Add doubly robust estimator + notebook #1201
Conversation
removed test file from commit Add notebook and data Signed-off-by: rahulbshrestha <[email protected]>
rahulbshrestha
force-pushed
the
doubly_robust_estimator
branch
from
9abeacc
to
ffa3a1d
Compare
|
@amit-sharma can you take a look? |
|
This PR is stale because it has been open for 60 days with no activity. |
|
This PR was closed because it has been inactive for 7 days since being marked as stale. |
|
@amit-sharma any chance to take a look? |
There was a problem hiding this comment.
@rahulbshrestha I've added some comments to iterate on the PR.
The estimator is still incomplete. So it needs some work before it can be merged.
| from dowhy.utils.api import parse_state | ||
|
|
||
|
|
||
| class DoublyRobustEstimator(CausalEstimator): |
There was a problem hiding this comment.
The comment for this class may need to be changed. This estimator is for the standard backdoor setting when you want to combine propensity and adjustment methods.
| estimation of conditional treatment effect over it. | ||
| :param first_stage_model: First stage estimator to be used. Default is | ||
| linear regression. | ||
| :param second_stage_model: Second stage estimator to be used. Default |
There was a problem hiding this comment.
change the default to propensityscore
| **kwargs, | ||
| ) | ||
| self.logger.warning("First stage model not provided. Defaulting to sklearn.linear_model.LinearRegression.") | ||
|
|
There was a problem hiding this comment.
do we need a similar custom initialization for the second stage model?
| ) | ||
|
|
||
| pass | ||
| #return estimate |
There was a problem hiding this comment.
these two estimates need to be combined using the doubly robust formula
| "metadata": {}, | ||
| "outputs": [], | ||
| "source": [ | ||
| "data = pd.read_csv(\"./learning_mindset.csv\")\n" |
There was a problem hiding this comment.
you can add the dataset to the datasets/ folder.
This draft PR addresses this issue, by introducing the doubly robust estimator to doWhy. It also contains a tutorial notebook applying this algorithm to a demo dataset. This is a rough draft, so there will be lots of changes!