摘要
针对Jordan神经网络的反馈网络的反馈信息表征能力不强的缺点,提出了一种新的反馈网络模型,对Jordan神经网络的缺点进行了改进,并且对原来的训练学习算法进行了改进,提出了一种提取绝对值最大权的训练学习算法来降低计算复杂性,最终给出了实验结果证明。
Aiming at the shortcomings of the recurrent network recurrent information characterization weak capacity about Jordan neural network, a new model of the recurrent network was proposed for improving the shortcomings about the Jordan neural network, and a a training learning algorithm extracting the greatest absolute value weight was proposed in order to reduce the complexity of calculating. For improving the original training learning algorithm, eventually this paper gives the experimental results to illustrate it.
出处
《贵州大学学报（自然科学版）》
2009年第1期36-39,共4页
Journal of Guizhou University:Natural Sciences
基金
国家自然科学基金资助项目(10671045)
关键词
反馈网络
Jordan神经网络
复杂性
表征能力
recurrent neutral network
jordan neural networks
complexity
characterization capacity
