基于残差分布度量的M矮星细分类研究
发布时间：2019-05-24 13:08
【摘要】：恒星光谱分类是天文数据处理中一项非常重要的工作,主要对海量光谱巡天数据按照其物理性质进行分类。利用残差分布度量的方法对LAMOST巡天中观测到的M矮星光谱进行细分类研究。残差分布度量是一种光谱间的距离度量方法,计算光谱之间的距离时,先将两条光谱进行归一化处理,之后计算对应波长采样点处的残差,最终以残差分布的标准差作为光谱之间的距离。使用LAMOST DR2中释放的M矮星光谱进行细分类实验。实验结果表明,残差分布度量方法能比较准确地对M矮星光谱数据进行细分类。还研究了信噪比、离群点以及残差标准化系数等因素对分类结果的影响。
[Abstract]:Star spectral classification is a very important work in astronomical data processing, which mainly classifies massive spectral patrol data according to their physical properties. The spectra of M dwarfs observed in LAMOST survey are classified by residual distribution measurement. Residual distribution measurement is a distance measurement method between spectra. When calculating the distance between spectra, the two spectra are first normalized, and then the residual errors at the corresponding wavelength sampling points are calculated. Finally, the standard deviation of residual distribution is used as the distance between spectra. The M dwarf spectrum released from LAMOST DR2 was used for fine classification experiment. The experimental results show that the residual distribution measurement method can subdivide the spectral data of M dwarfs more accurately. The effects of signal-to-noise ratio (SNR), outlier and residual standardization coefficient on the classification results are also studied.
【作者单位】： 山东大学(威海)机电与信息工程学院;中国科学院光学天文重点实验室国家天文台;
【基金】：国家自然科学基金项目(U1431102,11473019)资助
【分类号】：P145.4
，
本文编号：2484887
[Abstract]:Star spectral classification is a very important work in astronomical data processing, which mainly classifies massive spectral patrol data according to their physical properties. The spectra of M dwarfs observed in LAMOST survey are classified by residual distribution measurement. Residual distribution measurement is a distance measurement method between spectra. When calculating the distance between spectra, the two spectra are first normalized, and then the residual errors at the corresponding wavelength sampling points are calculated. Finally, the standard deviation of residual distribution is used as the distance between spectra. The M dwarf spectrum released from LAMOST DR2 was used for fine classification experiment. The experimental results show that the residual distribution measurement method can subdivide the spectral data of M dwarfs more accurately. The effects of signal-to-noise ratio (SNR), outlier and residual standardization coefficient on the classification results are also studied.
【作者单位】： 山东大学(威海)机电与信息工程学院;中国科学院光学天文重点实验室国家天文台;
【基金】：国家自然科学基金项目(U1431102,11473019)资助
【分类号】：P145.4
，
本文编号：2484887
本文链接：https://www.wllwen.com/kejilunwen/tianwen/2484887.html
