Information Gain and Entropy

Entropy-Based Pruning

Post-pruning technique using entropy-based criteria to evaluate whether removing a branch improves the model's bias-variance tradeoff. Compares the potential information gain to the added complexity cost.

← Terug