期刊文献+

多层人工神经网络合理结构的确定方法 被引量:36

Rational Structure of Multi-Layer Artificial Neural Network
下载PDF
导出
摘要 隐层结构如何选择是多层人工神经网络应用中一个关键问题·基于多层神经网络优化算法原理和非线性方程理论,建立了多层神经网络计算输出和理想输出关系的非线性方程组,分析了权阈变量、标准样本数量和输出层单元数量的内在关系,给出隐层层数和每个隐层单元数量选取应该满足的基本条件·提出多层神经网络合理结构,即隐层层数和每个隐层单元数量选取的一般原则,给出隐层结构定量求解的直接计算方法和间接优化计算方法·对具体算例进行了合理结构分析,通过神经网络优化算法对多种结构组合比较,表明所提出的合理结构分析方法的正确性·这种方法将为多层神经网络在工程应用中如何选取合理结构提供理论依据和选取有效方法· Selecting rational hidden structure is very important question on multilayer neural network in application. Applying optimum algorithm of neural network and algebra equation theory, nonlinear equations expressing the relationships between computing output and ideal output of neural network were developed. The immanent connection of weights and threshold variables, standard sample numbers and node numbers of output layer are suggested. Hidden structure basis equation is given. A kind of analyzing method of rational construct of multilayer neural network is proposed. Direct and indirect computing method were studied to compute quantitatively the numbers of hidden layers and the unit numbers per hidden layer. The analysis validity was showed through comparison of rational structure with various possible structures of neural network examples.
出处 《东北大学学报(自然科学版)》 EI CAS CSCD 北大核心 2003年第1期35-38,共4页 Journal of Northeastern University(Natural Science)
基金 辽宁省博士起动基金资助项目(2001102017).
关键词 多层人工神经网络 隐层结构分析 隐层层数 隐层单元数量 非线性方程组 优化算法 multi-layer neural network structure of hidden layer hidden layers numbers unit numbers per hidden layer nonlinear equations optimum algorithm of multi-layer neural network
  • 相关文献

参考文献9

  • 1Rumelhart D E, Hinton G E,Williams R J. Learning Internal representations by error propagation in parallel distributed processing[M]. Gambridgef:MIT Press, 1986.318-362.
  • 2Huang S H, Zhang H C. Artificial neural network in manufacturing:concepts,applications,and perspectives[J]. IEEE Transaction on Components,Packaging and Manufacturing Technology, 1994,17(2):212-228.
  • 3Rumelhart D E, Hinton G E. Learning internal representations by back-propagation error[J]. Nature,1986, 323(9):533-536.
  • 4Magnitskii N A. Some new approaches to the construction and learning of artificial neural networks[J]. Computational Mathematics and Modeling, 2001,12(4):293-304.
  • 5Guh R S, Tannock J D T. A neural network approach to characterize pattern parameters in process control charts[J]. Journal of Intelligent Manufacturing, 1999,10(5):449-46.
  • 6Alexander G P,Bentito F A. An accelerated learning algorithm for multi-layer perceptron network[J]. IEEE Trans on Neural Networks, 1994,5(3):493-497.
  • 7侯祥林,陈长征,虞和济,王铁光,纪盛青.神经网络权值和阈值的优化方法[J].东北大学学报(自然科学版),1999,20(4):447-450. 被引量:49
  • 8侯祥林,张春晖,徐心和.多层神经网络共轭梯度优化算法及其在模式识别中的应用[J].东北大学学报(自然科学版),2002,23(1):20-23. 被引量:16
  • 9Auziner W, Stetter H J. An elimination algorithm for the computation of all zeros of a system of multivariate polynomial equations[J]. Series of Number Math, 1988,(86):11-30.

二级参考文献7

  • 1虞和济.机械设备故障诊断的人工神经网络识别法[J].机械强度,1995,17(2):48-54. 被引量:15
  • 2侯祥林.非线性系统故障的分形和神经网络智能诊断方法研究[M].沈阳:东北大学,1990.60-69,102-103.
  • 3Huang S H,IEEE Trans Comput Packag Manuf Tech,1994年,17卷,2期,212页
  • 4施鸿宝,神经网络及其应用,1993年,38页
  • 5焦李成,神经网络计算,1993年,151页
  • 6崔华林,机械优化设计方法与应用,1989年,63页
  • 7侯祥林,陈长征,虞和济,王铁光,纪盛青.神经网络权值和阈值的优化方法[J].东北大学学报(自然科学版),1999,20(4):447-450. 被引量:49

共引文献55

同被引文献239

引证文献36

二级引证文献203

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