Prediction of foF2 frequency based on BP neural network with single point extrapolation

Abstract

In this paper, a BP neural network prediction model based on single point extrapolation is used to improve the prediction accuracy of ionospheric foF2. According to the optimization of parameters of input layer analysis to build appropriate training samples, by trial and error method to determine the hidden layer structure, using the experiment and simulation to train network to forecast, adopt the method of average test analysis prediction error, the final analysis of single point extrapolation and prediction of IRI2016 prediction model of BP neural network.