利用Rademacher随机变量,本文讨论了学习函数f的亏损函数及f的样本误差的估计问题,给出了f的亏损函数及样本误差的估计,同时也给出了f的亏损函数的期望值的估计,这些估计都是O(m-12),这里m为样本容量.
In this paper,we study the error estimation in learning theory.We apply Rademacher random variables to discuss estimation of the defect function and sample error of learning function f.We obtain an estimation bounds of the defect function and sample error of f,and we also obtain an estimation bounds of the expected value of the defect function of f.These bounds are O(m-12),where m is the sample size.