郑州大学学报（工学版）2013，Vol.34Issue(3)：94-97,120,5.DOI:10.3969/j.issn.1671-6833.2013.03.023
基于小波和神经网络模型的邮电业务总量预测
Business Total of Posts and Telecommunications Prediction Based on Wavelet and Artificial Neural Network Model
摘要
Abstract
This paper introduces the concept of Maximal Overlap Discrete Wavelet Transform (MODWT) for business total of posts and telecommunications prediction.The characteristic of MODWT multi-resolution analysis is used to decompose the business total of posts and telecommunications.Then the smooth term and details term are separated from business total of posts and telecommunications series.The smooth term and details term are modeled and predicted by applying wavelet neural network model.Finally,the predictive value of business total of posts and telecommunications series is obtained by restructuring.The results tested with data indicate:this forecasting method introduced in this paper can realize multistep prediction and the forecasting accuracy is higher than the way that business total of posts and telecommunications is predicted just by wavelet neural network model or BP neural network model.关键词
BP神经网络模型/极大重叠离散小波变换/小波神经网络/邮电业务总量Key words
BP neural network model/ maximal overlap discrete wavelet transform/ wavelet neural network/business total of posts and telecommunications分类
管理科学引用本文复制引用
单锐,代海波,刘文..基于小波和神经网络模型的邮电业务总量预测[J].郑州大学学报（工学版）,2013,34(3):94-97,120,5.基金项目
国家自然科学基金资助项目(51175448) （51175448）
河北省教育厅科学研究计划项目(2009159) （2009159）
