Dynamical Factor Models (DFM) Implementation (GSOC 2025) #446

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.

Sign up for GitHub

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

Open

andreacate wants to merge 21 commits into pymc-devs:main from andreacate:DFM_draft_implementation

Contributor

andreacate commented Mar 31, 2025 •

edited

Loading

Dynamical Factor Models (DFM) Implementation

This PR provides a first draft implementation of Dynamical Factor Models as part of my application proposal for the PyMC GSoC 2025 project. A draft of my application report can be found at this link.

Overview

Added DFM.py with initial functionality

Current Status

This implementation is a work in progress and I welcome any feedback

Next Steps

Vectorize the construction of the transition and selection matrices (possibly by reordering state variables).
Add support for measurement error.

Contributor

zaxtax commented Apr 1, 2025

Looks interesting! Just say when you think it's ready for review

Member

fonnesbeck commented Apr 5, 2025

cc @jessegrabowski

review-notebook-app bot commented Apr 7, 2025

Check out this pull request on

See visual diffs & provide feedback on Jupyter Notebooks.

Powered by ReviewNB

Contributor Author

andreacate commented Apr 7, 2025

Thanks for the feedback!

I'm still exploring the best approach for implementing Dynamic Factor Models.
I've added a simple custom DFM model in a Jupyter notebook, which I plan to use as a prototype and testing tool while developing the main BayesianDynamicFactor class.

jessegrabowski requested changes

View reviewed changes

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

jessegrabowski reviewed

View reviewed changes

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

jessegrabowski reviewed

View reviewed changes

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

andreacate force-pushed the DFM_draft_implementation branch 2 times, most recently from 21560db to a459a1a Compare

July 25, 2025 10:44

Member

jessegrabowski commented Jul 25, 2025

Some tests are failing due to missing constants. You might have lost some changes in the reset/rebasing process

andreacate force-pushed the DFM_draft_implementation branch from 1c04f65 to bc3fcf2 Compare

July 25, 2025 13:51

jessegrabowski requested changes

View reviewed changes

Member

jessegrabowski left a comment

Left some comments. I didn't look over the tests because they still seem like WIP, but seem to be on the right track!

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

andreacate force-pushed the DFM_draft_implementation branch 3 times, most recently from 7846f15 to e15cdd3 Compare

July 29, 2025 07:59

andreacate force-pushed the DFM_draft_implementation branch from e15cdd3 to 3b8bfe4 Compare

August 8, 2025 12:36

andreacate force-pushed the DFM_draft_implementation branch from 6496f38 to 615960b Compare

August 15, 2025 21:20

jessegrabowski requested changes

View reviewed changes

Member

jessegrabowski left a comment

I did a deeper pass on everything except the build_symbolic_graph method. I need to spend more time on that because it's gotten quite complex.

I'll finish ASAP.

tests/statespace/models/test_DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

andreacate and others added 12 commits

August 26, 2025 12:51


          Added new file DFM.py for GSOC 2025 Dynamical Factor Models

38f16a3


          Add initial notebook on custom DFM implementation

a07f9e6


          Update of DFM draft implementation

ced49ce

In the notebook a comparison between the custom DFM and the implemented DFM (which has an hardcoded version of make_symbolic_graph, that work just in this case)


          Aligning the order of vector state with statsmodel and updating the test

67efb3d


          Added test_DFM_update_matches_statsmodels and small corrections to DF…

85c96ec

…M.py


          Updating test following test_ETS.py and small adjustment for exog var…

4d5fcc5

…iables in DFM.py


          Added support for joint VAR modelling (error_var=True)

7fc5e11


          Adding a first implemntation of exogeneous variable support based on …

…pymc_extras/statespace/models/structural/components/regression.py


          Completing the implementation of exogeneous varibales support

5cd01da


          Small adjustments and improvements in DFM.py

ad8c0af


          Small adjustments and improvements in DFM.py

ba6ec2e


          Adjustments after Jesse review

c864af2

andreacate and others added 4 commits

August 26, 2025 12:51


          Adjustments following Jesse suggestions and added tests for exog support

d10fce8


          Added new DFM example notebook and deleted an old version of custom D…

3204a19

…FM example


          Add tests for names/dims/coords

1c35fc7


          De-duplicate exogenous dim between DFM and SARIMAX

ca1a86e

jessegrabowski force-pushed the DFM_draft_implementation branch from 8a03197 to ca1a86e Compare

August 26, 2025 05:20

jessegrabowski requested changes

View reviewed changes

pymc_extras/statespace/models/DFM.py Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated

+                              pt.zeros((self.k_endog, self.k_endog * (self.error_order - 1)), dtype=floatX)
+                          )
+                      if len(matrix_parts) == 1:
+                          design_matrix = factor_loadings * 1.0

Member

jessegrabowski Aug 26, 2025

What does this line do?

pymc_extras/statespace/models/DFM.py Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Outdated Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py Show resolved Hide resolved

pymc_extras/statespace/models/DFM.py

+                              self.ssm["state_cov", :, :] = factor_cov
+                      # Observation covariance matrix (H)
+                      if self.error_order > 0:

Member

jessegrabowski Aug 26, 2025

This seems backwards? The first case looks like it assumes error_order == 0.

This comment is based on error_sigma appearing in the else branch

pymc_extras/statespace/models/DFM.py

+                              sigma_obs = self.make_and_register_variable(
+                                  "sigma_obs", shape=(self.k_endog,), dtype=floatX
+                              )
+                              total_obs_var = error_sigma**2 + sigma_obs**2

Member

jessegrabowski Aug 26, 2025

Are you sure error_sigma should appear here? It's already appearing in Q, so is this double-counting?

jessegrabowski reviewed

View reviewed changes

tests/statespace/models/test_DFM.py Show resolved Hide resolved

jessegrabowski reviewed

View reviewed changes

tests/statespace/models/test_DFM.py Outdated Show resolved Hide resolved

jessegrabowski reviewed

View reviewed changes

tests/statespace/models/test_DFM.py Show resolved Hide resolved

andreacate and others added 5 commits

August 26, 2025 15:50


          Small adjustments and refactoring after code review

7252d46


          Allow exogenous regressors in BayesianVARMAX (pymc-devs#567)

b1a3c27

* First pass on exogenous variables in VARMA

* Adjust state names for API consistency

* Allow exogenous variables in BayesianVARMAX

* Eagerly simplify model where possible

* Typo fix


          Small adjustments in the tests after review

a329450


          Harmonizing names for EXOG dimension between DFM and VARMAX

4b6f804


          Merge branch 'main' into DFM_draft_implementation

b659d61

andreacate marked this pull request as ready for review

August 27, 2025 09:21

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet