Prophet-Elman残差改正电离层TEC短期预报模型

针对电离层TEC非线性、非平稳的特点，建立一种基于Prophet与Elman神经网络相结合的残差改正电离层短期预报模型。利用该模型对IGS提供的不同太阳活动程度期间的电离层TEC时间序列进行建模预报。结果显示，改正模型能够反映电离层TEC的变化特征，在太阳活动低年和太阳活动高年预报的平均相对精度分别为92.9%和92.2%，均方根误差分别为0.94 TECu和1.77 TECu，精度较Prophet-Elman模型及单一Elman模型有显著提高。

Short Term Prediction Model of Ionospheric TEC Based on Residual Correction of Prophet-Elman

According to nonlinear and nonstationary characteristics of ionospheric TEC, we establish a residual correction ionospheric prediction model based on the combination of Prophet and Elman neural network. The model is used to model and forecast the ionospheric TEC time series provided by IGS during different degrees of solar activity. The results show that the modified model can reflect the variation characteristics of ionospheric TEC. The average relative accuracy of the model in low solar activity years and high solar activity years is 92.9% and 92.2%, and the root mean square error is 0.94 TECu and 1.77 TECu, respectively. The accuracy of the modified model is significantly higher than that of the Prophet-Elman model and the single Elman model.