On 29 May 2003 13:55:10 -0700, [EMAIL PROTECTED] (christoph) wrote:

> sorry, maybe an all too trivial question. But we have power data from J
> frequency spectra and to have the same range for the data of all our
> subjects, we just transformed them into % values, pseudo-code:
> 
> power[i,j]=power[i,j]/sum(power[i,1:J])
> 
> of course, now we have perfect collinearity in our x design-matrix,
> since all power-values for each subject sum up to 1.
> 
> How shall we solve this problem: just eliminate one column of x, or
> introduce a restriction which says exactly that our power data sum up to
> 1 for each subject?

Compositional data.
A few days ago, someone posted news of a new edition (I 
think it was this book) by Aitchison.  My stats-FAQ at  
  http://www.pitt.edu/~wpilib/statfaq/compfaq.html
includes about 8 other references.

@Article{aitchison82,
  author =       "J. Aitchison",
  title =        "The statistical analysis of compositional data",
  journal =      jrssb,
  year =         1982,
  volume =       44,
  number =       2,
  pages =        "139-177",
  annote =       "With discussion."

I my own experience with power-spectra, I found it 
stabilizing  to use the logit of the relative power.  Of course,
that got rid of the sum-to-1  problem.

-- 
Rich Ulrich, [EMAIL PROTECTED]
http://www.pitt.edu/~wpilib/index.html
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to