基于自动特征提取方法的太阳耀斑预报模型

发布时间：2018-03-13 22:27

本文选题：太阳耀斑　切入点：太阳活动区　出处：《科学通报》2016年36期 　论文类型：期刊论文

【摘要】：在太阳耀斑预报模型中,首先需要从原始观测数据中提取刻画太阳活动区特性的物理特征参量,然后使用统计或机器学习方法寻找物理特征参量与太阳耀斑发生的关系,以达到建立太阳耀斑预报模型的目的.其中,太阳活动区物理特征的提取在整个建模过程中发挥着重要的作用,活动区物理特征的优劣直接决定着预报模型性能的高低.然而,随着机器学习技术的发展,机器学习方法中的深度学习算法能够从原始数据中自动提取特征,并建立预报模型.本文利用深度学习方法建立了一个太阳耀斑预报模型.与先提取活动区物理参量、再建立预报模型的传统机器学习方法相比较,本文所建立的预报模型具有更好的预报性能.
[Abstract]:In the solar flare prediction model, the physical characteristic parameters describing the characteristics of the solar active region should be extracted from the original observation data, and then the relationship between the physical characteristic parameters and the solar flares is found by using statistical or machine learning methods. In order to establish the solar flare prediction model, the extraction of the physical features of the solar active region plays an important role in the whole modeling process, and the physical characteristics of the active region directly determine the performance of the prediction model. With the development of machine learning technology, depth learning algorithms in machine learning methods can automatically extract features from raw data. In this paper, a solar flare prediction model is established by using the depth learning method. Compared with the traditional machine learning method, the physical parameters of the active region are extracted first and then the prediction model is established. The prediction model established in this paper has better prediction performance.
【作者单位】：北京物资学院;中国科学院国家天文台;中国科学院空间科学与应用研究中心;
【基金】：国家自然科学基金(11273031,11303051) 智能物流系统北京市重点实验室项目(BZ0211) 北京市智能物流系统协同创新中心项目资助
【分类号】：P182.52

【相似文献】