[Enhancement] Add predict_proba support for Random Forest with `criterion="entropy"`

[Innixma]

Version used: 2021.5.1

The following model gets a speedup with predict_proba:

from sklearnex.ensemble import RandomForestClassifier
model = RandomForestClassifier(..., criterion="gini")

However, this does not get a speedup:

from sklearnex.ensemble import RandomForestClassifier
model = RandomForestClassifier(..., criterion="entropy")

Could the entropy criterion of RandomForestClassifier be accelerated in predict_proba? When ExtraTrees support is added, this question also applies to specifying "entropy" for ExtraTrees.

[agorshk]

Hi, @Innixma, Thanks for request, we will analyze and plan this work, after that we will update the information