主成分分析法是从观测数据中获取主 要信息的一种多变量统计方法。它用数目少得多的新变量代替原有观测量,以寻 找原观测量之间的相互关系,且不损失原始数据的主要信息,尤其是对于大样本、 多参量的情况,该方法更为简捷而有效。目前,主成分分析法被广泛应用于天体 物理的诸多研究领域中。介绍了主成分分析法的原理和它在天体物理中的广泛应 用。
Principal Component Analysis (PCA) is a main multivariate statistical method for getting principal information from observational data. It uses few new variables instead of initial parameters, in order to find out the relations among the inital parameters, without losing the main information of intial data. Especially for the case of large sample and multivariate, this method is simpler and more efficient. In the present day, PCA is applied widely in many research fields of astrophysics. The main principle and applications to astrophysics of PCA are reviewed in this paper.