摘要
本文给出了一种五层的模糊神经网络。这种神经网络的特点是训练数据可以从网络的输入和输出两端馈入。网络的学习过程分为参数辨识和结构辨识两个阶段,这样可以减少网络参数调整过程中的相互影响,加快学习过程,仿真结果表明了该学习算法可以获得比其他学习算法更好的辨识效果。
This paper presents a five - layer fuzzy neural network. Its characteristic is training data can be input from both input and output end. The learning processing of the network can be divided into parameter identification and structure identification. This can reduce the effect each other in the processing of parameter adjusting and speed up learning process. Results of simulation show the algorithm is better than others.
出处
《现代机械》
2007年第4期24-25,27,共3页
Modern Machinery
关键词
模糊神经网络
参数辨识
结构辨识
算法
fuzzy neural network
parameter identification
structure identification
algorithm
