Message-ID: <200411301047.40509.deepayan@stat.wisc.edu>
Date: 2004-11-30T16:47:40Z
From: Deepayan Sarkar
Subject: adding regression curve to xyplot
In-Reply-To: <E7D5AB4811D20B489622AABA9C53859104E0DBD9@teal-exch.amgen.com>
On Tuesday 30 November 2004 10:18, Austin, Matt wrote:
> Inside the panel, you have to only use the parts of x and y for each
> unique combination of
> factor(lonLabels[whichLon100])*factor(latLabels[whichLat100])
>
>
>
> So use "thislm <- lm(x[subscripts]~y[subscripts])" in the panel
> function.
That's not true (at least in this situation). The 'x' and 'y' passed to
the panel function by xyplot are already the appropriate subsets.
('groups' would have been passed whole, had it been there.)
Deepayan