期刊文献+

神经网络的本质逼近阶 被引量:14

原文传递
导出
摘要 运用多元函数逼近工具,对三层前向人工神经网络逼近连续和可积函数的本质逼近阶进行了定量研究.证明了当激活函数满足一定条件时,对任意的连续或可积函数,能具体构造有明确隐层单元下界的三层网络使之对被逼近函数任意逼近.给出该类神经网络逼近的上、下界估计和本质逼近阶估计,刻画所构造网络的逼近性能与网络隐层拓扑结构之间的关系.特别地,当被逼近函数为二阶Lipschitz函数时,所建立的神经网络其逼近速度完全取决于被逼近函数的光滑性.所获结果对逼近连续或可积函数类的前向神经网络具体构造及逼近能力刻画有重要的理论指导意义.
出处 《中国科学(E辑)》 CSCD 北大核心 2004年第4期361-373,共13页 Science in China(Series E)
基金 国家"八六三"基金资助项目(20001AA113182) 国家博士后基金资助项目 教育部科技重点资助项目(03412)
  • 相关文献

参考文献21

  • 1Cybenko G. Approximation by superpositions of sigmoidal function. Math Control Signals, System,1989, 2(4): 303-314
  • 2Funahashi K. On the approximate realization of continuous mappings by neural networks. Neural Networks, 1989, 2:183-192
  • 3Hornik K, Stinchombe M, White H. Multilayer feedforward networks are universal approximators. Neural Networks, 1989, 2:359-366
  • 4Ito Y. Approximation of functions on a compact set by finite sums of Sigmoid function without scaling.Neural Networks, 1991, 4:817-826
  • 5Chui C K, Li X. Approximation by ridge functions and neural networks with one hidden layer. J Approx Theory, 1992, 70:131-141
  • 6Leshno M, Lin V Y, Pinks A, et al. Multilayer feedforward networks with a nonpolynomial activation function can approximate any function. Neural Networks, 1993, 6: 861-867.
  • 7Kurkova V, Kainen P C, Kreinovich V. Estimates of the number of hidden units and variation with respect to half-space. Neural Networks, 1997, 10:1068-1078
  • 8Suzuki S. Constructive function-approximation by three-layer artificial neural networks. Neural Networks, 1998, 11:1049-1058
  • 9Maiorov V, Meir R S. Approximation bounds for smooth functions in C(Rd) by neural and mixture networks. IEEETran Neural Networks, 1998, 9:969-978
  • 10Chen X H, White H. Improve rates and asmptotic normality for nonparametric neural network estimators. IEEE Tran Information Theory, 1999, 45:682-691

二级参考文献1

  • 1G. Cybenko. Approximation by superpositions of a sigmoidal function[J] 1989,Mathematics of Control, Signals, and Systems(4):303~314

共引文献49

同被引文献131

引证文献14

二级引证文献73

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部