Classification and Regression Trees

Authors: Leo Breiman, Jerome H. Friedman, Richard A. Olshen, Charles J. Stone Year: 1984 Citations: 20,000+

Summary

Foundational decision tree methodology. CART algorithm with binary splitting, pruning strategies, and missing value handling. Trees became fundamental interpretable ML models.

Key Concepts

• Recursive Binary Splitting: Greedy feature space partitioning
• Gini Index: Impurity measure for splitting
• Pruning: Reduces overfitting
• Interpretability: Direct decision rule structure
• Non-parametric Learning: No distribution assumptions

Impact

• 20,000+ citations
• Basis for Random Forests, Gradient Boosting
• Foundation for tree-based ensemble methods
• Influential for interpretable ML
• Practical interpretability advantages

Related Papers

• Random Forests (Breiman, 2001)
• Gradient Boosting (Friedman, 2001)