首页 | 本学科首页   官方微博 | 高级检索  
     

非线性时间序列建模与预测的神经网络法
引用本文:雷鸣,吴雅,杨叔子.非线性时间序列建模与预测的神经网络法[J].华中科技大学学报(自然科学版),1993(1).
作者姓名:雷鸣  吴雅  杨叔子
作者单位:华中理工大学机械工程一系 (雷鸣,吴雅),华中理工大学机械工程一系(杨叔子)
摘    要:简述了人工神经网络的基本概念及计算特性,提出了用神经网络方法进行非线性时序建模的基本思路,分析了传统的BP学习过程,得出了BP学习算法的改进算法.然后讨论具有一层隐层的前向式神经网络,基于预测误差分析,提出了权值估计方法,使非线性时序建模的神经网络法规范化.

关 键 词:非线性时间序列  人工神经网络  建模  BP学习算法  权值估计

Non-Linear Time Series Modelling and Forecasting Using the Neural Network Approach
Lei Ming Wu Ya Yang Shuzi.Non-Linear Time Series Modelling and Forecasting Using the Neural Network Approach[J].JOURNAL OF HUAZHONG UNIVERSITY OF SCIENCE AND TECHNOLOGY.NATURE SCIENCE,1993(1).
Authors:Lei Ming Wu Ya Yang Shuzi
Affiliation:Lei Ming Wu Ya Yang Shuzi
Abstract:The basic concepts and mathematical calculation characteristics of artificial neural networks are described. The idea of using neural networks for non-linear time series modelling is put forward. The conventional BP learning algorithm is analyzed and some improvements are suggested. The forward neural network of a single hidden layer is discussed and a method of weight estimation based on errors predicted is proposed.
Keywords:non-linear time series  artificial neural network  modelling  BP learning algo rithm  weight estimation  
本文献已被 CNKI 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号