Consider decision tree algorithms, which splitting criterion should be used to bal$-$ ance the ability to create balanced branches in sizes and the ability to create pure children nodes?

A$.$ Minimize training error rate

B$.$ Maximize goodness of split

$(12)$ Which statement is INCORRECT about decision tree algorithms?

A$.$ In decision tree algorithm, pre$-$classified target variables must be included in training set.

B$.$ In decision tree algorithm, the fully grown tree always has higher or equal' training accuracy than pruned trees.

Q Search

C$.$ Minimize Gini impurity

D$.$ Maximize information gain