摘要
通过对ＲＢＦ网络结构的分析，结合奇异值分解方法，提出了一种ＲＢＦ网络结构优化的标准，并在此基础上实现ＲＢＦ网络结构的优化。
A kind of optimization criterion and an optimization method based on singular value decomposition (SVD) is proposed, which combines SVD and clustering method for selecting significant basis function centers, it can decrease the network complexity and the computation load under almost no lose in network's performance. Numeral simulation illustrated the effective of the proposed method.
出处
《控制与决策》
EI
CSCD
北大核心
1996年第6期667-671,共5页
Control and Decision
基金
国家自然科学基金
关键词
径向基函数网络
最佳化
网络结构
神经网络
radial basis function network (RBFN), singular value decomposition, recursive least squares(RLS), K-mean clustering method
