Message-ID: <4B6056C5.1050308@vanderbilt.edu>
Date: 2010-01-27T15:07:49Z
From: Frank E Harrell Jr
Subject: control of scat1d tick color in plot.Predict?
In-Reply-To: <ba66e7051001270628y726d4dcn5771bf2881002aa3@mail.gmail.com>
Mike Babyak wrote:
> Hi All,
>
> I have a quick question about using plot.Predict now that the rms package
> uses lattice. I'd like to add tick marks along the regression line, which
> is given by data=llist(variablename) in the plot call. The ticks show up
> fine, but I'd like to alter the color. I know the ticks are produced by
> scat1d, but after spending a fair bit of time going through documentation,
> it still isn't clear to me how to do this in the context of lattice.
> Guidance would be greatly appreciated.
>
> Thanks,
>
> Mike Babyak
> Duke University Medical Center
>
> My code using R 2.10.1/windows XP
>
> myx<-c(1,2,3,4)
> myy<-c(1,2,3,5)
>
> library(rms)
>
> d<-datadist(myx)
> options(datadist="d")
>
> myfit<-ols(myy~myx,x=T,y=T)
>
> p1<-Predict(myfit,myx =.)
>
> library(lattice)
>
> #change line to black
> line <- trellis.par.get("plot.line")
> line$col <- 1
> trellis.par.set("plot.line", line)
>
> plot(p1, data=llist(myx),col.fill="lightgray", lwd=1.5)
That's a good request. I've added a new argument scat1d.opts to
plot.Predict in rms. Default is list(frac=0.025, lwd=0.3) with col
computed. If you specify col inside the list, it will be used.
source('http://biostat.mc.vanderbilt.edu/tmp/plot.Predict.s') to get the
new version.
Frank
--
Frank E Harrell Jr Professor and Chairman School of Medicine
Department of Biostatistics Vanderbilt University