摘要
提出了一种新型的广义径向基函数(RBF)神经网络,并研究了该网络的学习方法。不同于传统三层结构的RBF网络,广义RBF网络增加了基函数输出加权层,并在输出层采用超曲面去逼近任意的非线性曲面。实例仿真结果表明,与传统的RBF网络相比,该网络具有良好的逼近性能,收敛速度快,可逼近任意多变量非线性函数。
A new type of general radial basis function neural network is proposed, and its training method is investigated. Unlike the traditional three-layer RBF network, the basis function output weight layer is added, and super curve is used to approximate any nonlinear curve surface. The simulation results of a function approximation show that compared with the traditional RBF neural network, this network has better approximation performance and faster convergence, and can approximate any multivariable non-linear functions.
出处
《计算机工程与设计》
CSCD
北大核心
2007年第12期2911-2913,共3页
Computer Engineering and Design
基金
北京化工大学青年教师自然科学基金项目(QN0408)
关键词
径向基函数神经网络
网络结构
学习方法
模式识别
仿真研究
radial basis function neural network
network structure
learning method
model identification
simulation research