[R] 2 samples KS-test...

[Prof Brian Ripley]

On Mon, 16 Apr 2007, Stéphane CRUVEILLER wrote:

> Dear R-users,
>
> I have some trouble to perform a 2-samples KS test.
> Apparently my 2 samples are numerical (see below) but R

That says there are data frames, not numeric vectors. That would be easier 
to see if you printed the value (rather than use str(), which is often 
helpful but not so here to you at least).

You want

x = subset(mydata,Identified=="NO")$KD

etc, I think.

> complains that the "y" term is not....
>
> -------------------------------------------------------------------------------------
> > str(subset(mydata,Identified=="NO",select=KD))
> 'data.frame':   2889 obs. of  1 variable:
> $ KD: num  -0.272 -0.080 -0.311 -0.222 -0.346 ...
> > str(subset(mydata,Identified=="YES",select=KD))
> 'data.frame':   443 obs. of  1 variable:
> $ KD: num  -0.0772 -0.6635 -0.1283 -0.0748 -0.3036 ...
> >
> ks.test(x=subset(mydata,Identified=="NO",select=KD),y=subset(mydata,Identified=="YES",select=KD))
> Erreur dans ks.test(x = subset(mydata, Identified == "NO", select = KD),  :
>        'y' doit être numérique ou bien une chaîne de caractères donnant
> le nom de la fonction adéquate
> -----------------------------------------------------------------------------------------
>
> Thanks for any hint,
>
> Stéphane.
>
> Conf: R-2.4.1 on win XP sp2.
>
>

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595