Evaluation and Comparison of Machine Learning Algorithms for Solar Flare Class Prediction

Due to powerful and sudden release of magnetic energy, solar flares pose a great threat to technological systems in space as well as on ground. This study explores the potential of machine learning in predicting the class of solar flare namely- B: weakest flare, C: weak flare, M: strong flare, X: strongest flare and N: no flare. The study aims to apply machine learning algorithms on SDO/HMI vector magnetic field data obtained by the Space-weather HMI Active Region Patches (SHARP) and assess the performance of different machine learning algorithms namely Logistic Regression, K-Nearest Neighbor (KNN), Support Vector Machine (SVM), Decision tree, Random Forest, Adaptive Boosting and Gradient Boosting with respect to different performance metrics. Of all applied algorithms, Random Forest was found to outperform other classification algorithms.