[Scipy-tickets] [SciPy] #435: Whitening full covariance gaussian models in pyem

SciPy scipy-tickets@scipy....
Tue Jun 5 06:20:59 CDT 2007


#435: Whitening full covariance gaussian models in pyem
--------------------+-------------------------------------------------------
 Reporter:  cdavid  |        Owner:  cdavid                                
     Type:  defect  |       Status:  new                                   
 Priority:  normal  |    Milestone:  pymachine: cleaning existing toolboxes
Component:  pyem    |      Version:                                        
 Severity:  normal  |   Resolution:                                        
 Keywords:          |  
--------------------+-------------------------------------------------------
Old description:

> For numerical reasons, it is often desirable to whiten data before doing
> clustering, etc... For full covariance matrices, this means changing the
> "basis" of the matrix, whiten there and going back to original basis.

New description:

 For numerical reasons, it is often desirable to whiten data before doing
 clustering, etc... For full covariance matrices, this means changing the
 "basis" of the matrix, whiten there and going back to original basis. See
 page 299/300 of Bishop Neural Network for Pattern recognition for more
 details

-- 
Ticket URL: <http://projects.scipy.org/scipy/scipy/ticket/435#comment:2>
SciPy <http://www.scipy.org/>
SciPy is open-source software for mathematics, science, and engineering.


More information about the Scipy-tickets mailing list