Hi Olivier,
Regarding your concern about the efficiency of the algorithms, as far as I
can tell, as the field is developing rapidly, the current state-of-the-art
RPCA algorithm is actually very scalable. In a nutshell, the cost of the
RPCA is about the cost of 20s SVD. For a 1000x1000 matrix, it is likely we
can get the decomposition within 1 minute or so. It seems that the matrix
completion task you mentioned can be solved in a similar manner using
convex optimization. I'm not very familiar with the SGD approach, though.
Is there any performance guarantee about this algorithm? Thanks.
Sincerely,
Kerui Min
------------------------------------------------------------------------------
Virtualization & Cloud Management Using Capacity Planning
Cloud computing makes use of virtualization - but cloud computing
also focuses on allowing computing to be delivered as a service.
http://www.accelacomm.com/jaw/sfnl/114/51521223/
_______________________________________________
Scikit-learn-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
