基于DBSCAN聚类算法的疏散星团NGC 188的3维运动学成员判定
发布时间：2018-11-26 08:44
【摘要】：利用疏散星团NGC 188所在天区的1046颗恒星样本的高精度3维(3D)运动学数据(自行和视向速度)测试了DBSCAN(Density-Based Spatial Clustering of Applications with Noise)聚类算法的成员判定效果.为了避免自行和视向速度的单位不一致带来的影响,在数据预处理阶段将3个分量的数据统一标准化至[0,1]区间.利用第k个最近邻点距离方法分析了1046颗恒星样本在标准化无量纲3D速度空间的分布特征,再根据第k个最近邻点距离随k值的变化趋势确定了DBSCAN聚类算法的输入参数(Eps,MinPts),最后利用DBSCAN聚类算法分离出497颗3D运动学成员星.分析结果表明得到的3D运动学成员星是可靠的.
[Abstract]:The high accuracy 3D kinematics data (self-motion and apparent velocity) of 1046 star samples in the sky region of open cluster NGC 188 are used to test the membership determination effect of DBSCAN (Density-Based Spatial Clustering of Applications with Noise) clustering algorithm). In order to avoid the effect of the unit inconsistency of self-motion and apparent velocity, the data of three components are standardized to the range of [0 / 1] in the stage of data preprocessing. The distribution characteristics of 1046 star samples in dimensionless 3D velocity space are analyzed by using the k-th nearest neighbor distance method, and the input parameters of DBSCAN clustering algorithm (Eps,) are determined according to the variation trend of the k-th nearest neighbor distance with k value. MinPts), finally uses DBSCAN clustering algorithm to separate 497 3D kinematics stars. The results show that the 3D kinematics star is reliable.
【作者单位】： 常州大学信息科学与工程学院;
【基金】：国家自然科学基金项目(11403004)资助
【分类号】：P154.11
，
本文编号：2358017
[Abstract]:The high accuracy 3D kinematics data (self-motion and apparent velocity) of 1046 star samples in the sky region of open cluster NGC 188 are used to test the membership determination effect of DBSCAN (Density-Based Spatial Clustering of Applications with Noise) clustering algorithm). In order to avoid the effect of the unit inconsistency of self-motion and apparent velocity, the data of three components are standardized to the range of [0 / 1] in the stage of data preprocessing. The distribution characteristics of 1046 star samples in dimensionless 3D velocity space are analyzed by using the k-th nearest neighbor distance method, and the input parameters of DBSCAN clustering algorithm (Eps,) are determined according to the variation trend of the k-th nearest neighbor distance with k value. MinPts), finally uses DBSCAN clustering algorithm to separate 497 3D kinematics stars. The results show that the 3D kinematics star is reliable.
【作者单位】： 常州大学信息科学与工程学院;
【基金】：国家自然科学基金项目(11403004)资助
【分类号】：P154.11
，
本文编号：2358017
本文链接：https://www.wllwen.com/kejilunwen/tianwen/2358017.html
