What is the difference between pre-pruning and post-pruning in Decision Trees?

Pre-pruning stops tree growth early using constraints, while post-pruning first grows a full tree and then removes unnecessary branches.

Pre-pruning (early stopping) restricts tree growth during construction using conditions like max_depth, min_samples_leaf, or min_impurity_decrease. It reduces computation and prevents overfitting but may stop useful splits too early. Post-pruning builds a fully grown tree first, then removes branches that do not improve validation performance. This is typically more optimal because it evaluates global structure rather than local decisions. Cost-complexity pruning is a common post-pruning method used in CART, where a complexity penalty is applied to subtrees to select the best simplified model.