基于RegEM算法的GPS坐标时间序列插值应用分析

将一种基于数据驱动的RegEM算法引入GPS坐标时间序列插值中,分别采用模拟不同比例连续缺失数据与实测含缺失数据,比较RegEM与拉格朗日方法、三次样条方法、正交多项式方法的插值效果与性能。结果表明,对于模拟不同比例连续缺失数据,RegEM算法插值效果均优于传统方法,且在大量数据连续缺失的情况下效果最优;对于实测含缺失数据,RegEM方法插值所得序列保留方差最大化效果最好,约为正交多项式方法的1.17倍、三次样条方法的1.38倍。

Application Analysis of GPS Coordinate Time Series Interpolation Based on RegEM Algorithm

In this paper, we introduce a data-driven RegEM algorithm into GPS coordinate time series. The interpolation effects and performances of RegEM and Lagrange method, cubic spline method and orthogonal polynomial method are compared by simulating continuous missing data of different proportions and measuring missing data respectively. The experimental results show that RegEM interpolation is superior to traditional interpolation methods in simulating continuous missing data with different proportions, and the best result is obtained in the case of continuous missing data. RegEM interpolation method has the best effect in maximizing the retention variance of the sequence with missing data, which is about 1.17 times that of orthogonal polynomial method and 1.38 times that of cubic spline method.