FastAI with other algorithms than Random Forest

Part 1 (2018)
1

I just wanted to know if all the best practices of FastAI works eqaully well if we are using any other classification or regression algorithm like XgBoost?
I will really appreciate any help. Thank you in advance.

2

HI @shivang.mathur
I have previously used fastai’s proc_df and add_datepart method for preprocessing before XGBoost and it improved the results by at least 2% for me.
FastAI’s best practices seem to have worked everywhere for me so far and there is no reason why a preprocessing step that works for RF should not work for XGBoost.

1 Like