Add initial explainer. #25
Merged
Conversation
This file contains 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
Merge Dev
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Outdated suggestions cannot be applied.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.
Additions
This PR adds mainly the first iteration of the base explainer to the
shapiqpackage which should serve as the main interface to the package. This PR also adds the KernelSHAP approximator, which is the first estimator to compute only Shapley values (SVs) ofmax_order=1.Explainer
The explainer supports marginal feature removal and is currently quite limited in its application and needs further testing. Main features include:
New Approximator: KernelSHAP
The
KernelSHAPapproximator is a regression estimator of order 1. It computes the SV according to Lundberg er al.‘s definition with multiple modernizations and tweaks that materialized over the recent years. One of those optimizations is paired sampling.Cleanup
The main code quality improvements include a refactoring of the tests into submodules and some tweaks with the imports.