ENH add post fit calibration to HGBT #26311
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
API wise, I have a questions: Should it be a separate method like Then, concerning the actual implementation, 2 important points:
|
|
@ogrisel @thomasjpfan You may be interested. |
|
Moving the milestone to 1.5 but this seems to be an important topic. |
…zenchr/scikit-learn into hgbt_post_fit_calibration
|
Moving in 1.7 but it would be in my priority to look at this for the next release. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Reference Issues/PRs
Closes #22435
What does this implement/fix? Explain your changes.
This PR adds the parameter
post_fit_calibrationtoHistGradientBoostingRegressor. In the case of a non-canonical link, i.e. only forloss="gamma", after the fit is more or less finished, a constant is added to the raw_predictions in link space such that on the training data (minus thevalidation_fraction), the balance property is fulfilled: the weighted average of predictions (predict) equals weighted average of observations, i.e.np.average(y, weights=sample_weight).Any other comments?
TODO list: