Describe how to implement and optimize ensemble methods.

Ensemble methods combine multiple models to achieve better performance than individual models through diversity and aggregation.\n\n• Bagging: Random Forest, Extra Trees - reduce variance\n• Boosting: XGBoost, AdaBoost - reduce bias sequentially\n• Stacking: Meta-learner combines base model predictions\n• Voting: Hard/soft voting for classification, averaging for regression\n\nExample: Create ensemble with Random Forest, XGBoost, and Neural Network. Use 5-fold CV for stacking, optimize base model diversity, tune ensemble weights. Monitor for diminishing returns and computational cost.