| 训练多层网络的样本数问题 |
| |
| 引用本文： | 张鸿宾.训练多层网络的样本数问题[J].自动化学报,1993,19(1):71-77. |
| |
| 作者姓名： | 张鸿宾 |
| |
| 作者单位： | 1.北京计算机学院计算机科学系 |
| |
| 摘 要： | 本文分析多层网络的映射增长函数,以经验风险最小和期望风险最小之间的偏差来定义
网络的泛化能力.基于Vapnik-Chervonenkis的事件出现频率一致收敛于其概率的理论,讨
论网络的结构、训练样本数和网络泛化能力间的关系.分析在最不利的情况下为保证一定泛
化能力所需要的训练样本数.
|
| 关 键 词： | 人工神经网络 泛化 Vapnik-Chervonenkis维数 |
| 收稿时间： | 1991-11-11 |
The Sample Size for Training Multi-Layered Neural Network |
| |
| Zhang Hongbin.The Sample Size for Training Multi-Layered Neural Network[J].Acta Automatica Sinica,1993,19(1):71-77. |
| |
| Authors: | Zhang Hongbin |
| |
| Affiliation: | 1.Dept.of Computer Science,Beijing Computer Institute,Beijing |
| |
| Abstract: | Based on the theory of uniform convergence of frequencies of events to their probabilities given by Vapnik and Chervonenkis, the relation-ship among the number of mapping functions, the size of training samples, and the ability of generalization of the multilayered neural network is discussed. The minimum training sample size, which guarantees valid generalization in the worst case, is analysed. |
| |
| Keywords: | Neural network generaliization Vapnik-Chervonenkis dimenstion |
| 本文献已被 CNKI 维普 等数据库收录！ |
| 点击此处可从《自动化学报》浏览原始摘要信息 |
|
点击此处可从《自动化学报》下载全文 |
