How do Decision Trees handle outliers?

Decision Trees are relatively robust to outliers, but extreme values can still distort split decisions.

Decision Trees are less sensitive to outliers compared to linear models because they rely on threshold-based splits rather than distances. However, outliers can still influence split thresholds, especially if they affect impurity calculations significantly. A single extreme value may create a split that isolates it, leading to unnecessary branches. Techniques like pruning or limiting depth help mitigate this effect.