0 Stars
Updated Last
6 Years Ago
Started In
April 2014


Build Status


This package implements canonical correlation in Julia. Someday it will also implement partial least squares. (If you'd like to contribute, please let me know!)


To perform canonical correlation analysis between matrices X and Y:

using CrossDecomposition
cc = canoncor(X, Y)

cc is a CanonicalCorrelation object that implements the following methods:

  • coef(cc) returns the canonical coefficients (loadings) of X and Y as a tuple of two matrices. The coefficients are scaled so that the scores have identity covariance.
  • cor(cc) returns the vector of canonical correlations.
  • scores(cc) returns the canonical variables (component scores) of X and Y as a tuple of two matrices.