Boosting with early stopping: Convergence and consistency

Tong Zhang, Bin Yu

Boosting is one of the most significant advances in machine learning for classification and regression. In its original and computationally flexible version, boosting seeks to mini…