MEEMD-ARMA残差修正电离层TEC预测模型

MEEMD-ARMA Residual Modified Ionospheric TEC Prediction Model

提出一种改进的集总平均经验模态分解与自回归移动平均模型组合的残差修正模型,并利用2015年不同经纬度同一时段以及相同经纬度不同时段IGS提供的电离层电子总量数据,用3种模型对5d内的数据进行预测。结果表明,改进模型5d内的平均相对精度为96%,而EMD-ARMA模型及ARMA模型分别为94.5%、93%。

To determine the characteristics of total electric contents data and the deficiencies of the original models, a model combining modified ensemble empirical model decomposition and ARMA residual correction model is proposed. We use the total electric contents data provided by IGS in 2015, and use the three models to predict the data within 5 days. The results indicate that the mean relative precision of TEC predicted is 96% by the improved model , while the EMD-ARMA model and ARMA model are 94.5%, 93% accurate, respectively.