摘要
恒星光谱分类是光谱分析的一种重要方法,是天体光谱数据挖掘的重要内容。针对从LAMOST(the Large Sky AreaMulti-Object Fiber Spectroscopic Telescope)Data Release 5(DR5)选取出的33 000条F、G和K型3种恒星光谱数据,采用一种基于深度信念网络的恒星光谱分类方法,通过在训练过程中对恒星光谱数据进行分层特征学习,从而建立深度信念网络模型。最后对此模型进行恒星光谱分类测试,得到F、G和K型3种恒星的分类精确率分别为0.93、0.90和0.98,从而验证了该模型对这3种恒星光谱的正确性,分类精确率较高,对海量天体光谱数据的处理有着重要意义。
Stellar spectral classification is an important method of spectral analysis, and it is an important part of celestial data mining. In accordance with the selected spectral data of 33 000 F, G and K type stars from LAMOST(the Large Sky Area Multi-Object Fiber Spectroscopic Telescope)Data Release 5(DR5), a method based on the deep belief network for classification was adopted. The deep belief network model is established by hierarchical feature learning of stellar spectral data during training, and finally, the stellar spectral classification test of this model shows that the classification accuracy of the three stars F, G and K are 0.93, 0.90 and 0.98 respectively, which verifies the correctness of the model and its high classification accuracy rate and is of great significance for the processing of massive celestial spectral data.
作者
张静敏
许婷婷
杜利婷
周卫红
Zhang Jingmin;Xu Tingting;Du Liting;Zhou Weihong(School of Mathematics and Computer Science,Yunnan Minzu University,Kunming 650500,China;Key Laboratory of theStructure and Evolution of Celestial Objects,Chinese Academy of Sciences,Kunming 650011,China)
出处
《大理大学学报》
CAS
2019年第6期10-14,共5页
Journal of Dali University
基金
国家自然科学基金资助项目(61561053)
关键词
光谱分类
特征学习
深度学习
深度信念网络
spectral classification
feature learning
deep learning
deep belief network