ML based model for Detection of Ionospheric Scintillations using PolaRx5S data

In this paper, machine learning based techniques namely Support Vector Machine (SVM), Random Forest (RF) are used for detection of the scintillations and compares their detection performance with classical models namely Semi Hard and Hard detection rules. For this, we used one week data from PolaRx5S scintillation monitoring receiver which receives huge amount of data from 72 to 95 visible satellites of seven constellations. The performance of ML algorithms is compared with traditional rules by using confusion matrix and statistical parameters. The overall detection performance of ML algorithms is considerably improved and is useful for aviation and communication applications.