Highlights The use of artificial neural networks increases the accuracy of ionosphere modeling. Neural network training can be based on Digital Ionogram Database measurements. Depending on latitudes, the accuracy of forecasting ionospheric parameters changes.
Abstract The article presents a formula apparatus for calculating the main parameters of the F2 layer using empirical coefficients recommended by ITU-R. An approach to recalculation of coefficients using artificial neural networks is proposed. Training, testing and validation of the neural network is based on the use of digital measurements of the ionogram database for the period from 2002 to 2019. The influence of the level of solar activity, geographic location, month and time of day on the efficiency of applying a new set of coefficients is considered. A comparative analysis showed an increase in the forecast accuracy for the critical frequency of the F2 layer by an average of 3.9% and by 15% for the maximum height of the F2 layer.
Improving forecasting accuracy the F2-layer peak characteristics using artificial neural network
Advances in Space Research ; 71 , 8 ; 3373-3381
2022-12-02
9 pages
Article (Journal)
Electronic Resource
English
Improving Artificial Neural Network Based Streamflow Forecasting Models through Data Preprocessing
Springer Verlag | 2021
|Daily Electric Load Forecasting Using Artificial Neural Network
Online Contents | 1995
|Development of Real Time Storm Surge Forecasting Using Artificial Neural Network
HENRY – Federal Waterways Engineering and Research Institute (BAW) | 2014
|River Stage Forecasting Using Artificial Neural Networks
British Library Online Contents | 1998
|