核心概念
The author utilizes machine learning strategies to predict Solar Energetic Particle events, emphasizing the importance of understanding underlying processes leading to these events.
要約
The study focuses on predicting Solar Energetic Particle (SEP) events using machine learning models. It highlights the significance of accurate forecasting due to radiation hazards in space operations. The research combines data from Solar Cycles 23 and 24 to improve predictive accuracy.
The study uses a dataset combining SMARP and SHARP data, training SVM and regression models for SEP event prediction. Key predictors like USFLUXZ, R VALUE, and ANG DIST show potential for distinguishing between positive and negative flares.
Different ML models are evaluated across various predictor combinations, with the R VALUE and ANG DIST pair consistently achieving high accuracy. The study emphasizes the importance of selecting relevant predictors for effective SEP event forecasting.
統計
Our study indicates that despite the augmented volume of data, the prediction accuracy reaches 0.7 ± 0.1.
A linear SVM model with training and testing configurations reveals a slight increase (+0.04 ± 0.05) in the accuracy of a 14-hour SEP forecast compared to previous studies.
Leveraging vast datasets, ML models are being trained to predict Solar Proton Events (SPEs) by considering diverse parameters such as magnetic field characteristics.
The SMARP-SHARP dataset includes maps of automatically-tracked active regions extracted from full-disk magnetograms.
The merged dataset keeps the original cadence of the data products: 96 minutes for MDI and 12 minutes for HMI observations.
引用
"Prediction of the Solar Energetic Particle (SEP) events garner increasing interest as space missions extend beyond Earth’s protective magnetosphere."
"These events pose significant radiation hazards to aviation, space-based electronics, and particularly space exploration."