方法对比

	鲁棒LightGBM	CatBoost	Huber回归
领域	机器学习	机器学习	统计学
方法族	Machine learning	Machine learning	Regression model
起源年份	2017 (LightGBM); robust variants widely adopted 2018–present	2018	1964
提出者	Ke, G. et al. (LightGBM); robust objectives adapted from Friedman, J. H.	Prokhorenkova, L. et al. (Yandex)	Peter J. Huber
类型	Ensemble (gradient boosted decision trees with robust loss)	Gradient boosting on decision trees	Robust linear regression (M-estimation)
开创性文献	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, 30, 3146–3154.	Prokhorenkova, L., Gusev, G., Vorobev, A., Dorogush, A.V. & Gulin, A. (2018). CatBoost: Unbiased Boosting with Categorical Features. In NeurIPS 2018.	Huber, P. J. (1964). Robust Estimation of a Location Parameter. Annals of Mathematical Statistics, 35(1), 73-101.
别名	Robust LGBM, LightGBM with Huber loss, outlier-resistant gradient boosting, robust gradient boosted trees	CatBoost (Categorical Boosting), categorical boosting, ordered boosting, kategorik gradyan artırma	Huber M-estimator, Huber loss regression, robust regression, Huber Regresyonu
相关	6	5	5
摘要	Robust LightGBM is a gradient boosting framework that pairs Microsoft's highly efficient LightGBM engine with outlier-resistant loss functions — most commonly Huber, quantile, or mean absolute error — so that predictions are not unduly distorted by extreme or erroneous observations. It retains LightGBM's speed and leaf-wise tree growth while providing resistance to heavy-tailed noise in the target variable.	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.	Huber regression is a robust linear regression method, introduced by Peter J. Huber in 1964, that resists the influence of outliers by treating small and large residuals differently. It applies a squared (OLS-like) loss to small residuals and a milder absolute-value loss to large ones, so extreme observations cannot dominate the fit.