摘要
介绍一种基于全局搜索策略的神经网络学习算法,并对该算法的学习性能进行比较和分析,关于建模过程中网络结构的选择与延迟数识别等方面也做了讨论与分析,给出相应的仿真结果。
A global optimal searching based on NN algorithm has been presented with functional analysis . Some comments about the determination of NN structure and lag space are also discussed, supported with the simulation results by CSTR process.
出处
《武汉科技学院学报》
2003年第1期66-70,共5页
Journal of Wuhan Institute of Science and Technology