Jeff Tessein

NWRA/CoRA
Mentors: KD Leka and Graham Barnes

 

Abstract:

The discriminant analysis is a statistical method used for discriminating groups. We apply this method to solar flare forecasting with two possible cases (flaring and non-flaring). Using data from the Imaging Vector Magnetograph at Mees Solar Observatory, the analysis can compare parameters and find the most powerful predictors of solar flares. If there is an event within a certain arbitrary time interval after the magnetogram is taken, then this interval will be defined as a flaring interval. Conversely, if no flare happens, it is a flare-quiet interval. We analyze what happens to forecasting ability when we change this time interval. We also apply a nonparametric discriminant analysis to the data in an attempt to improve prediction.