RSFM网络及其在信号滤波中的应用

介绍了一种适用于模式识别的新型神经网络模型--局部有监督特征映射（Reglonal Supervised Feature Mapping,RSFM）网络,描述了该网络的拓扑结构和学习算法,研究了网络的基本性能,最后将其应用到了信号滤波中.理论研究和仿真实验表明,该网络结构简单、算法简洁,收敛速度快、识别精度高,适用于需要大样本训练、随机干扰严重的复杂模式的分类与识别.

Regional supervised feature mapping network and its application in signal filtering

A new neural network named regional supervised feature mapping(RSFM) network was introduced in this paper.The topology structure and training algorithm of this network were represented,and its basic performance was studied.Signal filtering based on this new network was employed at last.Theoretical reasoning and numerical simulation results show this model possesses many advantages,such as simple structure and training algorithm,quick training and good recognition performance.It has been proven to be suitable for pattern classification and recognition,especially for the situation which need large training sample and associate with serious random disturbance.