Update exhaustive_feature_selector.py

[Haider010]

AttributeError: np.NINF was removed in the NumPy 2.0 release. Use -np.inf instead. Fixed this Error.

Code of Conduct

Description

Related issues or pull requests

Pull Request Checklist

• Added a note about the modification or contribution to the ./docs/sources/CHANGELOG.md file (if applicable)
• Added appropriate unit test functions in the ./mlxtend/*/tests directories (if applicable)
• Modify documentation in the corresponding Jupyter Notebook under mlxtend/docs/sources/ (if applicable)
• Ran PYTHONPATH='.' pytest ./mlxtend -sv and make sure that all unit tests pass (for small modifications, it might be sufficient to only run the specific test file, e.g., PYTHONPATH='.' pytest ./mlxtend/classifier/tests/test_stacking_cv_classifier.py -sv)
• Checked for style issues by running flake8 ./mlxtend

[d-kleine]

Has been fixed with #1107 already, this issue can be closed

[gedeck]

@d-kleine This fixes another occurrence of np.NINF in ExhaustiveFeatureSelector

However, not sure if this should be fixed in the same way as was done in #1107.

[d-kleine]

@gedeck Thanks for pointing that out - yeah, you are right. I overlooked that there are two different feature selector implementations (exhaustive vs. sequential). About how this issue can be fixed, I would like to leave the final decision to @rasbt.

[rasbt]

Thanks for the PR, and I might have missed that previously, sorry!

The fix looks good, but I think it might break in earlier NumPy versions then (not everyone has transitioned to 2.0 yet I assume). The fix in #1107 by @d-kleine was actually ideal. If we could apply it here too, that'd be great.

[d-kleine]

@Haider010 or @rasbt As I cannot contribute to this PR, please feel free to take my changes from #1107 for fixing this issue 🙂

[d-kleine]

Could you please release a new package version for pip users?