rcurtin_irc changed the topic of #mlpack to: mlpack: a scalable machine learning library (https://www.mlpack.org/) -- channel logs: https://libera.irclog.whitequark.org/mlpack -- NOTE: messages sent here might not be seen by bridged users on matrix, gitter, or slack
<jonpsy[m]> XGBoost is awesome, isn't it?
<jonpsy[m]> can XGBoost handle extreme data imbalance in multi-class? Like one has 20k instances, while other has only 1.
<RishabhGarg108Ri> @jonpsy:matrix.org It is indeed awesome. :)
<RishabhGarg108Ri> You might wanna use `scale_pos_weight` for class imbalance. Not sure though if it will be good enough for the kind of imbalance that you are talking about.
<RishabhGarg108Ri> Can we have this in our implementation too?
<rcurtin[m]> RishabhGarg108 (RishabhGarg108): sure, I didn't look closely at it but I think at a quick glance it is similar to what we are already doing for the decision tree, just parallelized (I could be incorrect). If you do choose to implement it, it'll need to be with OpenMP 👍️
<jonpsy[m]> <RishabhGarg108Ri> "You might wanna use `scale_pos_w" <- correct, we cna use ```samlple_weihgts``` for multiclass.