摘要
提出了一种优化选择径向基神经网络数据中心的算法,该算法结合了Kohonen网络的模式分类能力,将初步分类结果用作RBFNN的初始数据中心,然后采用OLS算法进行优化,对比仿真实验表明该算法效果比单独使用OLS算法生成的RBFNN性能更好。
This article proposes a kind of algorithm of optimized selection radial basis function neural network data.Combining the pattern classification ability of the Kohonen network,this algorithm used the initial classification results as the initial data center of RBFNN,and then optimized the OLS algorithm.The simulation experiments showed that this algorithm had better effect than that of using OLS algorithm only.
出处
《微型电脑应用》
2008年第9期10-13,4,共4页
Microcomputer Applications
关键词
RBFNN
径向基中心
KOHONEN网络
OLS方法
Radial basis function neural network
Basis center
Kohonen network
Orthogonal Least Squares Method (OLS)
