手法を比較

	アンサンブルサポートベクターマシン	バギング（ブートストラップ集約）	ブースティング
分野	機械学習	機械学習	機械学習
系統	Machine learning	Machine learning	Machine learning
提唱年	2000–2003	1996	1990–1997
提唱者	Kim, H.-C. et al.; Dietterich, T. G.	Breiman, L.	Schapire, R. E.; Freund, Y.
種類	Ensemble of SVMs (bagging, voting, or stacking)	Ensemble meta-algorithm (variance reduction via bootstrap aggregation)	Sequential ensemble (iterative reweighting)
原典	Kim, H.-C., Pang, S., Je, H.-M., Kim, D., & Bang, S. Y. (2002). Constructing support vector machine ensemble. Pattern Recognition, 36(12), 2757–2767.	Breiman, L. (1996). Bagging Predictors. Machine Learning, 24(2), 123–140.	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.
別名	Ensemble SVM, SVM ensemble, bagged SVM, SVM committee machine	Bootstrap Aggregating, bootstrap aggregation, bagged ensemble, bagged predictor	AdaBoost, gradient boosting, iterative reweighting ensemble, sequential ensemble
関連	5	5	6
概要	Ensemble Support Vector Machine combines multiple independently trained SVM classifiers or regressors — each fitted on a different data partition, bootstrap sample, or feature subset — and aggregates their outputs via voting, averaging, or stacking. The approach mitigates the high computational cost and sensitivity to kernel hyperparameters inherent in a single large-scale SVM, while improving generalisation on complex or high-dimensional datasets.	Bagging, short for Bootstrap Aggregating, is an ensemble meta-algorithm introduced by Leo Breiman in 1996 that trains multiple copies of a base learner on independently drawn bootstrap samples of the training data and combines their predictions — by averaging for regression or majority vote for classification — to produce a final predictor with substantially lower variance than any single base learner.	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.