Asymptotic distributions of principal components based on robust dispersions

Hengjian Cui

Biometrika, 2003, vol. 90, issue 4, 953-966

Abstract: Algebraically, principal components can be defined as the eigenvalues and eigenvectors of a covariance or correlation matrix, but they are statistically meaningful as successive projections of the multivariate data in the direction of maximal variability. An attractive alternative in robust principal component analysis is to replace the classical variability measure, i.e. variance, by a robust dispersion measure. This projection-pursuit approach was first proposed in Li & Chen (1985) as a method of constructing a robust scatter matrix. Recent unpublished work of C. Croux and A. Ruiz-Gazen provided the influence functions of the resulting principal components. The present paper focuses on the asymptotic distributions of robust principal components. In particular, we obtain the asymptotic normality of the principal components that maximise a robust dispersion measure. We also explain the need to use a dispersion functional with a continuous influence function. Copyright Biometrika Trust 2003, Oxford University Press.

Date: 2003
References: Add references at CitEc
Citations: View citations in EconPapers (10)

There are no downloads for this item, see the EconPapers FAQ for hints about obtaining it.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:oup:biomet:v:90:y:2003:i:4:p:953-966

Ordering information: This journal article can be ordered from
https://academic.oup.com/journals

Access Statistics for this article

Biometrika is currently edited by Paul Fearnhead

More articles in Biometrika from Biometrika Trust Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, UK.
Bibliographic data for series maintained by Oxford University Press ().