-02-24 11:26:23 +00:00
mu	55a9c5a423	xt	2014-02-17 09:18:35 +00:00
mu	f956590fdd	xt	2014-02-14 12:15:40 +00:00
mu	4b3930a52e	UY	2014-02-04 19:47:51 +00:00
mu	a243a8eabe	Merge branch 'kalman' of https://github.com/SheffieldML/GPy into kalman	2014-02-04 14:13:15 +00:00
mu	dc2ff00938	xt	2014-02-04 13:17:00 +00:00
Arno Solin	4e8c92407e	Added a new way of calculating the likelihood gradient. The code is still a draft, but looks promising.	2014-02-04 13:48:35 +02:00
mu	267e0b4f1b	xt	2014-02-04 10:48:27 +00:00
Arno Solin	51cbaced94	Initial version of the likelihood gradient evaluation.	2014-01-30 18:52:26 +02:00
Arno Solin	388ba68dcf	Added the derivatives of the state space matrices for likelihood gradient evaluation.	2014-01-30 18:51:42 +02:00
Arno Solin	c7a09de80c	Now also the derivative matrices can be summed.	2014-01-30 18:50:54 +02:00
mu	b980734fb8	st	2014-01-17 15:41:43 +00:00
mu	de39a22a4b	xt	2014-01-14 13:19:22 +00:00
mu	4fe0583c7d	xt	2014-01-14 08:43:06 +00:00
mu	1aa060439f	xt	2014-01-10 18:19:05 +00:00
mu	7072cfe7a0	state1	2014-01-09 17:14:31 +00:00
mu	8638e05d81	sep xt	2013-12-10 11:07:07 +00:00
mu	d44e16e615	sep xt	2013-12-09 17:59:33 +00:00
Arno Solin	9729a9a3a5	Now several trajectories can be simulated in one go.	2013-12-03 20:50:22 +02:00
mu	94c7047b82	xt	2013-11-29 18:09:43 +00:00
mu	1e01d79072	xt	2013-11-29 17:14:49 +00:00
Arno Solin	59afe14434	Removed unused code.	2013-11-28 15:17:21 +02:00
Arno Solin	68af1db76a	Posterior and prior samples are now handled.	2013-11-28 15:08:51 +02:00
Arno Solin	95948c760e	fixed bug that prevented the smoother from running.	2013-11-28 14:11:44 +02:00
Arno Solin	5fc80202ab	predict_raw can take arbitary Ys.	2013-11-28 14:03:45 +02:00
mu	71891a95d3	xt	2013-11-28 11:09:40 +00:00
mu	7247849441	change state_xt	2013-11-27 12:55:55 +00:00
mu	d0ff97f37e	statespace xt	2013-11-26 15:34:21 +00:00
mu	242307e6ed	stateSpace xt	2013-11-26 15:33:39 +00:00
Arno Solin	9fff81b7ce	Merge branch 'kalman' of ssh://github.com/SheffieldML/GPy into kalman	2013-11-22 17:01:26 +00:00
Arno Solin	8ed9bbcb4d	Now you can plot the filter estimate.	2013-11-22 17:01:13 +00:00
James Hensman	d70d8f6146	added todo	2013-11-22 11:59:08 +00:00
Arno Solin	27a79945a6	SDE kernels can now be summed.	2013-11-19 16:31:25 +00:00
Arno Solin	3b24991c81	SDE representation of the exp covaraince function.	2013-11-19 16:30:30 +00:00
Arno Solin	12bde7aeb4	Avoid Cholesky in 1D models	2013-11-19 16:29:35 +00:00
Arno Solin	15a8a270d4	Small fixes.	2013-11-13 18:14:11 +00:00
Arno Solin	a25844ac71	Round to 14 decimals to avoid problems.	2013-11-13 17:26:00 +00:00
Arno Solin	2303132933	Now the model for each dt is calculated only once	2013-11-11 23:48:38 +00:00
Arno Solin	53f0fbcbff	Use solve in lti_disc	2013-11-11 22:53:45 +00:00
Arno Solin	f4425efdae	Ensure postive params, added plot function.	2013-11-11 17:12:40 +00:00
Arno Solin	1b0cb8fd77	Fixed bug in smoother. Now working.	2013-11-10 22:46:36 +00:00
Arno Solin	d6bae3364e	Fixed bug in log_likelihood evaluation	2013-11-10 20:38:41 +00:00