用AR模型建立变形预测模型的研究  

A Study of the Establishment of Deformation Forecasting Model with AR Model

作　　者：许国辉[1] 张新长[2] 

机构地区：[1]广州大学土木工程学院,广东广州510405 [2]中山大学城市与资源规划系,广东广州510275

出　　处：《中山大学学报（自然科学版）》2004年第1期107-109,共3页Acta Scientiarum Naturalium Universitatis Sunyatseni

基　　金：测绘遥感信息工程国家重点实验室开放基金资助项目(WKL(020)0302)

摘　　要：在用时间序列分析方法建立变形测量的预测模型时,基于动态数据系统方法的传统F检验定阶法,由于样本的随机性可能过早地退出对模型的循环检验,从而不能找到合适的预测模型。在探讨如何用AR模型建立变形测量预测模型的基础上,提出了一种改进的建模方法,并在实例的应用中取得了较好的效果。When time series analysis method is used to establish forecasting model of deformation survey, the traditional F-test rank determining method which is based on DDS method may withdraw from the cycle of testing model too early to find the appropriate model because of the random characteristics of the sample. Based on the establishment of forecasting model of deformation survey with AR model, a improved modeling method is proposed.

关 键 词：AR模型 时间序列 预测 阶次 

分 类 号：P228[天文地球—大地测量学与测量工程]