Study on the forecasting method of relativistic electron flux at geostationary orbit based on support vector machine

In this paper the Support Vector Machine (Classification/Regression) is applied to predict the relativistic electron flux at geostationary orbit. The parameters of model are chosen by Mean Impact Values (MIV), including the electron flux, solar wind speed, solar wind density, Dst index on the previous day and AE index during the preceding two days. This model forecasts the level of relativistic electron flux event and the magnitude of electron flux on the coming day. Based on the comparison with original data in 2008, this model can normally categorize active and quite intervals. For predicting the magnitude of relativistic electron flux, the linear correlation coefficient and prediction efficiency is 0.85 and 0.71; and the model can correctly predict the level of energetic electron enhancement event at most of the time (82%). Our result demonstrates this forecasting technique based on SVM is viable and maybe applicable to other subjects.