| 多层前向网络研究进展及若干问题 |
| |
| 引用本文： | 董聪,郦正能.多层前向网络研究进展及若干问题[J].力学进展,1995,25(2):186-196. |
| |
| 作者姓名： | 董聪 郦正能 |
| |
| 作者单位： | 清华大学; 北京航空航天大学 |
| |
| 摘 要： | 本文概述了多层前向网络研究的发展历史，对其中有代表性的若干成就进行了较为系统的介绍和评论，分析了当前研究工作中存在的一些问题，提出了解决这些问题的几种可行方案。在对多层前向网络的有效逼近机理进行深入剖析的基础上，提出了合理的有限规模多层前向网络应当遵循的若干构造原则。
|
| 关 键 词： | 多层前向网络 机理分析 最小二乘逼近 |
ADVANCES AND PROBLEMS IN THE STUDY OFMULTILAYER FEEDFORWARDN EURAL NETWORKS |
| |
| Dong Cong,Li Zhengneng,Xia Renwei,HeQingzhi.ADVANCES AND PROBLEMS IN THE STUDY OFMULTILAYER FEEDFORWARDN EURAL NETWORKS[J].Advances in Mechanics,1995,25(2):186-196. |
| |
| Authors: | Dong Cong Li Zhengneng Xia Renwei HeQingzhi |
| |
| Affiliation: | Beijing University of Aeronautics and Astronauics |
| |
| Abstract: | In this paper,①a brief introduction is given to the history of research on multilayer feedforward neural networks;②a systematic review is given to some important achievements in the field of multilayer feedforward networks;③some existing problems and tendencies in the study of multilayer feedforward networks are pointed out,and a few of feasible methods for solving those questions are suggested;④a systematic analysis of approximation mechanism of feedforward networks with a scale,limited discrete point set ... |
| |
| Keywords: | multilayer feedforward neural networks mechanism analysis least square approximation learning algorithm |
| 本文献已被 CNKI 维普 等数据库收录！ |
| 点击此处可从《力学进展》浏览原始摘要信息 |
|
点击此处可从《力学进展》下载全文 |
