Predicting ionospheric storm-time fof2 using Support Vector Machine

Using data from two ionosonde stations, Haikou and Chongqing, based on the knowledge gained from the variability of low latitude ionospheric storms, we have developed an empirical model using a new technique, Support Vector Machine, to predict the storm time F2 layer critical frequency, fof2. The model is driven by Dst, AE index and the historical data of fof2. Ionosonde data was sorted as a function of season, and the intensity of the storm, to obtain the corresponding dependencies. It indicated that the model described here can capture the low latitude storm time F2 layer variability at most times.