组合模型的电离层总电子短期预报研究  

Research on combined model in short term ionospheric total electron prediction

作　　者：王建敏[1] 唐彦 吕楠 李特 WANG Jianmin;TANG Yan;LYU Nan;LI Te(School of Geography,Liaoning Liaoning Technical University,Fuxin,Liaoning 123000,China)

机构地区：[1]辽宁工程技术大学测绘与地理科学学院,辽宁阜新123000

出　　处：《测绘科学》2022年第4期34-43,67,共11页Science of Surveying and Mapping

基　　金：国家自然科学基金项目(41474020)。

摘　　要：针对电离层总电子含量(TEC)是受到众多影响因素的非平稳性、非线性时间序列的问题,该文提出一种基于小波分解与埃尔曼(Elman)神经网络模型和差分自回归移动平均(ARIMA)模型组合的方法。利用db4小波对电离层TEC样本序列分解得到低频信息和高频信息,对高频信息采用ARIMA模型进行预报,对低频信息采用Elman神经网络模型进行预报,将ARIMA模型和Elman神经网络模型的预报值进行重构,从而得到电离层TEC的预测值。实验表明,组合模型在电离层平静期和活跃期预报的均方根误差分别为0.83、1.08TECu,残差小于1TECu的比例分别为80.28%、68.00%,较单一模型有了大幅的提升。In view of the problem that the total electron content(TEC)of ionosphere is a non-stationary and nonlinear time series affected by many factors,in this paper,a new method based on wavelet decomposition,Elman neural network model and auto regressive integrated moving average(ARIMA)model was proposed.The db4wavelet was used to decompose the ionospheric TEC sample sequence to obtain the low and high-frequency information,wherein that high-frequency information was predicted by an ARIMA model,and the low-frequency information was predicted by an Elman neural network model.Then,the predicted values of the ARIMA model and the Elman neural network model were reconstructed to obtain the predicted values of the ionospheric TEC.The results showed that the root means square error of the combined model in the quiet and active periods of the ionosphere was 0.83TECu and 1.08TECu,and the proportions of residual errors less than 1TECu were 80.28%and 68.00%,which were significantly improved compared to the single model.

关 键 词：电离层 总电子含量 小波分解 埃尔曼神经网络 差分自回归移动平均模型 

分 类 号：P228[天文地球—大地测量学与测量工程]