DBSCAN聚类算法在Gaia-DR2中检测疏散星团的研究

DBSCAN Clustering Algorithm for Detection of Nearby Open Clusters Based on Gaia-DR2

基于Gaia Data Release 2 (Gaia-DR2)星表,采用数据挖掘技术中的DBSCAN(Density-Based Spatial Clustering of Applications with Noise)算法进行邻近疏散星团成员检测.从Gaia-DR2中选取了594284颗恒星(距离太阳<100 pc)作为样本,使用恒星的五维数据(三维空间位置和两维自行)进行聚类分析.在数据预处理阶段,将每一维数据标准化到[0, 1]区间内,避免了单位不一致对聚类效果的影响.然后,利用k-dist图确定了DBSCAN算法的输入参数(Eps, MinPts).最终,使用DBSCAN算法获取了133颗成员星,它们在五维相空间中可以被分成两组,分别对应于疏散星团Hyades和Coma.分析结果表明得到的成员星是可靠的.根据两个星团的成员星, Hyades和Coma的距离分别确定为(46.5±0.3) pc和(84.9±0.4) pc.

In this paper,we attempt to use the DBSCAN(Density-Based Spatial Clustering of Applications with Noise)clustering algorithm to detect nearby open clusters based on Gaia Data Release 2(Gaia-DR2).We select 594284 stars(within a distance of 100 pc to the sun)from the Gaia-DR2 catalog,and construct a five dimensional phase space(three dimensional space position and two dimensional proper motions)in order to obtain reliable cluster members.At the data preprocessing stage,we normalize each dimension of data to the[0;1]interval in order to avoid the effect of inconsistent units.Then,we use k-dist graph to determine the input parameters of the DBSCAN Algorithm.Finally,we obtain 133 reliable members using the DBSCAN algorithm,which correspond to two open clusters--Hyades and Coma.According to these cluster members,the distances to the Hyades and Coma clusters are determined to be(46.5  0.3)pc and(84.9  0.4)pc,respectively.