Calculation Method of Ionospheric Amplitude Scintillation Index Based on Random Forest Regression

According to the information obtained from satellite navigation, the traditional calculation method of the amplitude scintillation index requires too many intermediate variables to estimate, which caused the low accuracy of results. The conventional ionospheric scintillation model and real-time receiver branch information was utilized to produce amplitude-related parameters by the simulated signal technique. The parameters with strong correlation with the scintillation index are selected through feature selection to form a training data set, and the amplitude scintillation index calculation model based on the random forest is trained. The experimental results show that, compared with the traditional method, the calculation model constructed by using the characteristics of random forest with high tolerance to outliers avoids the estimation of too many intermediate variables, and can effectively and accurately calculate the ionospheric amplitude scintillation index.