If n > p + q
, should I just set lambda1=0, lambda2=0
?
also, does the method = "ridge"
means the regularization is similar to that used in ridge regression?
Thanks!
If n > p + q
, should I just set lambda1=0, lambda2=0
?
also, does the method = "ridge"
means the regularization is similar to that used in ridge regression?
Thanks!
Hi @blueskypie,
Yes, if n > p+q you can use a classical CCA and that would be the same as setting the lambda parameters to 0.
Ridge penalties are indeed similar to those used in ridge regression, see details in: https://www.jstatsoft.org/article/view/v023i12