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
It would be good if PCA supported any type of array that support the required methods.
and sored what ever projection matrix those methods came up with.
For example:
NamedDims.jl supports svd and will give you a partially named projection matrix.
Which can be pretty useful; since it will enforce and propagate it's names/
But if you use it with PCA here, then it works but the result is converted to a Matrix
#140 would help with this, though not solve the full problem.
Almost all fit functions accept AbstarctArray data (expept Whitening & MulticlassLDA). Problems arise from particular implementations (e.g. #140). We need to review the code to remove any type constrains, and allow to fail on missing implementations of the specific dependent functions for various array types (e.g. svd or cov)
It would be good if PCA supported any type of array that support the required methods.
and sored what ever projection matrix those methods came up with.
For example:
NamedDims.jl supports
svd
and will give you a partially named projection matrix.Which can be pretty useful; since it will enforce and propagate it's names/
But if you use it with PCA here, then it works but the result is converted to a
Matrix
#140 would help with this, though not solve the full problem.
#139 is also related
The text was updated successfully, but these errors were encountered: