论文部分内容阅读
讨论了网络学习过程中的假饱和现象，并给出了克服方法，同时，也讨论了学习样本输入编排机理，给出了一种避免网络学习出现的局部极小的算法。通过实例验证此方法非常有效
This paper discusses the phenomenon of false saturation in the process of network learning and gives some solutions. At the same time, it also discusses the mechanism of inputting orchestration of learning samples, and presents a local minimum algorithm to avoid the network learning. Validation of this method is very effective through examples