Describe the issue linked to the documentation

There's no where in the documentation that explains what method is used to identify which values to consider as candidate splits. For example, for regression, an exhaustive method would be to sort each feature and then use the halfway point between each feature value. In this case there would be O(N*F) candidate splits where N=# data and F= # features. Another way would be to randomly choose K values between the min and max of each feature.

Does anyone know what method is used? Could someone point me to where in the code this occurs?

Suggest a potential alternative/fix

No response