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
The modularised implementation in #3029 makes it very easy to introduce alternative mechanisms for PCA threshold selection. One possibility suggested here by @jelleveraart is to simply fix the output rank, irrespective of the eigenvalue spectrum. This should be reasonably trivial code-wise.
I'm not sure whether this should be made mutually exclusive with respect to optimal shrinkage, or whether they should just be left as independent mechanisms and users can configure however they wish.
The text was updated successfully, but these errors were encountered:
The modularised implementation in #3029 makes it very easy to introduce alternative mechanisms for PCA threshold selection. One possibility suggested here by @jelleveraart is to simply fix the output rank, irrespective of the eigenvalue spectrum. This should be reasonably trivial code-wise.
I'm not sure whether this should be made mutually exclusive with respect to optimal shrinkage, or whether they should just be left as independent mechanisms and users can configure however they wish.
The text was updated successfully, but these errors were encountered: