Compara mètodes

	Gradient Boosting	LightGBM	Aprenentatge en línia
Camp	Aprenentatge automàtic	Aprenentatge automàtic	Aprenentatge automàtic
Família	Machine learning	Machine learning	Machine learning
Any d'origen	2001	2017	1958–2000s
Autor original	Friedman, J. H.	Ke, G. et al. (Microsoft)	Rosenblatt, F.; Littlestone, N.; Shalev-Shwartz, S. (key contributors)
Tipus	Ensemble (sequential boosting of decision trees)	Gradient boosting decision tree ensemble	Learning paradigm (sequential model update)
Font seminal	Friedman, J. H. (2001). Greedy Function Approximation: A Gradient Boosting Machine. Annals of Statistics, 29(5), 1189–1232.	Ke, G., Meng, Q., Finley, T., Wang, T., Chen, W., Ma, W., Ye, Q. & Liu, T.-Y. (2017). LightGBM: A Highly Efficient Gradient Boosting Decision Tree. Advances in Neural Information Processing Systems (NeurIPS) 30, 3146–3154.	Shalev-Shwartz, S. (2011). Online Learning and Online Convex Optimization. Foundations and Trends in Machine Learning, 4(2), 107–194.
Àlies	Gradient Boosting (GBM), GBM, gradient boosted trees, gradient boosting machine	LightGBM, Light Gradient Boosting Machine, lgbm, leaf-wise gradient boosting	incremental learning, sequential learning, streaming learning, online machine learning
Relacionats	5	5	6
Resum	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.	LightGBM is Microsoft's gradient boosting decision tree implementation, introduced by Ke and colleagues in 2017, that grows trees leaf-wise and bins features into histograms for speed. On large datasets it is much faster than XGBoost while retaining strong predictive accuracy.	Online learning is a machine learning paradigm in which a model is updated incrementally as each new data point arrives, rather than being trained once on a fixed dataset. It is essential when data streams continuously, storage is limited, or the underlying distribution shifts over time. Theoretical performance is measured by cumulative regret relative to the best fixed predictor in hindsight.