| 基于非监督权值预学习加速网络的训练 |
| |
| 引用本文： | 方斌,朱明星.基于非监督权值预学习加速网络的训练[J].信号处理,2000(Z1). |
| |
| 作者姓名： | 方斌 朱明星 |
| |
| 作者单位： | 安徽大学自动化系 |
| |
| 摘 要： | 本文提出一种基于非监督学习的Ｏｊａ规则和有监督学习的ｄｅｌｔａ规则的快速学习算法．用改进后的结构和学习方法初始化前馈神经网络的权值；使该算法性能显著提高．最后本文通过仿真实例验证了该方法．
|
| 关 键 词： | 神经网络 Oja规则 delta规则 权值初始化 收敛速度 |
Accelerating the Training of Feedforward Neural Networks Based on Unsupervised Learning Principle Initializing Synaptic Weights |
| |
| Fang Bin Zhu Minxing.Accelerating the Training of Feedforward Neural Networks Based on Unsupervised Learning Principle Initializing Synaptic Weights[J].Signal Processing,2000(Z1). |
| |
| Authors: | Fang Bin Zhu Minxing |
| |
| Affiliation: | Dep of Automation Anhui University |
| |
| Abstract: | In this paper, a fast learning algorithm was proposed on the Oja's rule of unsupervised learning principle and delta rule of supervised learing principle The synaptic weights of the networks is initiated with them. It is advanced notability for the networks learing , At hot the simulation is given to show the properties of the method |
| |
| Keywords: | Neural Networks Oja's Rule Delta Rule Initialization of the synaptic Weights Convergent Speed |
| 本文献已被 CNKI 等数据库收录！ |
