一种改进Elman神经网络的电离层TEC预报方法

A prediction model of ionospheric TEC based on improved Elman neural network

为了建立更高精度的电离层TEC预报模型,利用IGS数据中心提供的平静期与磁暴期电离层TEC原始序列,提出基于奇异谱分析法(SSA)与Elman神经网络结合的电离层TEC预报模型。实验结果表明,在电离层平静期TEC的预报精度上,SSA-Elman组合模型的精度更加稳定,预测残差值在2 TECu以内;在电离层磁暴期TEC的预报上,SSA-Elman组合模型较单一Elman模型的预测精度有显著提升,预测值残差在3~4 TECu以内。通过对比分析,文中提出的组合模型方法有助于电离层TEC预报精度的提升。

This paper,in order to establish a more accurate ionospheric TEC prediction model,uses the ionospheric TEC data during the quiet period and the magnetic storm period provided by IGS Center,and proposes an ionospheric TEC prediction model based on singular spectrum analysis(SSA)and Elman neural network.The result shows that the accuracy of the SSA-Elman combined model is more stable in the prediction of TEC during the quiet period,and the residual error is less than 2 TECu;the prediction precision of SSA-Elman combined model is much higher than that of single Elman model during the storm time,and the residual precision of prediction value is less than 3-4 TECu.By comparison and analysis,the combined model method proposed in this paper is helpful to improve the forecast accuracy of ionospheric TEC.