欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
从学习算法自身的角度研究算法的泛化能力，给出了一个新的学习算法稳定性框架．在该框架下，我们通过独立随机变量函数和的尾概率估计集中不等式，获得了一个更加严格的学习算法泛化误差指数界估计．
英文摘要：
This paper researches on the generalization ability of algorithms from the perspective of learning algorithm providing a new framework of stability in learning algorithm. In this framework, a more rigid estimation of generalization error is got through a concentration inequality which can be used to derive exponential inequalities for functions of independent random variables.
同期刊论文项目
同项目期刊论文
A comparative study for content-based dynamic spam classification using four machine learning algori
A critical global convergence analysis of recurrent neural networks with general projection mappings
Study on the interrelation of efficient portfolios and their frontier under distribution and various
期刊信息
位置：