K-FOLD CROSSVALIDATION IN CANONICAL-ANALYSIS

Authors
LIANG KH KRUS DJ WEBB JM
Citation
Kh. Liang et al., K-FOLD CROSSVALIDATION IN CANONICAL-ANALYSIS, Multivariate behavioral research, 30(4), 1995, pp. 539-545
Citations number
7
Categorie Soggetti
Social Sciences, Mathematical Methods","Psychologym Experimental","Statistic & Probability","Mathematical, Methods, Social Sciences
Journal title
Multivariate behavioral research → ACNP
ISSN journal
00273171
Volume
30
Issue
4
Year of publication
1995
Pages
539 - 545
Database
ISI
SICI code
0027-3171(1995)30:4<539:KCIC>2.0.ZU;2-Y
Abstract
A computer-assisted, K-fold crossvalidation technique is discussed wit hin the framework of canonical correlation analysis of randomly genera ted data sets. Results of the analysis suggest that this technique of multi-crossvalidation can be an effective method to reduce the contami nation of canonical variates and canonical correlations by sample-spec ific variance components.