[R] Odd behaviour with scale()

[Liaw, Andy]

The problem is that scale() returns a matrix, even if only a vector is
supplied.  Thus the regression model actually has a matrix (scaled xx) as
the predictor.  See:

> str(Data1)
`data.frame':   10 obs. of  2 variables:
 $ xx: num [1:10, 1] -1.486 -1.156 -0.826 -0.495 -0.165 ...
  ..- attr(*, "scaled:center")= num 5.5
  ..- attr(*, "scaled:scale")= num 3.03
 $ yy: num  2.1 3.1 4.1 5.1 6.1 7.1 8.1 9.1 10.1 11.1

and note that xx is not `num [1:10]',  but `num [1:10, 1]'.  If you give a
single column matrix as xx to predict(), it would work:

> predict(reg1, data.frame(xx=I(matrix(2:4, ncol=1))), se.fit=T)
$fit
       1        2        3 
12.65530 15.68295 18.71060 

$se.fit
           1            2            3 
4.582710e-16 6.513913e-16 8.510747e-16 

$df
[1] 8

$residual.scale
[1] 6.210772e-16

HTH,
Andy

> From: Anon.
> 
> Moi!
> 
> A student here has been getting a bit irritated with some 
> side effects 
> of scale() (OS is Windows XP, the behaviour occurs in R 
> 2.0.0, but not 
> 1.7.1).  The problem is that she scales a variable in a data 
> frame, then 
> does a regression, and tries to get some predictions for some 
> new data.  
> However, at this point she gets an error (see the example 
> below).  This 
> seems to be because the scaled variable in the new data frame 
> does not 
> have the center and scale attributes, but the one in the old 
> data frame 
> does.
> 
> The work-around is to put the scaled variable intro a new data frame, 
> which again won't have the attributes.  But it seems odd to me that 
> whether a scale()'d variable has attributes depends on where it's 
> placed.  I presume that this is because I'm not understanding 
> something 
> about the way R is working, rather than it being a bug.  Would anyone 
> care to enlighten me?
> 
>  > Data1=data.frame(xx=1:10, yy=2.1:12)
>  > Data1$xx=scale(Data1$xx)
>  >
>  > reg1=lm(yy~xx, data=Data1)
>  > New=data.frame(xx=2:4)
>  > b=predict(reg1, New, se.fit=T)
> Error: variable 'xx' was fitted with nmatrix.1 but numeric 
> was supplied
>  >
>  > New=data.frame(xx=scale(2:4, center=5.5, scale=3.02))
>  > b=predict(reg1, New, se.fit=T)
> Error: variable 'xx' was fitted with nmatrix.1 but numeric 
> was supplied
>  >
> 
> Bob
> 
> -- 
> Bob O'Hara
> Department of Mathematics and Statistics
> P.O. Box 68 (Gustaf Hällströmin katu 2b)
> FIN-00014 University of Helsinki
> Finland
> 
> Telephone: +358-9-191 51479
> Mobile: +358 50 599 0540
> Fax:  +358-9-191 51400
> WWW:  http://www.RNI.Helsinki.FI/~boh/
> Journal of Negative Results - EEB: www.jnr-eeb.org
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide! 
> http://www.R-project.org/posting-guide.html
> 
>