Forecasting the AE index based on neural networks

• DOI

Energy from the interplanetary space, when penetrating into geosphere shells, makes significant impact on the technosphere and human life (disturbances of geomagnetic field and ionosphere, malfunctions of radio communication and navigation, excitation of geo-induced currents). The arrived energy volume can be estimated from AE index. The paper solves the problem of constructing a neural network model for AE index and forecasting its fluctuations. Application of wavelet analysis allowed us to distinguish wavelet decomposition components from the data of interplanetary magnetic field and AE index, to train neural network based on them and to forecast the AE index variation change. The results of network operation during calm and disturbed periods were shown.

Polozov Y. Forecasting the AE index based on neural networks // AIP Conference Proceedings, Vol. 2910, Issue 1, P. 020130. International Conference on SMART Automatics and Energy. Vladivostok, 2023. Doi: https://doi.org/10.1063/5.0166561