CANONICAL CORRELATION ANALYSIS IN THE CASE OF MULTIVARIATE REPEATED MEASURES DATA
CANONICAL CORRELATION ANALYSIS IN THE CASE OF MULTIVARIATE REPEATED MEASURES DATA
Author(s): Mirosław Krzyśko, Wojciech Łukaszonek, Waldemar WołyńskiSubject(s): Economy, Public Finances, Socio-Economic Research
Published by: Główny Urząd Statystyczny
Keywords: canonical correlation analysis; repeated measures data (doubly multivariate data); Kronecker product covariance structure; maximum likelihood estimates
Summary/Abstract: In this paper, we present, in the real example, canonical variables applicable in the case of multivariate repeated measures data under the following assumptions: (1) multivariate normality for the vector of observations and (2) Kronecker product structure of the positive definite covariance matrix. These variables are especially useful when the number of observations is not large enough to estimate the covariance matrix, and thus the traditional canonical variables fail. Computational schemes for maximum likelihood estimates of required parameters are also given.
Journal: Statistics in Transition. New Series
- Issue Year: 19/2018
- Issue No: 1
- Page Range: 75-85
- Page Count: 11
- Language: English