Method for detecting anomalous changes in the speed of arrival of cosmic rays to the Earth using machine learning

• DOI

An automated method for detecting anomalous changes in the speed of arrival of galactic cosmic rays to the Earth is proposed. The method is based on the use of packet wavelet decompositions and adaptive stochastic thresholds. The determination of thresholds is performed with a given confidence probability based on the α-quantiles of Student's distribution. The construction of wavelet packet trees and the use of thresholds make it possible to suppress natural and hardware noise in cosmic ray data and to detect local anomalous variations of various shapes and durations. To estimate the power of the selected anomalous variations, a discrete wavelet transform is used. The operations of the method are described and a scheme for its implementation is proposed. Using neutron monitor data as an example, it is shown that the proposed method provides efficient detection of anomalies in cosmic rays during increased solar activity and magnetic storms.

Liss, A., Mandrikova, B. Method for detecting anomalous changes in the speed of arrival of cosmic rays to the Earth using machine learning. In: Problems of Geocosmos—2022. ICS 2022. Springer Proceedings in Earth and Environmental Sciences. Springer, Cham. 2023, pp. 419-431. Doi: https://doi.org/10.1007/978-3-031-40728-4_32