Selecting the key control parameters for the ionospheric total electron content nowcasting
Nowcasting the dynamics of ionospheric parameters is an actual and at the same time rather complicated task. One of the main issues is the selection of control parameters for constructing accurate predictive model (feature selection). The approach is based on the machine learning technology for this problem solution. The vertical absolute total electron content (TEC) with a time resolution of 30 minutes is used as experimental data. The data were obtained using phase and group measurements of TEC at the mid-latitude IRKJ station (52 N, 104 E) for 2014. The results showed that the key control parameters for TEC nowcasting model are the current value of the TEC, the estimated TEC derivative, the local time, the current values of F10.7 and the SYM/H, and the exponentially weighted moving averaged TEC values with periods of 12 and 24 hours and SYM/H with periods of 24 and 96 hours, as well as previously received data with some lag, such as the vertical TEC with 12 hours lag, F10.7 with 3 and 15 days lags. The proposed empirical nowcasting models are based on parameters selected by recursive selection of characteristics with determination of their significance using random forest and support vectors methods. Using these key parameters, the linear regression model allows obtaining an estimate on the interval of 4-7 hours with RMS ~4.5 TECU. The machine learning methods such as random forest, support vector method and gradient boosting allow to reduce RMS to 3-3.5 TECU. © 2018 Space Research Institute of the Russian Academy of Sciences. All rights reserved.
Библиографическая ссылка
Zhukov A.V., Sidorov D.N., Mylnikova A.A., Yasyukevich Yu.V. Selecting the key control parameters for the ionospheric total electron content nowcasting // Sovremennye Problemy Distantsionnogo Zondirovaniya Zemli iz Kosmosa. Vol.15. №3. 2018. P.263-272. DOI: 10.21046/2070-7401-2018-15-3-263-272
