TaylorBoost: First and second-order boosting algorithms with explicit margin control
TaylorBoost: First and second-order boosting algorithms with explicit margin control
复制标题
DOI：
10.1109/cvpr.2011.5995605
复制
发表时间：
2011-06
期刊：
影响因子：
--
通讯作者：
M. Saberian;Hamed Masnadi-Shirazi;N. Vasconcelos
中科院分区：
文献类型：
--
作者：
M. Saberian;Hamed Masnadi-Shirazi;N. Vasconcelos
提出了一类新的Boosting算法，记为Taylor-Boost。它支持损失函数和一阶或二阶优化的任何组合，并包括经典算法，如AdaBoost，Question-Boost或LogitBoost作为特例。它对规范损失集的限制使得有可能具有显式裕度控制的提升算法。一个新的大家庭的损失与此属性，基于零均值随机变量的累积分布，然后提出。在这个家庭中的一个新的损失函数，拉普拉斯损失，最后得出。该损失和二阶TaylorBoost的组合产生具有显式裕度控制的升压算法。
A new family of boosting algorithms, denoted Taylor-Boost, is proposed. It supports any combination of loss function and first or second order optimization, and includes classical algorithms such as AdaBoost, Gradient-Boost, or LogitBoost as special cases. Its restriction to the set of canonical losses makes it possible to have boosting algorithms with explicit margin control. A new large family of losses with this property, based on the set of cumulative distributions of zero mean random variables, is then proposed. A novel loss function in this family, the Laplace loss, is finally derived. The combination of this loss and second order TaylorBoost produces a boosting algorithm with explicit margin control.