动态全参数自调整BP神经网络模型的改进

中国管理科学 ›› 2004, Vol. ›› Issue (6): 68-72.

动态全参数自调整BP神经网络模型的改进

李晓峰, 徐玖平

四川大学工商管理学院成都 610064

收稿日期:2003-03-20 修回日期:2004-06-20 出版日期:2004-12-28 发布日期:2012-03-07
基金资助:
国家自然科学基金资助项目(70171021);四川大学哲学、社科研究青年启动项目资助.

The Improvement of Forecasting Model Based on BP Neural Network of Self-Adjusted All Parameters

LI Xiao-feng, XU Jiu-ping

School of Business and Administration, Sichuan University, Chengdu 610064, China

Received:2003-03-20 Revised:2004-06-20 Online:2004-12-28 Published:2012-03-07

摘要/Abstract

摘要： 针对BP网络存在的缺点,有多种改进方法。本文在文献[14]的基础上,从算法和网络结构设计方面又进行了综合改进,这不仅加快了网络的收敛速度,而且优化了网络的拓扑结构,从而增强了BP神经网络的适应能力。将新改进的BP网络应用于我国能源消费预测,取得了令人满意的效果。

关键词: 人工神经网络, BP算法, 网络结构, 自组织算法

Abstract: In accordance with the shortcoming that BP neural network exists,there are many kinds of improvement methods.This paper has carried on the synthetical improvement at the aspect of algorithm and the network structure design on the foundation of document^[3] work.It not only improves the rate of studying but also improves the network structure,and increases the adaptive ability of BP neural network.Application to forecasting China’s energy consumption has acquired good impact.

Key words: artificial neural networks, BP algorithm, network structure, GMDH

中图分类号:

C931

李晓峰, 徐玖平. 动态全参数自调整BP神经网络模型的改进[J]. 中国管理科学, 2004, (6): 68-72.

LI Xiao-feng, XU Jiu-ping. The Improvement of Forecasting Model Based on BP Neural Network of Self-Adjusted All Parameters[J]. Chinese Journal of Management Science, 2004, (6): 68-72.

参考文献

[1] Hagan M T, Menhaj M B. Training feed forward networks with Marquart algorithm [J]. IEEE Transaction on Neural Networks, 1994,5(6): 989-993.
[2] Hecht-Nielsen R. Theory of back propagation neural network[M]. Proc of IJCNN, 1989,1:593-603.
[3] Sawles G S, Yoon Y, Applying artificial neural networks to investment analysis [J]. Financial Analysis Journal, 1992,Septenbet/Octombet: 78-80.
[4] Anderson J A. An introduction to neural networks[M]. The MIT Press, 1995.
[5] 张立明.人工神经网络的模型及其应用[M].上海:复旦大学出版社,1995:1-92.
[6] Hagan M T, Menhaj M B. Training feed forward networks with Mar quart algorithm[J]. IEEE Trans. on Neural Networks, 1994,5(6): 989-993.
[7] Ergezinger S, Tomsen E. An accelerated algorithm for multiplayer perceptions optimization layer by layer [J]. IEEE Transactions Neural Networks, 2001,(6): 31-42.
[8] 胡守仁,余少波,戴葵.神经网络导论[M].国防科技大学出版社,1993.
[9] 王文剑.一种确定网络结构的自适应方法[C].中国人工智能学会第八届年会论文集,1994.
[10] 胡建军,等.BP网络的权值诱导与层次训练算法[J].计算机科学,1998,25(1):60-63.
[11] 何耀华,夏志忠.BP网络的快速自适应学习算法[J].系统工程理论与实践,2000,20(1):56-63.
[12] 高大启.有教师的线性基本函数前向三层神经网络结构研究[J].计算机学报,1998,21(1):80-85.
[13] 李士勇.模糊控制、神经控制和智能控制论[M].哈尔滨:哈尔滨工业大学出版社,2001.
[14] 李晓峰.动态全参数自调整BP神经网络预测模型的建立[J].预测,2001,20(3):69-71.
[15] Hema Rao and Alexey G. Inductive Learning Algorithms for Complex System modeling[M]. CRC Press, Inc. 1994.
[16] Farlow S J. Self-organizing Method in Modeling [M].Marcel Decker,New York, 1996.