中国科学院云南天文台机构知识库

太阳物理研究已经进入大数据时代,而机器学习作为大数据研究的一种良好工具已经获得越来越多的认可.本文评述了自2007年以来机器学习在太阳物理中的应用.从结果上看,最近4年这一领域的研究明显增加.所利用的数据包括地面和空间的各种仪器、各种类型和波段的太阳观测资料.研究领域涵盖太阳耀斑、日冕物质抛射、太阳黑子等太阳物理研究的主要方面.目前虽然获得一些很好的结果,但尚未有突破性的进展.使用的机器学习方法涉及分类、回归、聚类、降维以及深度学习等手段,但经典的算法,尤其是分类方法依然占据主导地位.这意味着机器学习在太阳物理的应用还处于起步阶段,但同样也意味着在这一领域还有很多工作可以深入开展. 

Solar physics has entered the era of big data, and machine learning has gained more and more recognition as a good tool for big data research. This paper reviews the application results of machine learning in solar physics since 2007. Our studies have shown that research in this field has increased significantly during the last four years. Massive solar observation data obtained from various instruments on the ground and in space have been applied, and the topics have covered major aspects of solar physics, such as solar flares, coronal mass ejections, sunspots. Although some good results have emerged and proved that machine learning is suitable for data analysis of solar physics, there has not been a
breakthrough yet. The machines learning methods that used in this field involve classification, regression, clustering, dimensionality reduction, and deep learning. However, classical algorithms, especially classical  lassification method is more popular. This means that the application of machine learning in solar physics is still in its infancy, but it also means
that there is still a lot of work in this field that can be studied in the future.