变形监测数据的时间序列分析

Time Series Analysis of Deformation Data

基于统计分析软件SAS的辅助,详细阐述时间序列建模分析的步骤,并结合某一具体工程的变形监测数据建立一个有效的线性模型来拟合该动态序列的发展。实例证明,采用时间序列分析方法建立的模型,其拟合效果较高。5期预测误差绝对值均在0.2 mm以内。最后提出将物理意义与数理统计相结合进行时序分析更有利于模型的建立等一些有益的建议。

Based on statistical software SAS, procedures of the time series modeling analysis were expounded in detail and an effective linear model was established to fit dynamic series development referring to deformation supervision data of certain concrete project. It was proved that the model which was established by using time series method had a higher fitting degree. In addition error absolute value of the fifth period was under 02. mm. Finally some suggests that combining physical meaning and mathernatieal Statistics to conduct time series analysis were much helpful to establish the model were put forward.