摘要
提出了一种用于模糊神经网络函数逼近的新方法,将分组的思想与模糊神经网络相结合应用于函数的逼近。实验结果表明,这种方法具有更好的逼近能力和泛化能力,同时在精度给定的情况下训练时间也有所减少。
A new method for function approximation based on packeted fuzzy neural networks was proposed ,which idea of packet and fuzzy neural networks are combined for function approximation. Experimental results show that this method has better approximation capability and generalization ability,while the training time was reduced in the case of precision was given.
出处
《微计算机信息》
2011年第2期173-175,共3页
Control & Automation
关键词
模糊系统
神经网络
模糊逼近
函数逼近
fuzzy system
neural networks
fuzzy approximation
function approximation
