Compară metode

	CatBoost Regularizat aplică controale explicite de regularizare	LightGBM Regularizat
Domeniu	Învățare automată	Învățare automată
Familie	Machine learning	Machine learning
Anul apariției	2018	2017
Autorul original	Prokhorenkova, L., Gusev, G., Vorobev, A., Dorogush, A. V., & Gulin, A. (Yandex Research)	Ke, G. et al. (Microsoft Research)
Tip	Regularized gradient boosting ensemble	Regularized gradient boosting ensemble
Sursa seminală	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.	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.
Denumiri alternative	CatBoost with regularization, regularized categorical boosting, CatBoost L2 regularization, penalized CatBoost	LightGBM with L1/L2 regularization, penalized LightGBM, LightGBM ridge/lasso, regularized LGBM
Înrudite	5	5
Rezumat	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.	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.