具有记忆环节的前向网络时空结构识别方法研究

A STADY ON THE FEED—FORWARD NETWORKS WITH MEMORY NEURON FOR RECOGNIZING SPATIO—TEMPORAL STUCTURE

提出一种具有记忆环节的前向网络时空结构识别方法，并给出了理论推导，其记忆环节使其处理时空模式结构和时序列系统的能力更加明显，通过仿真证明了该方法的有效性。

This paper proposed the feed-forward networks with memory neuron for recoghizing Spatio-Temporal structure.The theory algorithhm was given as well.The ability of process Spation-Temporal struction and time series was increased.The simulation shows the effectiveness of this algoritum.