R help - Mar 2003 - Measure of Redundancy In Variables

Hi all,
  I have a question which I guess is more of a general stats question than a
specific R quetions.
I have a data set that contains a large number of numerical variables (in the
hundreds). What I
would like to do is quantify the redundancy in those variables. Let me explain
what I mean by
that.
If I use Principle Component Analysis (PCA) to reduce the amount of variables,
the process
measures the relationship between the different variables and reorganises it so
that each variable
provides unique information and removes any redundancy between different
variables. What I would
like to do is a kind of measure between the data before PCA and after PCA. For
example, if there
is no redundancy, i.e. all of the pre-PCA variables provide unique information,
the redundancy
rate would be 100%. On the other hand if all the pre-PCA variables provide the
same information
than the redundancy rate would be 1%.
Could anyone tell me if there is a method of measuring this redundancy rate or
something similar
in R.
If somebody could help me with this issue it would be greatly appreciated. Many
Thanks

Rishabh

Have you considered "factanal" in library "mva"?  How do the
"uniquenesses" in the "factanal" object relate to what you
want?  From
what I read of your question, it sounds like they estimate exactly what 
you want.

Hope this helps.
Spencer Graves

Rishabh Gupta wrote:
> Hi all,
>   I have a question which I guess is more of a general stats question than
a specific R quetions.
> I have a data set that contains a large number of numerical variables (in
the hundreds). What I
> would like to do is quantify the redundancy in those variables. Let me
explain what I mean by
> that.
> If I use Principle Component Analysis (PCA) to reduce the amount of
variables, the process
> measures the relationship between the different variables and reorganises
it so that each variable
> provides unique information and removes any redundancy between different
variables. What I would
> like to do is a kind of measure between the data before PCA and after PCA.
For example, if there
> is no redundancy, i.e. all of the pre-PCA variables provide unique
information, the redundancy
> rate would be 100%. On the other hand if all the pre-PCA variables provide
the same information
> than the redundancy rate would be 1%.
> Could anyone tell me if there is a method of measuring this redundancy rate
or something similar
> in R.
> If somebody could help me with this issue it would be greatly appreciated.
Many Thanks
> 
> Rishabh
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://www.stat.math.ethz.ch/mailman/listinfo/r-help