Tree-based Methods

Key concepts:

• CART splitting criteria, max depth, pruning.
• Random Forests: bagging, OOB error, feature importance caveats.
• Gradient Boosting: weak learners, learning rate, regularization, XGBoost/LightGBM.

Common interview checks:

• Compare bias/variance across trees, RF, and GBMs.
• Handle imbalance and monotonic constraints.
• Interpretability vs. performance trade‑offs.