摘要
本文把联想记忆网络中要全部记住一组训练样本的充要条件的求解等价地转化为一组混合不等式方程组的求解问题，并采用序列最小二乘技术求解。
In this paper, the problem of determining a group ofweights of an associative memory network so as to satisfy the necessary and sufficient conditions for the network to store entirely a given set of training samples is first converted in an equivelent way to that of solving some mixed-type inequality equations. The latter problem is then solved by applying sequential least mean square technique. The network thus obtained is proved, through our numerical experiments, to be of good storage capacity.
出处
《福州大学学报（自然科学版）》
CAS
CSCD
1997年第1期19-23,共5页
Journal of Fuzhou University(Natural Science Edition)
基金
福建省自然科学基金
关键词
最小二乘技术
联想记忆网络
训练样本
associative memory
least mean square
inequality equations
