Reply With QuoteApologize if this seems idiotic, but have you Google searched that exact phrase online? The phase was in this book when I searched, not sure where your issue is.
http://books.google.com/books?id=Fjk...20zero&f=false
I'm running the CANCORR procedure in SAS and I keep getting a warning:
"Note: The correlation matrix for the WITH Variables is less than full rank. Therefore, some canonical correlations and coefficients will be zero."
and indeed the final canonical correlation is zero.
However, I can't see that the correlation matrix I entered has less than full rank. Matlab tells me it has full rank, and I can't see any rows or columns that are linear combinations of others. The correlation matrix is below, and the WITH variables are the final four.
Does anyone know what might be going on here?
Code:n . 138 138 138 138 138 138 138 138 138 CORR x1 1 . . . . . . . . CORR x2 0.754 1 . . . . . . . CORR x3 -0.69 -0.712 1 . . . . . . CORR x4 -0.446 -0.515 0.323 1 . . . . . CORR x5 0.692 0.412 -0.444 -0.334 1 . . . . CORR x6 -0.605 -0.722 0.737 0.527 -0.383 1 . . . CORR x7 -0.479 -0.419 0.361 0.461 -0.505 0.251 1 . . CORR x8 0.78 0.542 -0.546 -0.393 0.737 -0.49 -0.434 1 . CORR x9 -0.152 -0.102 0.172 -0.019 -0.148 0.25 -0.79 -1.63
Apologize if this seems idiotic, but have you Google searched that exact phrase online? The phase was in this book when I searched, not sure where your issue is.
http://books.google.com/books?id=Fjk...20zero&f=false
Posting Permissions
|
|