| BP学习算法的局限性与对策 |
| |
| 引用本文： | 林雅梅,郑文波.BP学习算法的局限性与对策[J].福州大学学报(自然科学版),1996(6):13-17. |
| |
| 作者姓名： | 林雅梅 郑文波 |
| |
| 作者单位： | 福州大学计算机科学与技术系 |
| |
| 摘 要： | 研究采用多神经网络组合结构、改进ＢＰ学习算法和训练策略等３个对策来克服ＢＰ学习算法的局限性．手写数字识别的实验研究表明，这些对策效果显著．
|
| 关 键 词： | BP算法 神经网络 组合结构 |
Limitations and Countermeasures of BP Learning Algorithm |
| |
| Lin Yamei,Zheng Wenbo.Limitations and Countermeasures of BP Learning Algorithm[J].Journal of Fuzhou University(Natural Science Edition),1996(6):13-17. |
| |
| Authors: | Lin Yamei Zheng Wenbo |
| |
| Institution: | Lin Yamei;Zheng Wenbo(Department of Computer Science and Technology,Fuzhou University,Fuzhou,350002) |
| |
| Abstract: | We present three countermeasures of using a composite structure of the mutiple neural network,improving BP learning algorithm and perfecting training tactics.By these countermeasures,the limitations of BP learning algorithm are overcome and good results in the recognition of hand-written digits arc obtained. |
| |
| Keywords: | BP algorithm neural network composite structure |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《福州大学学报(自然科学版)》浏览原始摘要信息 |
| 点击此处可从《福州大学学报(自然科学版)》下载免费的PDF全文 |
