In a decision tree used for

classification, what is the primary goal when choosing a feature for a

split at any node?

Question

In a decision tree used for 
 
classification, what is the primary goal when choosing a feature for a 
 
split at any node?

Answer

To select the feature with the 
 
fewest unique values.

Answer

To increase the depth of the tree as much as 
 
possible.

Answer

To create child nodes that are as pure as possible (i.e., 
 
containing samples from a single class).

Answer

To create child nodes that are as mixed and 
 
diverse as possible.

Crowdly