Prediction of Ionospheric Total Electron Content Based on Deep Learning

A new prediction model for the total electron content of the global ionosphere is presented by combining the long short-term memory neural network and empirical mode decomposition. The model can predict the total electron content of Nanchang Station up to 24 hours in advance. The experimental results show that the combination model performs well under different experimental conditions, such as the varying modeling time and dependent days. Based on experimental results and assessment criteria, the combination model is feasible and efficient.