Investigation the Possibility of Increasing the Accuracy of Predicting the Critical Frequency of the F2 Layer Using Artificial Neural Networks

Abstract

The article presents the results of evaluating the effectiveness of using artificial neural networks to predict the critical frequency of the ionosphere. When calculating the quantitative values of the errors, the vertical sounding database were used for 18 years from 2002 to 2019. Modeling of the critical frequency is based on the recommendations of the radio sector of the International Telecommunication Union (ITU-R). In order to predict the ionosphere in the ITU-R model, the more accurate index F10.7 was used instead of the Wolf number. The data obtained make it possible to determine the advantages of the method for correcting model calculations throughout the entire 11-year solar cycle.