高次差法求解时间序列中趋势项的方法研究及其应用

Resolving Time-Series Tendency Part with Higher Difference Method and Its Application

介绍了高次差法求解时间序列中趋势项的基本原理与方法。通过实例对某沉降监测点的观测序列进行分析处理、预报 ,并与最小二乘法进行了比较 ,结果表明该法在建模的精度、抵抗粗差等方面均优于最小二乘法。

The paper introduces the principle and method of resolving time-series tendency part with higher difference method. By analyzing, processing and forecasting some settlement monitoring point's observation and comparing with Least Square method , the result showed that the method is better than the Least Square method in precision of building model and anti-gross-error. The method can apply in resolving time-series tendency part and polynomial regressive analysis.