Machine Learning Approach for Solar Wind Categorization

Abstract

Solar wind classification is conducive to understand the physical processes ongoing at the Sun and solar wind evolution in the interplanetary space, and furthermore, it is helpful for early warning of space weather events. With rapid developments in the field of artificial intelligence, machine learning approaches are increasingly being used for pattern recognition. In this study, an approach from machine learning perspectives is developed to automatically classify the solar wind at 1 AU into four types: coronal-hole-origin, streamer-belt-origin, sector-reversal-region-origin, and ejecta. By exhaustive enumeration, an eight-dimensional scheme ($B_T$, $N_P$, $T_P$, $V_P$, $N_{αp}$, $T_{exp}/T_P$, $S_p$, and $M_f$) is found to perform the best among 8191 combinations of 13 solar wind parameters. 10 popular supervised machine learning models, namely $k$ Nearest Neighbors (KNN), Support Vector Machines with linear and Radial Basic Function kernels, Decision Tree, Random Forest, Adaptive Boosting, Neural Network, Gaussian Naive Bayes, Quadratic Discriminant Analysis, and Extreme Gradient Boosting, are applied to the labeled solar wind data sets. Among them, KNN classifier obtains the highest overall classification accuracy, 92.8%. It significantly improves the accuracy by 9.6% over existing manual schemes. No solar wind composition measurements are needed, permitting our classification scheme to be applied to most solar wind spacecraft data. Besides, two application examples indicate that solar wind classification is helpful for the risk evaluation of predicted magnetic storms and surface charging of geosynchronous spacecrafts.