GridSearchCV does not seem to recognize whether estimators from StackingClassifier are fitted or not #24409
Comments
|
I believe this is because the search clones its estimator
See also the broader issue #8370. Is there precedent for overriding |
|
Indeed, the issue comes from cloning where any nested estimator will be cloned regardless of some parameters of the meta-estimator. Here, the We should be able to overwrite the cloning for such case. I think that the following SLEP would be relevant: https://github.com/scikit-learn/enhancement_proposals/pull/67/files |
Describe the bug
There seems to be a bug with the combination of
GridSearchCVandStackingClassifierwhen the parametercvofStackingClassifieris set to 'prefit'. With this option, the estimators of theStackingClassifiershould be fitted before fitting the stacked model, and only the final_estimator would then be fitted. When including theStackingClassifierwithinGridSearchCVhowever, the fact that estimators have already been fitted does not seem to be recognized.Steps/Code to Reproduce
Expected Results
In the above code, the
stackmodel works fine and no error related to the estimators' previous fitting is thrown.However, when included in the GridSearchCV, the fact that estimator models have been fitted already does not seem to be recognized and a
sklearn.exceptions.NotFittedErrorerror is prompted.Actual Results
Versions
System: python: 3.10.0 (tags/v3.10.0:b494f59, Oct 4 2021, 19:00:18) [MSC v.1929 64 bit (AMD64)] executable: C:\Users\levesque\Documents\Python\AMEX\env\Scripts\python.exe machine: Windows-10-10.0.17763-SP0 Python dependencies: sklearn: 1.1.2 pip: 22.2.2 setuptools: 57.4.0 numpy: 1.23.1 scipy: 1.9.0 Cython: None pandas: 1.4.3 matplotlib: 3.5.2 joblib: 1.1.0 threadpoolctl: 3.1.0 Built with OpenMP: True threadpoolctl info: user_api: openmp internal_api: openmp prefix: vcomp filepath: C:\Users\levesque\Documents\Python\AMEX\env\Lib\site-packages\sklearn\.libs\vcomp140.dll version: None num_threads: 8 user_api: blas internal_api: openblas prefix: libopenblas filepath: C:\Users\levesque\Documents\Python\AMEX\env\Lib\site-packages\numpy\.libs\libopenblas.FB5AE2TYXYH2IJRDKGDGQ3XBKLKTF43H.gfortran-win_amd64.dll version: 0.3.20 threading_layer: pthreads architecture: Haswell num_threads: 8 user_api: blas internal_api: openblas prefix: libopenblas filepath: C:\Users\levesque\Documents\Python\AMEX\env\Lib\site-packages\scipy\.libs\libopenblas.PZA5WNOTOH6FZLB2KBVKAURAKVTFSNNU.gfortran-win_amd64.dll version: 0.3.18 threading_layer: pthreads architecture: Haswell num_threads: 8The text was updated successfully, but these errors were encountered: