What action should be taken when a decision tree has too many nodes or leaves? (alternative wording)

• A. expansion
• B. generalization
• C. pruning
• D. reduction

Answer: (C)

Pruning is the action taken when a decision tree becomes too large with many nodes or leaves. The goal is to simplify the model to prevent overfitting and improve how well it generalizes to new data. Pruning removes parts of the tree that don’t add meaningful predictive power—cutting back branches or merging leaves based on validation performance or a cost-complexity criterion. This reduces complexity, helping the tree resist noise in the training data and perform better on unseen examples. Expanding the tree would increase size and likely hurt generalization, while generalization describes the desired outcome rather than a specific operation. Reduction isn’t the standard term for this process in decision trees.