학술논문
A new method of reduction of dimensionality in disctrminant analysis: application to the membership problem in open clusters
Document Type
TEXT
Author
Source
Subject
Language
Czech
Multiple Languages
Multiple Languages
Abstract
A method of reduction of dimensionality in discriminant analysis has been specially designed for this study of separation between two populations and it is only based on the heterocedasticityof both groups. For this reason, it is of interest in the case of zero mean difference vector. Two examples of application, concerning the membership problem in open clusters from uvby andHβ photometry and the separation of globular clusters into Oosterhoff's groups are shown.