Forecasting the Critical Frequency of the Ionospheric F2 Layer by using a Neural Network with the Particle Swarm Optimization Algorithm

Abstract

This paper considers the prediction of the critical frequencies of the ionospheric F2 layer, foF2, by using two models: a backpropagation neural network (BPNN) model and a BPNN combined with particle swarm optimization (BPNN–PSO) model for different states of solar activity: low, medium and high. Nine-year critical frequency data from an ionosonde installed at the Universiti Tun Hussein Onn Malaysia in Johor (1.86° N, 103.80° E) were used. The efficiency of the models in predicting foF2 under different states of solar activity was explored. The output of the models was evaluated using root-mean-square error (RMSE) and mean average percentage error (MAPE). The BPNN–PSO model provided a better result compared with the BPNN model during low, medium and high solar activity. The BPNN–PSO model had RMSE and MAPE equal to 0.50 MHz and 5.27%, respectively, during low solar activity and RMSE and MAPE of 0.32 MHz and 4.16%, respectively, during medium solar activity. In addition, the BPNN–PSO model had the lowest RMSE (0.24 MHz) and MAPE (2.45%) during high solar activity. Overall, the performance of the BPNN–PSO model was higher than that of the BPNN model during any state of solar activity.