Please use this identifier to cite or link to this item:
http://hdl.handle.net/1942/314
Title: | Transformation of Non Positive Semidefinite Correlation Matrices | Authors: | Rousseeuw, Peter J. MOLENBERGHS, Geert |
Issue Date: | 1993 | Source: | Communications in Statistics, Theory and Methods, 22(4), p. 965-984 | Abstract: | In multivariate statistics, estimation of the covariance or correlation matrix is of crucial importance. Computational and other arguments often lead to the use of coordinate-dependent estimators, yielding matrices that are symmetric but not positive semidefinite. We briefly discuss existing methods, based on shrinking, for transforming such matrices into positive semidefinite matrices. A simple method based on eigenvalues is also considered. Taking into account the geometric structure of correlation matrices, a new method is proposed which uses techniques similar to those of multidimensional scaling. | Keywords: | EIGENVALUE METHOD; MISSING DATA; MULTIDIMENSIONAL SCALING; MULTIVARIATE PROBIT MODEL; ROBUST CORRELATIONS; SHRINKING | Document URI: | http://hdl.handle.net/1942/314 | Link to publication/dataset: | https://www.researchgate.net/publication/232858104_Transformation_of_Non_Positive_Semidefinite_Correlation_Matrices https://www.academia.edu/8393364/Transformation_of_non_positive_semidefinite_correlation_matrices |
ISSN: | 0361-0926 | e-ISSN: | 1532-415X | DOI: | 10.1080/03610928308831068 | ISI #: | A1993KZ73700002 | Rights: | (C) 1993 by Marcel Dekker Inc. | Type: | Journal Contribution |
Appears in Collections: | Research publications |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
1993_Transformation_nonpositive_corr_CommStat.pdf Restricted Access | Published version | 666.88 kB | Adobe PDF | View/Open Request a copy |
SCOPUSTM
Citations
79
checked on Sep 2, 2020
WEB OF SCIENCETM
Citations
84
checked on May 8, 2024
Page view(s)
90
checked on Sep 7, 2022
Download(s)
74
checked on Sep 7, 2022
Google ScholarTM
Check
Altmetric
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.