基于LAMOST，Gaia和红外巡天数据的Ae/Be恒星样本分类和演化特性研究

本项目拟采用机器学习方法，基于LAMOST，Gaia，2MASS和WISE红外巡天数据，通过对多源数据信息的学习和挖掘，开展对Ae/Be恒星（包括赫比格Ae/Be和典型Ae/Be恒星）的搜寻，样本分类，谱线形态分析及其演化特性的研究工作。其主要内容包括：对已证认样本的多源数据进行采集与分析，采用相关子空间的机器学习方法对Ae/Be恒星进行特征学习与分析；基于提取的特征子空间和深度森林方法对Ae/Be星作样本分类，并且对其发射谱线的轮廓形态进行系统分析；将已分类的赫比格Ae/Be和典型Ae/Be恒星点在赫罗图上，估算质量、年龄等参数，探究相应的演化轨迹。

Based on LAMOST, Gaia, 2MASS and WISE infrared survey data, this project aims to search Ae/Be stars (including Herbig Ae/Be and classical Ae/Be stars) by using the machine learning method. The physical classification and spectral line morphology are investigated through learning the multi-source data, and the evolutionary properties are also analyzed for each type. The detailed researches specifically include the following three aspects. Firstly, using the method of relevant subspace, the features of Ae/Be stars are extracted from identified samples with multi-source data. Secondly, based on the extracted feature subspaces, Ae/Be stars and their spectral line profiles are classified and analyzed using the method of depth forest. Finally, based on the astrometric data of Gaia survey, we place Herbig Ae/Be and classical Ae/Be stars in the HR diagram, and endeavor to study the evolutionary tracks according to stellar parameters such as masses and ages.