Why scaling is not needed for Decision Tree based algorithms like Random forest and Xgboost?
Very good question.
Think about the micro decision in decision tree 10 < X < 20
If we scale the features down by say 100, then the above condition will become 0.01 < X < .02
Even after scaling down the decision will not change right.
Hence, decision tree models like Random forest and XGboost don’t require feature scaling.
I hope this answers your question.