So sánh phương pháp
Xem các phương pháp đã chọn cạnh nhau; những hàng khác biệt được làm nổi bật.
| CatBoost Chính quy hóa× | Gradient Boosting× | LightGBM Chính quy hóa× | |
|---|---|---|---|
| Lĩnh vực | Học máy | Học máy | Học máy |
| Họ | Machine learning | Machine learning | Machine learning |
| Năm ra đời≠ | 2018 | 2001 | 2017 |
| Người khởi xướng≠ | Prokhorenkova, L., Gusev, G., Vorobev, A., Dorogush, A. V., & Gulin, A. (Yandex Research) | Friedman, J. H. | Ke, G. et al. (Microsoft Research) |
| Loại≠ | Regularized gradient boosting ensemble | Ensemble (sequential boosting of decision trees) | Regularized gradient boosting ensemble |
| Công trình gốc≠ | Prokhorenkova, L., Gusev, G., Vorobev, A., Dorogush, A. V., & Gulin, A. (2018). CatBoost: unbiased boosting with categorical features. Advances in Neural Information Processing Systems, 31. link ↗ | Friedman, J. H. (2001). Greedy Function Approximation: A Gradient Boosting Machine. Annals of Statistics, 29(5), 1189–1232. DOI ↗ | 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. link ↗ |
| Tên gọi khác | CatBoost with regularization, regularized categorical boosting, CatBoost L2 regularization, penalized CatBoost | Gradient Boosting (GBM), GBM, gradient boosted trees, gradient boosting machine | LightGBM with L1/L2 regularization, penalized LightGBM, LightGBM ridge/lasso, regularized LGBM |
| Liên quan | 5 | 5 | 5 |
| Tóm tắt≠ | Regularized CatBoost applies explicit regularization controls — L2 leaf regularization, tree depth constraints, shrinkage rate, and model size penalties — on top of CatBoost's ordered gradient boosting framework, reducing overfitting while retaining CatBoost's native handling of categorical features and its low prediction latency on tabular datasets. | 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. | Regularized LightGBM applies L1 (lasso) and L2 (ridge) penalty terms to the leaf weight objective of LightGBM — Microsoft's highly efficient gradient boosting framework — to control model complexity, reduce overfitting, and improve generalization on tabular classification and regression tasks with high-dimensional or noisy feature sets. |
| ScholarGateBộ dữ liệu ↗ |
|
|
|