So my question is essentially the same as this one:

Why do we generate out-of-fold predictions for meta-ensembling/stacking?

However, I am not entirely satisfied with the answer (not detailed enough for my understanding).

Essentially, what's wrong with predicting on the training set for each model and using these predictions as metafeatures in our meta_training set? Which would be like seeing\training on the training target twice instead of once in the meta training set.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.