Prediction of the Ionospheric foF2 Parameter Using R Language Forecasthybrid Model Library Convenient Time Series Functions

Abstract

Two different approaches of two different time-series models were used to predict the critical frequency (foF2) of the ionospheric F2 layer over Athens (38.0 degrees N, 23.5 degrees E), Greece. Experimental foF2 data were obtained for the Athens station between 2004 and 2018. For the foF2 prediction, the R language forecasthybrid model library time-series convenient functions were used. Root mean square error (RMSE), mean absolute percent error (MAPE) and mean absolute error (MAE) performance metrics were used to examine the performances of the models. According to these tests, the predictions made with the cross-validation error approach are somewhat better than the equal-weighted-prediction approach. Besides, the predictions of foF2 were compared with those of the IRI-2016 model. According to the RMSE, MAE and MAPE analysis, the values of IRI-2016 model were 1.17 MHz, 1.00 MHz and 29.67%, whereas those of EWP and cross-validation errormodel were 0.40, 0.28 and 5.60 and 0.40, 0.27 and 5.56, respectively. Therefore, it can be said that the predictions of these time-series approaches used for the first time for the prediction of ionospheric foF2 are better than those of the IRI-2016. As a result, time-series algorithms, which are now living in the golden age, offer new opportunities for the prediction of ionospheric parameters as in other fields.