摘要
在Hebb学习规则的基础上,提出一种设计离散型Hopfield神经网络权值的新方法,用该方法设计的网络能够对多个非正交记忆样本进行正确的回忆.通过理论证明和Matlab仿真,说明了新设计方法的正确性和可行性.
On the base of Hebb learning rule, this paper puts forward a novel method of designing the weight of discrete hopfield neural network. Using this method, the network could recall non-orthogonal memory samples accurately. Through theoretical proof and Mat- lab simulation, the novel method is proved correct and practicable.
出处
《福建师范大学学报(自然科学版)》
CAS
CSCD
北大核心
2012年第3期38-42,共5页
Journal of Fujian Normal University:Natural Science Edition
基金
福建省自然科学基金资助项目(2008J0211)