Quantifying the Intensity of Solar Flares by Sigmoid Pattern Analyses

Solar flares are among the leading causes of our planet's magnetosphere disruption; they often induce major geomagnetic storms, and can lead to consequences, chiefly telecommunications and satellite failure. The current method of prediction involves sunspot analyses to locate peak hotspots on the corona, but is time consuming and not entirely accurate. This study assessed the possibility of a correlation between solar sigmoids, a fairly recent discovery, with the intensities of arising flares for implementation in forecasting models. It was hypothesized that at sigmoid rupture, the magnitude of the sigmoids’ magnetic complexity varies proportionately to the intensities of the flares. Solar cycle peak periods were parsed through to demonstrate which properties of the sigmoids demonstrated influence to resulting flare intensities.Using this data, an artificial intelligence-based framework for an automatic classification of solar attributes for flare prediction was presented. A set of twelve classes of sigmoids and sunspots were inputted into a total of 26 machine-learning models. The optimized algorithm showed an accuracy rate of 90.5 within 3 minutes, as opposed to the professional 3 days.