In Exercise-Classification the largest eigenvalue does not always have index 0.
Seems like this depends on the Numpy and Python versions.
Therefore I added some code that sorts the eigenvalues and eigenvectors in increasing order. This ensures the index 1 eigenvector is always the correct one to select.
In Exercise-Classification the largest eigenvalue does not always have index 0. Seems like this depends on the Numpy and Python versions.
Therefore I added some code that sorts the eigenvalues and eigenvectors in increasing order. This ensures the index 1 eigenvector is always the correct one to select.