In classification and regression trees (CART), it is performed by the model itself based on impurity measures. Features

[answerhappygod]

In classification and regression trees (CART), it is performed
by the model itself based on impurity measures. Features that are
used in CART are considered the most important ones. Some analysts
suggested that users do not have to select features prior to the
construction of CART. However, some other analysts disagreed and
argued that, as long as we need to run models, feature selection
remains a critical step before model construction. Do you agree
that feature selection by users before running CART models is
important? Why or why not?