Skip to content

/ statsmodels

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

ENH: State space models #2250

Merged
merged 173 commits into from Feb 12, 2015
Merged

ENH: State space models #2250

merged 173 commits into from Feb 12, 2015

Conversation

@ChadFulton
Copy link
Member

@ChadFulton ChadFulton commented Feb 12, 2015

Adds state space model estimation via a Cython Kalman filter implementation along with Python wrappers and an SARIMAX estimation class.

See #1698 for development details, discussion.
ChadFulton and others added 30 commits Nov 2, 2013
@ChadFulton
Adding Cython and pure Python general Kalman Filter estimators.
42eb719
@ChadFulton
Added methods separating the iteration of the Kalman filter from its …
7f7fc0d 
…setup.
@ChadFulton
Fixed issues with Kalman filter and added the initial values as optio…
0092983 
…nal arguments.
@ChadFulton
Refactored python kalman filter; added versions for single precision,…
e37004b 
… double precision, complex, and double precision complex.
@ChadFulton
FIX: Fixed problem with calculating initial values.
977eaca
@ChadFulton
ENH: Added blas functions for copying data (dcopy) and for multiplyin…
db79fce 
…g symmetric matrices (dsymm) for double precision reals.
@ChadFulton
DOC: Fix typo
ea6c0ba
@ChadFulton
ENH: Added *copy and *symm BLAS functions for faster data copying and…
4463abb 
… symmetric matrix multiplication.
@ChadFulton
FIX: *symm BLAS functions not available prior to scipy 0.13. Replaced…
ede921a 
… by *gemm calls for now.
@ChadFulton
FIX: Created single precision arrays correctly.
7673e8c
@ChadFulton
FIX: Typo in zcopy function declaration.
4b5903d
@ChadFulton
ENH: Made weakly exogenous data z and associated parameter matrix A o…
e355192 
…ptional.
@ChadFulton
ENH: Allowed iterations to use simpler inversion calculation when n=1.
9511e1c
@ChadFulton
FIX: Fixed error in vector length using *copy.
f5b74be
@ChadFulton
ENH: Added Cholesky decomposition.
114db4a
@ChadFulton
ENH: Improved performance by considering steady-stage convergence
84a022a
@ChadFulton
FIX: Updated convergence tolerance so that unit tests pass.
818d894
@ChadFulton
ENH: Added storage of inverse prediction error covariance matrix.
5490a14 
This matrix is useful for smoothers.
@ChadFulton
ENH: Added support for non-positive definite covariance matrix of ful…
82dc0a4 
…l state error term.
@ChadFulton
ENH: Added support for non-positive definite Q matrices.
4a82017
@ChadFulton
MAINT: Added readme and added license
2c2fa78
@jseabold @ChadFulton
COMPAT: Allow Python 3 compatibility. Closes #2.
6ac82fd
@ChadFulton
Refactored python kalman filter; added versions for single precision,…
3614ec1 
… double precision, complex, and double precision complex.
@ChadFulton
ENH: Added blas functions for copying data (dcopy) and for multiplyin…
47ea32e 
…g symmetric matrices (dsymm) for double precision reals.
@ChadFulton
ENH: Added *copy and *symm BLAS functions for faster data copying and…
d21cfb7 
… symmetric matrix multiplication.
@ChadFulton
DOC: Documented *copy arguments
d5849dd
@ChadFulton
FIX: Improved descriptions of arguments.
dcca501
@ChadFulton
ENH: Added Cholesky decomposition.
9a9e3b4
@jseabold @ChadFulton
COMPAT: Allow Python 3 compatibility. Closes #2.
6b68bdb
@ChadFulton
ENH: Added Kalman filter to build
2d7db3a
ChadFulton added 13 commits Feb 11, 2015
@ChadFulton
DOC: Fix confidence intervals in notebook.
a0c0c63
@ChadFulton
DOC: Fix local linear trend notebook Heading.
711d869
@ChadFulton
DOC: Fix sarimax stata notebook Heading.
3424df0
@ChadFulton
ENH: Add ad-hoc MLE estimation.
535e357
@ChadFulton
DOC: Improve documentation
Loading status checks…
a3025b9
@ChadFulton
TST: Reverse 3ad4a35; skip single precision tests.
e80d6a6
@ChadFulton
TST: Increase tolerance to pass cross-platform
Loading status checks…
b0f9388
@ChadFulton
TST: Skip remaining single precision tests.
Loading status checks…
4dd63a7
@ChadFulton
REF: rename params_included to param_terms
7a3e954
@ChadFulton
REF: Rename params_names to param_names
2b16a96
@ChadFulton
REF: Add parameter names to model data
a5965c9
@ChadFulton
REF: Rename loglikelihood attribute to llf_obs
2fc71e3
@ChadFulton
REF: Moved model attributes to specification dict
Loading status checks…
1479460
@josef-pkt

This comment has been minimized.

Sign in to view
Copy link

@josef-pkt josef-pkt commented on statsmodels/tsa/statespace/sarimax.py in 1479460 Feb 12, 2015

This formatting looks very "weird" to me
@josef-pkt josef-pkt mentioned this pull request Feb 12, 2015
Merged
@josef-pkt
Copy link
Member

@josef-pkt josef-pkt commented Feb 12, 2015

merging this

Thanks to Chad and everyone that helped and reviewed
josef-pkt added a commit that referenced this pull request Feb 12, 2015
@josef-pkt
Merge pull request #2250 from ChadFulton/statespace
Loading status checks…
f88cd69 
ENH: State space models
@josef-pkt josef-pkt merged commit f88cd69 into statsmodels:master Feb 12, 2015
2 checks passed
2 checks passed
continuous-integration/travis-ci The Travis CI build passed
Details
coverage/coveralls Coverage increased (+0.63%) to 83.88%
Details
@josef-pkt josef-pkt added this to the 0.7 milestone Feb 12, 2015
@bashtage
Copy link
Member

@bashtage bashtage commented Feb 12, 2015

@josef-pkt @ChadFulton FWIW this has broken Travis with Cythonize errors.
@bashtage
Copy link
Member

@bashtage bashtage commented Feb 12, 2015

This appears to be an issue with cython 0.22. It built fine on cython 0.21.2 on my machine.
@josef-pkt
Copy link
Member

@josef-pkt josef-pkt commented Feb 12, 2015

@bashtage where do you see those errors?
The test runs are green and I don't see anything spot checking the logs
@bashtage
Copy link
Member

@bashtage bashtage commented Feb 12, 2015

https://travis-ci.org/statsmodels/statsmodels
@bashtage
Copy link
Member

@bashtage bashtage commented Feb 12, 2015

Cython was updated on Anaconda in the last couple of hours.

Maybe these should be Py_ssize_t ?

Also reasonably likely Cython is broken.
@josef-pkt
Copy link
Member

@josef-pkt josef-pkt commented Feb 12, 2015

Cython was updated on Anaconda in the last couple of hours.

That explains why it didn't fail a few hours ago.

Maybe these should be Py_ssize_t ?

something like this for indices.

@jseabold ?
@ChadFulton
Copy link
Member Author

@ChadFulton ChadFulton commented Feb 12, 2015

Simple fix, which I will submit a PR for in a sec.
@hxu hxu mentioned this pull request Jul 7, 2015
Open
@ChadFulton ChadFulton deleted the ChadFulton:statespace branch Jan 16, 2016
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees
No one assigned
Labels
comp-tsa type-enh
Projects
None yet
Milestone
0.7
Linked issues

Successfully merging this pull request may close these issues.

None yet

4 participants
@ChadFulton @josef-pkt @bashtage @jseabold