[R] scaling advice

Peter Langfelder peter.langfelder at gmail.com
Sat Jul 16 00:45:09 CEST 2011


On Fri, Jul 15, 2011 at 2:05 PM, Data Analytics Corp.
<walt at dataanalyticscorp.com> wrote:

>  But then he apparently rescaled this 44x13
> matrix so that the rows all sum to zero and the columns all sum to zero.
>  None of the row and column standard deviations are 1.0.  This I can't see
> how to do.  How can I rescale the rows and columns so that they all sum to
> zero?  Any suggestions?

Well, he could have used the Gower's centering transformation
described for example in this pdf:

www.stat.auckland.ac.nz/~mja/prog/PCO_UserNotes.pdf

As you can convince yourself very easily, given any matrix A, the
matrix G calculated as on page 3 of the document will have zero sum
rows and columns.

HTH,

Peter



More information about the R-help mailing list