[R] rms package-superposition prediction curve of ols and data points

[Sarah Goslee]

You could use points() instead of plot() for the second command.

Sarah

On Tue, Jun 26, 2012 at 8:37 AM, achaumont <agnes.chaumont at live.be> wrote:
> Hello,
>
> I have a question about the “plot.predict” function in Frank Harrell's rms
> package.
> Do you know how to superpose in the same graph the prediction curve of ols
> and raw data points?
> Put most simply, I would like to combine these two graphs:
>
>>  fit_linear <- ols (y4 ~ rcs(x2,c(5,10,15,20,60,80,90)), x=TRUE, y=TRUE)
>> p <- Predict(fit_linear,x2,conf.int=FALSE)
>> plot (p, ylim =c(-2,0.5), xlim=c(0,100))              # graph n°1
>
>> z <- plot (x2,y4,ylim=c(-2,0.5),xlim=c(0,100),type="p",lwd=6,col="blue")
>> # graph n°2
>
> Thanks all,
>
> Agnès
>
>
>


-- 
Sarah Goslee
http://www.functionaldiversity.org