变形观测数据时间序列建模中的几个问题

针对在变形观测数据时间序列建模中所遇到的问题,对变形观测数据时间序列建模中的数据预处理、模型 选择、模型定阶与系统稳定性检验等问题进行了研究,提出了分析数据趋势项提取的AR模型方法,编写了C语言 计算程序,实现了样条函数插值方法、周期项提取的差值法,建立ARMA分析模型,对变形预测数据与实际观测数 据进行了比较。结果表明:该方法的正确性以及时间序列分析方法在变形数据处理与分析中的适用性与可行性。

Several Important Problems in Time Series Modelling of Deformation Measurement Datum

According to the real problems that are encountered in setting the time series model, this paper studies such problems as the issues of deformation datum pre-proposing, model selecting, model setting and dynamic system stability judging. The methods to calculate the trend datum, periodical datum of deformation are put forward, and the C-programs of ARMA model to analyze the real problem is developed. Comparison between the surveying datum and calculated ones shows the correctness and availability of time series analysis in the data analysis of deformation survey.