ИСТИНА

Solar, galactic and magnetospheric energetic particles can create a serious radiation danger during space activities and affect astronauts and equipment, up to and including loss. Anomalous changes in the dynamics of cosmic rays occur during periods of nonstationary solar phenomena and heliospheric disturbances. At the moment, the problem of identifying and forecasting anomalies in cosmic rays is open. We propose an automated method for detecting and identifying non-periodic variations in cosmic rays, based on a combination of nonlinear adaptive schemes and deep learning. The Autoencoder paradigm and orthogonal package decompositions are used. The effectiveness of the method is confirmed by numerical calculations using data from neutron monitors of a network of ground stations.