I have been reading material that suggesting that, after each grid search you do on a single parameter (e.g say on learning rate), you should update the number of boosting rounds afterwards. This is very expensive (time). I also intend to use GridSearchCV, and from my understanding, you cannot use early stopping rounds and GridSearchCV simultaneously.

Why is it necessary to update the number of boosting rounds each time you update a parameter?

  • $\begingroup$ maybe this approach can help speed things up $\endgroup$ – oW_ Aug 13 '20 at 16:57

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.