GBM

Gradient Boosting Machine

Ensemble of trees built sequentially to correct prior errors. XGBoost, LightGBM.

When you'd see it: Structured-data modeling and ML competitions. XGBoost and LightGBM are GBM implementations.

Why it matters: A GBM builds trees one after another, each correcting the last one's errors, which often makes it the most accurate option on tabular data. The cost is more tuning and more overfitting risk than a Random Forest.