Total Electron Content PCA-NN Prediction Model for South-European Middle Latitudes

A regression-based model was previously developed to forecast total electron content (TEC) at middle latitudes. We present a more sophisticated using neural networks (NN) instead of linear regression. This regional prototype simulates and forecasts TEC variations in relation space weather conditions. The development consisted the selection best set predictors, NN architecture, length input series. Tests made data from December 2014 June 2018 show that PCA-NN based on simple feed-forward with very limited number (up six) predictors performs better than PCA-MRM uses up 27 predictors. is series obtained GNSS receiver Lisbon airport tested three other locations latitudes Eastern North Atlantic. Conclusions dependence quality longitude latitude are made.