9.4.2 - Prunning

• An alternative to explicitly specifying the depth of a decision tree is to grow a very large, complex tree and then prune it back to find an optimal subtree.

• Prunning is activate by the cost complexity parameter (α) that penalizes our objective function.

\[\begin{equation} \operatorname{minimize}\{S S E+\alpha|T|\} \end{equation}\]