Add Propensity Score DAG to Knowledge Base #337
Conversation
Signed-off-by: Nathaniel <[email protected]>
|
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
Signed-off-by: Nathaniel <[email protected]>
|
This should be good now @drbenvincent! |
There was a problem hiding this comment.
Comments about the DAG notebook
- Please add the
remove-inputcell tag to the new cell which builds the propensity DAGs inquasi_dags.ipynb - Could be worth briefly explaining
$Y(0), Y(1)$ because so far we've got mentioned it and have nothing from the potential outcomes framework. Can be very brief - we'll likely end up with a specific page on the potential outcomes framework - Nit: Use of PS as an acronym without defining it
| @@ -63,7 +63,7 @@ def test_regression_kink_gradient_change(): | |||
|
|
|||
| def test_inverse_prop(): | |||
| df = cp.load_data("nhefs") | |||
| sample_kwargs = {"tune": 100, "draws": 100, "chains": 2, "cores": 2} | |||
| sample_kwargs = {"tune": 100, "draws": 500, "chains": 2, "cores": 2} | |||
There was a problem hiding this comment.
Could add a random_seed kwarg to sample_kwargs in order to make test output deterministic.
So far this is not generally done in tests because we don't test for specific results, but it might be relevant here if you're getting numerical stability issues.
We do do it however in the docstring examples (under test by doctest) because there we do examine numerical results.
I'll leave it up to you, but this might be the more defensive play.
There was a problem hiding this comment.
Updated. Added Random seed and made the above changes
Signed-off-by: Nathaniel <[email protected]>
Signed-off-by: Nathaniel <[email protected]>
Signed-off-by: Nathaniel <[email protected]>
drbenvincent
changed the title
Adding random seed to fix bug #336
and adding DAG write up re propensity scores as per #327