Learning algorithms, in statistical learning theory, can be described as greedy procedures for stagewise minimization of an appropriate cost funcation.

• 摘要学习理论中，许多学习算法可以描述为一个最小化适当损失函数的贪婪过程。