[R] Pearson correlation coefficient matrix with permutation test

herbert8686 at gmx.de herbert8686 at gmx.de
Sun Mar 20 18:14:36 CET 2011


Hello Gabor,

Thanks for your response. There are certainly several offers in R to construct confidence envelopes.

But here I am specifically interested in sort of a summary output such as provided by the function cor().

I have more than a thousand correlations and I just need for each correlation the output "Pearson's r" combined with the significance of permutation results. So I am interested in a simple function that provides a quick overview on results (with the option to set the number of permutations and whether it's a one- or two-tailed test).

Kind regards,

Herbert 




-------- Original-Nachricht --------
> Datum: Sun, 20 Mar 2011 12:53:04 -0400
> Von: Gabor Grothendieck <ggrothendieck at gmail.com>
> An: herbert8686 at gmx.de
> CC: r-help at r-project.org, Pierre.Legendre at umontreal.ca
> Betreff: Re: [R] Pearson correlation coefficient matrix with permutation test

> On Sun, Mar 20, 2011 at 10:45 AM,  <herbert8686 at gmx.de> wrote:
> > Hello,
> >
> > I found an interesting program on Pierre Legendre's webpage:
> >
> > http://www.bio.umontreal.ca/casgrain/en/labo/corr_permute.html
> >
> > With this program one can compute a "Pearson correlation coefficient
> matrix with permutation test".
> >
> > This is exactly what I need as an R-package because so far I have only
> analyzed my data with the function cor(). However, I need additional
> permutation tests for each individual correlation summarized in a matrix of
> correlation coefficients.
> >
> > Does anyone know an R-package where this is included as a function?
> >
> > Thanks for help!
> >
> 
> Section 3.2 of the vignette in the proto package shows computations in
> R for calculating confidence intervals for the correlation
> coefficient.   There are calculations shown for both the Fisher
> transform approach and bootstrapping.
> 
> 
> -- 
> Statistics & Software Consulting
> GKX Group, GKX Associates Inc.
> tel: 1-877-GKX-GROUP
> email: ggrothendieck at gmail.com

--



More information about the R-help mailing list