Skip to content

Releases: fipelle/MessyTimeSeriesOptim.jl

Simplified initialisation

14 Feb 15:50
333164f
Compare
Choose a tag to compare
Merge pull request #24 from fipelle/dev

Further simplifications in the initialisation

Multivariate initialisation

04 Dec 21:43
c4f24b6
Compare
Choose a tag to compare
Merge pull request #23 from fipelle/dev

Multivariate initialisation

Bug fix for new DFM initialisation

18 Oct 22:50
64ca8e0
Compare
Choose a tag to compare
Merge pull request #22 from fipelle/dev

Update to 0.2.1

Fix in initialisation for DFM models

19 Sep 20:53
1a3c7ce
Compare
Choose a tag to compare
Merge pull request #21 from fipelle/dev

New version: fix in init

Patch release: initialise idio cycle as ar(1)

24 Dec 15:05
008faca
Compare
Choose a tag to compare
Merge pull request #20 from fipelle/dev_init

Initialise idio cycle as ar(1)

Use of compute_scaling_factors(...) within validation + bug fixes

09 Dec 18:14
32e783a
Compare
Choose a tag to compare
Merge pull request #19 from fipelle/fipelle-patch-1

Update LICENSE

v0.1.6

27 Jul 12:31
770d5b3
Compare
Choose a tag to compare

Added support the new sequential approach for filtering and smoothing in MessyTimeSeries.jl

v0.1.5

21 Jul 23:22
080a9ca
Compare
Choose a tag to compare

Patch release

v0.1.4

18 Jul 23:48
96a49d4
Compare
Choose a tag to compare

Implemented Kitagawa's trend for non-stationary DFMs

v0.1.3

23 Mar 13:46
edf4ced
Compare
Choose a tag to compare

MessyTimeSeriesOptim v0.1.3

Diff since v0.1.2

Closed issues:

  • Get rid of StableRNG? (#6)

Merged pull requests:

  • Logo (#8) (@fipelle)
  • Implemented N as a sparse matrix -> notable memory improvements for high-dimensional problems (#9) (@fipelle)