My suggestion is that you look at the code for panel.loess, and then ask
yourself what made you think this could possibly work.
- what happens to arguments in ... ?
- does panel.loess call loess?
- does loess accept a se=TRUE argument? Does loess.smooth?
You need to write your own panel and prepanel functions, calling loess and
then predict -- see ?predict.loess.
On Sat, 8 May 2004, Luca De Benedictis wrote:
> Dear R-friends,
> in plotting x and y, conditioning on z, I am trying to find a way to
> make the standard errors appearing together with the loess in the
> trellis plot.
> I have tried the following code without success.
>
> x <- x
> y <- y
> z.two.groups <- equal.count(z, number=2, overlap=.1)
> > xyplot(x~y | z.two.groups,
> + prepanel = function(x, y)
> + prepanel.loess(x, y, span = .75, se = TRUE),
> + panel = function(x, y){
> + panel.grid(h = 2, v = 2, lwd = 2)
> + panel.xyplot(x, y, cex = 0.6)
> + panel.loess(x, y, span = .75, se = TRUE)
> + },
> + par.strip = list(cex = .75),aspect = 2,
> + xlab = list("x",cex = 1),
> + ylab = list("y",cex = 1),
> + main=list("", cex=2)
> + )
>
> Any help would be appreciated
--
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