| 随机梯度下降法的一些性质(英文) |
| |
| 引用本文： | 汪宝彬,汪玉霞.随机梯度下降法的一些性质(英文)[J].数学杂志,2011,31(6):1041-1044. |
| |
| 作者姓名： | 汪宝彬 汪玉霞 |
| |
| 作者单位： | 1. 中南民族大学数学与统计学学院,湖北武汉,430074
2. 黄石理工学院数理学院,湖北黄石,435003 |
| |
| 基金项目： | Supported by the Scientific Research Fund of Central South University for Nationalities(YZZ09005) |
| |
| 摘 要： | 本文研究了一般核空间下的随机梯度下降法.通过迭代方法,给出了该算法的一些重要性质,这些性质对于研究收敛速度起到至关重要的作用.
|
| 关 键 词： | 梯度下降法 核空间 随机逼近 逼近能力 |
SOME PROPERTIES RELATING TO STOCHASTIC GRADIENT DESCENT METHODS |
| |
| WANG Bao-bin,WANG Yu-xia.SOME PROPERTIES RELATING TO STOCHASTIC GRADIENT DESCENT METHODS[J].Journal of Mathematics,2011,31(6):1041-1044. |
| |
| Authors: | WANG Bao-bin WANG Yu-xia |
| |
| Institution: | WANG Bao-bin~1,WANG Yu-xia~2 (1.School of Mathematics and Statistics,Central South University for Nationalities,Wuhan 430074,China) (2.School of Mathematics and Physics,Huangshi Institute of Technology,Huangshi 435003,China) |
| |
| Abstract: | In this paper,we study the stochastic gradient descent methods in general kernel spaces.By using the iteration method,we provide some important properties for the stochastic gradient descent methods,which are essential to approximate the target function. |
| |
| Keywords: | approximation ability gradient descent methods kernel spaces stochastic approximation |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
