二重趋势时间序列的灰色组合预测模型  

Grey combined prediction models for double trend time series

作　　者：宋仙磊[1] 刘业政[1] 陈思凤[1] 许波[1] 

机构地区：[1]合肥工业大学管理学院过程优化与智能决策教育部重点实验室,合肥230009

出　　处：《计算机工程与应用》2011年第8期115-117,142,共4页Computer Engineering and Applications

基　　金：国家高技术研究发展计划(863)(No.2008AA042302)~~

摘　　要：神经网络、ARIMA等广泛应用于具有趋势变动性和周期波动性的二重趋势特征的时间序列预测,而这些单一的模型难以达到满意的预测效果。提出一种针对该特征的灰色组合模型,其基本思想是:从二重趋势时间序列中分离趋势变动项和周期波动项后,用灰色G(1,1)模型预测趋势变动项,引用BP网络和ARIMA的组合模型预测周期波动项,用乘积模型合成两部分预测值为灰色组合模型的最终预测值。实验表明:该灰色组合模型适应了二重趋势时间序列的特征,具有很好的预测效果。Neural networks,Autoregressive Integrated Moving Average Model（ARIMA） and other methods have been used comprehensively to predict double trend time series,which have trend change nature and periodic fluctuation characteristic.The prediction performances by a sole model are as still far from satisfactory.Taking into consideration this characteristic,a grey combined model is proposed.After the trend and period are separated from double trend time series,the trend is forecasted by grey G（1,1） model,the period is predicted by the combined model which is comprised by BP neural network model and ARIMA model,the two parts predictive value are combined to the final predictive value by the multiplicative model.Experiments show that the grey combined model is adapted to the characteristics of double trend time series,and it has got the best prediction performance.

关 键 词：灰色理论 反向传播(BP)神经网络 自回归滑动平均(ARIMA) 二重时间序列 预测 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]