Prediction of major solar flares using interpretable class-dependent reward framework with active region magnetograms and domain knowledge
文献类型：期刊论文
| 作者 | Wu, Zixian5,7; Li, Xuebao5,7; Zheng, Yanfang5,7; Wang, Rui7; Zhang, Shunhuang5; Wei, Jinfang4,5; Lv, Yongshang5; Dong L(董亮)3; Abidin, Zamri Zainal1,6; Shah, Noraisyah Mohamed2 |
| 刊名 | MONTHLY NOTICES OF THE ROYAL ASTRONOMICAL SOCIETY
 |
| 出版日期 | 2026-04 |
| 卷号 | 547期号:2 |
| 关键词 | Sun: activity Sun: flares Sun: magnetic fields |
| ISSN号 | 0035-8711 |
| DOI | 10.1093/mnras/stag349 |
| 产权排序 | 第4完成单位 |
| 文献子类 | Article |
| 英文摘要 | In this work, we develop, for the first time, a supervised classification framework with class-dependent rewards (CDR) to predict >= M flares within 24 h. We construct multiple data sets, covering knowledge-informed features and line-of-sight (LOS) magnetograms. We also apply three deep learning models (CNN, CNN-BiLSTM, and Transformer) and three CDR counterparts (CDR-CNN, CDR-CNN-BiLSTM, and CDR-Transformer). First, we analyse the importance of LOS magnetic field parameters with the Transformer, then compare its performance using LOS-only, vector-only, and combined magnetic field parameters. Second, we compare flare prediction performance based on CDR models versus deep learning counterparts. Third, we perform sensitivity analysis on reward engineering for CDR models. Fourth, we use the SHAP method for model interpretability. Finally, we conduct performance comparison between our models and NASA/CCMC. The main findings are: (1) Among LOS feature combinations, R_VALUE and AREA_ACR consistently yield the best results. (2) Transformer achieves better performance with combined LOS and vector magnetic field data than with either alone. (3) Models using knowledge-informed features outperform those using magnetograms. (4) While CNN and CNN-BiLSTM outperform their CDR counterparts on magnetograms, CDR-Transformer is slightly superior to its deep learning counterpart when using knowledge-informed features. Among all models, CDR-Transformer achieves the best performance. (5) The predictive performance of the CDR models is not overly sensitive to the reward choices. (6) Through SHAP analysis, the CDR model tends to regard TOTUSJH as more important, while the Transformer tends to prioritize R_VALUE more. (7) Under identical prediction time and active region number, the CDR-Transformer shows superior predictive capabilities compared to NASA/CCMC. |
| 学科主题 | 天文学 ; 太阳与太阳系 |
| URL标识 | 查看原文 |
| 出版地 | GREAT CLARENDON ST, OXFORD OX2 6DP, ENGLAND |
| WOS关键词 | CONVOLUTIONAL NEURAL-NETWORK ; MAGNETIC-FIELD ; FORECAST MODEL ; PERFORMANCE ; PROXIES ; FUSION ; IMAGES ; SHARP ; CNN |
| 资助项目 | National Natural Science Foundation of China[12473056]; Natural Science Foundation of Jiangsu Province[BK20241830]; Qing Lan Project; Specialized Research Fund for State Key Laboratories |
| WOS研究方向 | Astronomy & Astrophysics |
| 语种 | 英语 |
| WOS记录号 | WOS:001709296400001 |
| 出版者 | OXFORD UNIV PRESS |
| 资助机构 | National Natural Science Foundation of China[12473056] ; Natural Science Foundation of Jiangsu Province[BK20241830] ; Qing Lan Project ; Specialized Research Fund for State Key Laboratories |
| 版本 | 出版稿 |
| 源URL | [http://ir.ynao.ac.cn/handle/114a53/29024]  |
| 专题 | 云南天文台_射电天文研究组 |
| 通讯作者 | Li, Xuebao; Zheng, Yanfang |
| 作者单位 | 1.Radio Cosmology Lab, Centre for Astronomy and Astrophysics Research, Department of Physics, Faculty of Science, Universiti Malaya, 50603 Kuala Lumpur, Malaysia; 2.Department of Electrical Engineering, Faculty of Engineering, Universiti Malaya, 50603 Kuala Lumpur, Malaysia 3.Yunnan Astronomical Observatory, Chinese Academy of Sciences, Kunming 650216, People’s Republic of China; 4.School of Software, Southeast University, Nanjing 211189, People’s Republic of China; 5.School of Computer Science, Jiangsu University of Science and Technology, Zhenjiang 212100, People’s Republic of China; 6.National Centre for Particle Physics, Universiti Malaya, 50603 Kuala Lumpur, Malaysia; 7.State Key Laboratory of Space Weather, Chinese Academy of Sciences, Beijing 100190, People’s Republic of China; |
| 推荐引用方式 GB/T 7714 | Wu, Zixian,Li, Xuebao,Zheng, Yanfang,et al. Prediction of major solar flares using interpretable class-dependent reward framework with active region magnetograms and domain knowledge[J]. MONTHLY NOTICES OF THE ROYAL ASTRONOMICAL SOCIETY,2026,547(2). |
| APA | Wu, Zixian.,Li, Xuebao.,Zheng, Yanfang.,Wang, Rui.,Zhang, Shunhuang.,...&Jin, Honglei.(2026).Prediction of major solar flares using interpretable class-dependent reward framework with active region magnetograms and domain knowledge.MONTHLY NOTICES OF THE ROYAL ASTRONOMICAL SOCIETY,547(2). |
| MLA | Wu, Zixian,et al."Prediction of major solar flares using interpretable class-dependent reward framework with active region magnetograms and domain knowledge".MONTHLY NOTICES OF THE ROYAL ASTRONOMICAL SOCIETY 547.2(2026). |
入库方式： OAI收割
来源：云南天文台
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。