[R] Regression on presence/absence matrix

[Ben Bolker]

Daniel Patón Domínguez <d.paton.d <at> gmail.com> writes:

> 
> > The library of packages that installs with R includes the stats
> > package, in the stats package is the glm function for fitting
> > generalized linear models.  Using glm with a binomial family will fit
> > a logistic regression which can be used as you describe.
> > 
> > If you really feel the need to use an additional package then lrm in
> > the rms package is another option (and I am sure there are others as
> > well).
> 
> Dear colleague:
 
> I know both possibilities but I am speaking about a relationship
> between two sets of presence/absence data. Usually I use Canonical
> Correspondence Analysis for this but I am not sure if a specific
> package of R is designed for this type of data
 
  If you know you want to do canonical correspondence analysis, you
could look for it as follows:

    library("sos")
    findFn("{canonical correspondence}")  ## whole phrase
    # or 
    findFn("canonical correspondence")  ## separate words

which indicates that you should be probably looking first
at the 'anacor' and 'vegan' packages.