作　　者： ; ;

机构地区： 华中理工大学电子与信息工程系

出　　处： 《自动化学报》 1997年第1期68-72,共5页

摘　　要： 该文提出一种适用于单层神经网络（ＳＮＮ）训练的新颖的广义误差函数，给出了ＳＮＮ新的快速学习算法（ＦＬＡ）．进一步提出了一种广义系统辨识模型，对ＦＬＡ的收敛性进行了理论分析． This paper proposes a new generalized criterion for the training of single layer neural networks, which leads to a novel fast learning algorithm for single layer neural network. In order to analyse the convergent properties of the fast algorithm we developed, a new generalized system identificaton model is also presented. Experiment results show that the fast algorithm proposed in this paper performs the training of neural nework faster than the corresponding learning algorithm given by Karayiannis.

关 键 词： 单层神经网络 快速 算法

领　　域： [自动化与计算机技术] [自动化与计算机技术]