What is the effect of dataset imbalance on impurity calculations?

Imbalanced datasets bias impurity measures toward majority classes, reducing sensitivity to minority patterns.

Gini impurity and entropy are weighted by class distribution. In imbalanced datasets, majority class dominates impurity reduction calculations, making it harder for the tree to prioritize minority class splits. As a result, decision trees may produce high accuracy but poor recall for minority classes. Techniques like class weighting, resampling, or custom loss adjustments are required to correct this bias.