I will make some for loop on to select the best features by my Data frame is big 10M row and about 50 columns so if i replaced xgb with a single Decision tree would it be the same results for the best features and i can choose them on xgb or it may differ, i just want to save time as much as possible.

  • $\begingroup$ It will differ. Why removing features? $\endgroup$
    – Michael M
    Apr 26, 2020 at 8:31
  • $\begingroup$ because i have added many like interaction variables and so on and i need to remove some to prevent overfitting $\endgroup$ Apr 26, 2020 at 13:54


Your Answer

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