Порівняння методів

	Ансамблевий градієнтний бустинг	CatBoost	Дерево рішень
Галузь	Машинне навчання	Машинне навчання	Машинне навчання
Родина	Machine learning	Machine learning	Machine learning
Рік появи	2001	2018	1984
Автор методу	Friedman, J. H.	Prokhorenkova, L. et al. (Yandex)	Breiman, Friedman, Olshen & Stone
Тип	Ensemble (sequential boosting of decision trees)	Gradient boosting on decision trees	Recursive partitioning (if-then rules)
Основоположне джерело	Friedman, J. H. (2001). Greedy function approximation: A gradient boosting machine. Annals of Statistics, 29(5), 1189–1232.	Prokhorenkova, L., Gusev, G., Vorobev, A., Dorogush, A.V. & Gulin, A. (2018). CatBoost: Unbiased Boosting with Categorical Features. In NeurIPS 2018.	Breiman, L., Friedman, J.H., Olshen, R.A. & Stone, C.J. (1984). Classification and Regression Trees. Wadsworth.
Інші назви	Gradient Boosting Machine, GBM, Gradient Tree Boosting, Stochastic Gradient Boosting	CatBoost (Categorical Boosting), categorical boosting, ordered boosting, kategorik gradyan artırma	Karar Ağacı (Decision Tree), karar ağacı, classification tree, regression tree
Пов'язані	6	5	5
Підсумок	Gradient Boosting is an ensemble method introduced by Jerome Friedman in 2001 that builds a strong predictive model by sequentially adding shallow decision trees, each correcting the errors of the previous ensemble. By framing the problem as gradient descent in function space, it achieves state-of-the-art accuracy on classification, regression, and ranking tasks across tabular data.	CatBoost is a gradient boosting algorithm, introduced by Prokhorenkova and colleagues at Yandex in 2018, that handles categorical variables natively and uses ordered target encoding to avoid label leakage. By building an additive ensemble of trees while permuting the data order at each iteration, it is often superior to XGBoost and LightGBM on category-heavy data.	A Decision Tree is an interpretable classification and regression method, formalised by Breiman, Friedman, Olshen and Stone in their 1984 CART framework, that partitions the data with hierarchical if-then rules. Each split sends observations down one branch or another until a prediction is read off the leaf.