基于Prophet-RF模型的GNSS高程坐标时间序列预测分析

针对GNSS高程坐标时间序列非平稳性与非线性等特点，在深入分析Prophet模型与随机森林（random forest，RF）模型特性的基础上，构建了Prophet-RF组合预测模型，解决了Prophet模型对时间序列非线性部分预测能力较弱的缺陷，且该组合模型具有较强的鲁棒性。本文选用BJFS站高程方向的连续观测数据进行分析，并设计多种组合方案检验组合模型的适用性与精度，实验结果表明，Prophet-RF组合模型较单一的Prophet模型能更好地表现高程坐标时间序列的变化趋势，并得到更高精度的预测数据。

Prediction and Analysis of GNSS Vertical Coordinate Time Series Based on Prophet-RF Model

In view of the characteristics of GNSS elevation coordinate time series,including non-stationarity and nonlinearity,based on the in-depth analysis of the characteristics of the Prophet model and random forest(RF),we construct the Prophet-RF combination forecasting model.The combination model solves the defect of weak predictive ability of Prophet model to the nonlinear part of time series,and has strong robustness.In this paper,the continuous observation data in the elevation direction of BJFS station is selected for analysis,and a variety of combination schemes are designed to test the applicability and accuracy of the combination model.The experimental results show that the Prophet-RF composite model is better than the single Prophet model in representing change trends of the time series of the elevation coordinates,and in getting more accurate prediction data.