Rotation argument for scikit-learn's factor analysis

Question

One of the hallmarks of factor analysis is that it allows for non-orthogonal latent variables.

In R for example this feature is accessible via the rotation parameter of factanal. Is there any such provision for sklearn.decomposition.FactorAnalysis? Clearly it's not among the arguments - but maybe there is another way to achieve this?

Sadly I have been unable to find many examples of usage for this function.

Answer 1

Interesting question. I think there are no rotations implemented indeed - see this issue.

Maybe this implementation is what you are looking for.