计算机工程与应用 ›› 2011, Vol. 47 ›› Issue (11): 207-210.

• 图形、图像、模式识别 • 上一篇    下一篇

超平面中心的RBF神经网络及其新方法

许亦男,王士同   

  1. 江南大学 信息学院,江苏 无锡 214122
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2011-04-11 发布日期:2011-04-11

Center-planed RBF neural network and its learning algorithm

XU Yinan,WANG Shitong   

  1. School of Information Technology,Jiangnan University,Wuxi,Jiangsu 214122,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2011-04-11 Published:2011-04-11

摘要: 在传统的径向基神经网络框架的基础上,通过引入中心超平面的概念,提出了超平面中心的径向基函数神经网络。在此网络中以点到中心超平面的距离代替传统的径向基神经网络中点到点的距离,其优势在于中心超平面作为数据中心包含了更多原始数据之间的信息。以函数逼近和数据分类的实验为例,证明了超平面中心的径向基神经网络相对于传统的网络有一定的优势。

关键词: 中心超平面, 径向基函数, 函数逼近, 分类

Abstract: A new Radial Basis Function(RBF) neural network learning algorithm named center-planed RBFN is presented by introducing plane into the network.In this neural network,the entity of the data center is changed from being a point to that of being a plane to compute the distance between data and data center.The proposed method has its superiority that more information between the data is contained in the center-plane.Experimental results show that the center-planed RBF neural network can achieve better results than traditional RBF neural network in the function approximation and classification.

Key words: center-plane, Radial Basis Function(RBF), function approximation, classification