Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Canonical correlations for test data? #8

Open
kimrants opened this issue Jul 13, 2018 · 2 comments
Open

Canonical correlations for test data? #8

kimrants opened this issue Jul 13, 2018 · 2 comments

Comments

@kimrants
Copy link

After running validate, is it possible to obtain the canonical correlations for the components?
As far as I can read, you can only get the correlation to the actual test data and/or predictions... but wouldn't it better for validating the model to obtain the canonical correlations?

Thanks!

@kimrants
Copy link
Author

I actually don't understand why the validation method that adds the attribute: cca.corrs is supposed to work?
As far as I understand, CCA produces weights w_a and w_b such that corr(Aw_a, Bw_b) is maximised.

If you just find the correlation corr(A_test*w_a, B_test) you are not using one of the mappings? I don't see why this should work?

Thank you!

@sdrakulich
Copy link

Any followup on this?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants