Use when working with LightGBM scikit-learn wrappers — LGBMClassifier, LGBMRegressor, LGBMRanker. Covers sklearn-compatible fit/predict, parameter aliases, categorical handling, and Pipeline/GridSearchCV integration.
