|
|||||
|
|
| 人工神经网络的一种改进的B-P学习算法及其应用 | |
| 引用本文： | 梁民,孙仲康. 人工神经网络的一种改进的B-P学习算法及其应用[J]. 系统工程与电子技术, 1991, 0(3) |
| 作者姓名： | 梁民 孙仲康 |
| 作者单位： | 国防科学技术大学电子技术系(梁民)，国防科学技术大学电子技术系(孙仲康) |
| 摘 要： | 本文首先简要地介绍了人工神经网络(以下简称神经网络)的B—P学习算法,继而分析了B—P算法收敛速度慢的内在原因,讨论一些加速B—P算法收敛的措施,提出了一种改进的B—P学习算法(MB—P)。将这种算法应用于两类含噪飞机图象目标识别系统,并进行了仿真实验。实验结果表明,MB—P学习算法的收敛速度比B—P算法的收敛速度快许多,而且分别用这两种学习算法训练的神经网络对目标具有大致相同的识别率。 |
| 关 键 词： | 神经网络 ~+B—P算法 ~+MB—P算法 模式识别 |
A Modified Back-propagation Learning Algorithm for Artificial Neural Network and Its Application |
|
| Liang Min and Sun ZhongkangDept. of Electronic Eng. National University of Defence Technology. A Modified Back-propagation Learning Algorithm for Artificial Neural Network and Its Application[J]. System Engineering and Electronics, 1991, 0(3) | |
| Authors: | Liang Min and Sun ZhongkangDept. of Electronic Eng. National University of Defence Technology |
| Affiliation: | Liang Min and Sun ZhongkangDept. of Electronic Eng. National University of Defence Technology |
| Abstract: | In this paper, the back-propagation (B-P) learning algorithm is reviewed at first. Some modifications of the B-P algorithm are discussed. A modified B-P (MB-P) algorithm is proposed and applied to noisy image object recognition. The simulation experiments show that the MB-P algorithm offers a much faster convergence speed, compared with the B-P algorithm. |
| Keywords: | Neural network B-P algorithm MB-P algorithm Image recognition. |
| 本文献已被 CNKI 等数据库收录！ | |