Can anyone please help me in understanding the effect of various bucketing techniques used in CatBoost Algorithm for categorical features? Like there is border, buckets, binarized target mean, counter encoding techniques, I am not able to get proper intuition of this, what is significance of different methods and how they affect model performances? I am referring the original documentation currently.



Your Answer

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