![]() |
ИСТИНА |
Войти в систему Регистрация |
ФНКЦ РР |
||
This paper investigates the possibilities of prediction of the time series of hourly average fluxes of relativistic electrons in the outer radiation belt of the Earth 1 to 24 hours ahead. Various machine learning methods were used, namely, multi-layer perceptron (MLP) type artificial neural networks, the decision tree method (random forest), gradient boosting. A comparison of the quality indicators of short-term forecasts with a horizon of one to 24 hours showed that the best result was demonstrated by MLP.