When using canonical correlation analysis (CCA), we can integrate the dataset and label information via transforming the class label matrix Y into the class indicator matrix T. Such as: $T = (YY^T)^½Y$ in [this article on LS CCA].
While for the numeric dataset and the binary dataset, the binary dataset can be viewed as the class label matrix and transformed into the class indicator matrix. After that, should I use the CCA?
Though without this transformation, [the CCA can be used], I'd like to know how to explian this kind of transformation once using it on the binary dataset (not label matrix). Thank you.