Confronta i metodi

	Boosting	Gradient Boosting	Gradient Boosting Robusto
Campo	Apprendimento automatico	Apprendimento automatico	Apprendimento automatico
Famiglia	Machine learning	Machine learning	Machine learning
Anno di origine	1990–1997	2001	2001
Ideatore	Schapire, R. E.; Freund, Y.	Friedman, J. H.	Friedman, J. H. (with Huber loss from Huber, P. J.)
Tipo	Sequential ensemble (iterative reweighting)	Ensemble (sequential boosting of decision trees)	Ensemble (boosted trees with robust loss)
Fonte seminale	Freund, Y. & Schapire, R. E. (1997). A decision-theoretic generalization of on-line learning and an application to boosting. Journal of Computer and System Sciences, 55(1), 119–139.	Friedman, J. H. (2001). Greedy Function Approximation: A Gradient Boosting Machine. Annals of Statistics, 29(5), 1189–1232.	Friedman, J. H. (2001). Greedy function approximation: A gradient boosting machine. Annals of Statistics, 29(5), 1189–1232.
Alias	AdaBoost, gradient boosting, iterative reweighting ensemble, sequential ensemble	Gradient Boosting (GBM), GBM, gradient boosted trees, gradient boosting machine	gradient boosting with Huber loss, robust GBM, outlier-robust boosting, robust gradient-boosted trees
Correlati	6	5	6
Sintesi	Boosting is a sequential ensemble technique that converts many simple, barely-better-than-chance learners into a single highly accurate model by repeatedly focusing training on the examples that previous learners got wrong, then combining all learners with weights proportional to their individual accuracy.	Gradient Boosting is an ensemble learning method, formalised by Jerome H. Friedman in 2001, that combines a sequence of weak learners — typically shallow decision trees — so that each new tree is fitted to minimise the residual errors of the trees before it. It is the core algorithm behind popular implementations such as XGBoost, LightGBM and CatBoost.	Robust Gradient Boosting is gradient boosting trained with outlier-resistant loss functions — most commonly the Huber loss or quantile (pinball) loss — instead of squared-error loss. Proposed in Friedman's seminal 2001 paper, this variant produces predictions far less distorted by extreme values or contaminated labels, while retaining the full predictive power of gradient-boosted trees.