canonical correlation analysis

(redirected from Canonical correlation)
Also found in: Encyclopedia, Wikipedia.

canonical correlation analysis

A statistical manoeuvre which defines covariation for 2 groups of random variables—x and y—and seeks to identify linear combinations of the x's and the y's which have maximum correlation with each other.
References in periodicals archive ?
To work on this set of dependent and independent variables the Canonical Correlation Analysis (CC) was tentatively chosen.
Results of a canonical correlation analysis identified 2 interpretable variates.
Non-linear canonical correlation technique was used to analyse the data.
Implementation is covered in Part II, which proceeds through multilinear principal component, discriminant, independent component, canonical correlation, and partial least squares analyses.
In a little more than 50 pages, the authors attempt to address three different analysis of variance (ANOVA) models (single factor, two factor with nonrepeated experiments, and two factor with repeated experiments), discriminant analysis, factor analysis, principal component analysis, multidimensional scaling, and canonical correlation analysis.
Age explained 80% of variation in the multivariate model, with a canonical correlation coefficient of 0.
The value of canonical correlation were also found to be significant at 0.
Canonical correlation and multiple regression analyses indicated that less educated, less hardy, and recently injured participants were more likely to use escape-avoidance coping and less likely to use social support, problem solving, and positive reappraisal coping behaviors (p < .
The methods used, like the Davidson and Mackinnon (1981) test for specification error and the canonical correlation analysis (Chen and Jordan, 1993; Cheng, 1995) have not been used in similar studies for the ASE.
To assess the extent to which religious variables predicted internalized outcomes, a canonical correlation analysis was conducted using religious variables (gathered through RES, Quest Scale, SWBS) as "predictors of the internalizing outcome variables (shame, guilt, internalized homophobia) in order to analyze the multivariate shared relationship between the two variable sets" (p.
Further, the analysis is performed using canonical correlation which allows the interpretation of the results without imposing an assumption of causality.