You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Given two view matrices X and Y, CCA should return the n_components linear combinations of the elements in X's and Y's vectors with the largest correlation, i.e the first component has the largest correlation, the second one has the second-largest correlation in a direction that is uncorrelated with the first one, etc. However, in the following code, the second columns are clearly less correlated than the third columns:
The correlation between the third columns of X_transfomed and Y_transformed is approximately 1, while the correlation between the second columns is only 0.389. Therefore, I expected the third column to appear in the place of the second column.
Actual Results
When I run CCA with n_components=2 I get the first and second columns, while I should have gotten the first and the third.
Description
Given two view matrices
X
andY
, CCA should return then_components
linear combinations of the elements inX
's andY
's vectors with the largest correlation, i.e the first component has the largest correlation, the second one has the second-largest correlation in a direction that is uncorrelated with the first one, etc. However, in the following code, the second columns are clearly less correlated than the third columns:Steps/Code to Reproduce
Expected Results
The correlation between the third columns of
X_transfomed
andY_transformed
is approximately 1, while the correlation between the second columns is only 0.389. Therefore, I expected the third column to appear in the place of the second column.Actual Results
When I run CCA with
n_components=2
I get the first and second columns, while I should have gotten the first and the third.Versions
Linux-4.13.0-16-generic-x86_64-with-Ubuntu-17.10-artful
Python 3.6.3 (default, Oct 3 2017, 21:45:48)
NumPy 1.14.3
SciPy 1.2.0
Scikit-Learn 0.19.1
The text was updated successfully, but these errors were encountered: