一种用于径向基函数(RBF)神经网络训练的有效方法  

A Useful Method for the Training of Radial Basis Function Neural Networks

作　　者：孙毅刚[1] 战强[1] 

机构地区：[1]哈尔滨工业大学机器人研究所

出　　处：《哈尔滨工业大学学报》1997年第4期103-106,共4页Journal of Harbin Institute of Technology

摘　　要：提出了一种用于径向基函数（RBF）神经网络训练的新方法,即Gauss-Jordan与求广义逆（genemalinverse）的复合法。仿真结果表明,此方法训练速度快,实时性强,其收敛性和收敛精度均比正交最小二乘算法（OLS）效果好。This paper proposes a new method for the training of Radial Basis Function (RBF) neural networks, which is a compound algorithm of Gauss-Jordan algorithm and General Inverse algorithm. The simulating results show that the training of the method is very quickly convergent, a good real-time way and has better quality than Orthogonal Least Squares Algothrism in Least Mean Square error and convergence.

关 键 词：径向基函数 神经网络 仿真 复合法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]