Comparer des méthodes

	Empilement bayésien (Bayesian stacking)	Empilement
Domaine	Apprentissage automatique	Apprentissage automatique
Famille	Machine learning	Machine learning
Année d'origine	2018	1992
Auteur d'origine	Yao, Y.; Vehtari, A.; Simpson, D.; Gelman, A.	Wolpert, D.H.
Type	Bayesian ensemble combination	Ensemble (heterogeneous meta-learning)
Source fondatrice	Yao, Y., Vehtari, A., Simpson, D., & Gelman, A. (2018). Using stacking to average Bayesian predictive distributions. Bayesian Analysis, 13(3), 917–1007.	Wolpert, D.H. (1992). Stacked Generalization. Neural Networks, 5(2), 241–259.
Alias	Bayesian stacking, Bayesian model stacking, stacking with Bayesian weights, predictive distribution stacking	Stacking (Yığınlama — Meta-Öğrenme), stacked generalization, meta-learning ensemble, super learner
Apparentées	6	5
Résumé	Bayesian stacking combines the predictive distributions of several base models by finding non-negative weights that maximise the leave-one-out log predictive score of the mixture. Formalised by Yao, Vehtari, Simpson, and Gelman (2018), it yields a single calibrated predictive distribution that is provably at least as good as any single constituent model under cross-validation.	Stacking, or stacked generalization, is an ensemble method introduced by David Wolpert in 1992 that combines the outputs of several different base models (Level-0) through a separate meta-model (Level-1). Unlike bagging and boosting, it deliberately uses heterogeneous model types, and it is the standard final-stage strategy in Kaggle competitions.