I have built a stacked model using mlxtend StakingCVClassifier. I want to know the feature importance scores now. Is there any way I can calculate feature importance scores for the stacked model? If not can I combine the feature importance of individual models somehow? I have tried SHAP values but StakingCVClassifier doesn't support that. Stacked model has RF, LightGBM , and MLP. Please let me know what are the possible solutions.



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.