A Model-Assisted Combined Machine Learning Method for Ionospheric TEC Prediction

In order to improve the prediction accuracy of ionospheric total electron content (TEC), a combined intelligent model (MMAdapGA-BP-NN) based on multi-mutation, multi-cross adaptive genetic algorithm (MMAdapGA) and back propagation neural network (BP-NN) was proposed. The combines international reference ionosphere (IRI), statistical machine learning (SML), BP-NN, MMAdapGA. Compared with IRI, SML-based, other models, MMAdapGA-BP-NN has higher more stable effect. Taking Athens station in Greece as an example, root mean square errors (RMSEs) 2015 2020 are 2.84TECU 0.85TECU, respectively, 52.27% 72.13% lower than IRI model. single model, reduced RMSE by 28.82% 24.11% 2020, respectively. Furthermore, compared optimized mutation algorithm, fewer iterations ranging from 10 30. results show that effect stability proposed have obvious advantages. As result, could be extended alternative scheme for parameters.