Boosting: AdaBoost and Gradient Boosting

Bagging and random forests fit many trees independently and average away variance. Boosting fits trees sequentially, each one deliberately targeting the mistakes of the ensemble built so far — attacking bias rather than variance, and closing out this topic's tree-based methods with the technique behind the gradient-boosted-tree models (XGBoost, LightGBM) that dominate tabular machine learning competitions and much of quant tabular alpha research today.