# [R] roc and lattice

Deepayan Sarkar deepayan.sarkar at gmail.com
Wed Jan 10 09:00:07 CET 2007

```On 1/9/07, rhk <mailing-lists at rhkoning.com> wrote:
> Hello, I am afraid I do not fully understand all intricacies of
> programming in lattice plots. In the code below I try to plot an ROC
> curve, following R-news 4(1). When I condition on the variable 'group' I
> get the error message below, when I plot the curve for all data (i.e., y
> ~ pred.prob), I get the plot I want. Can someone point out why
> conditioning gives that message? Thanks, Ruud
>
>  > plot.a <- xyplot(y ~ pred.prob|group, data=x.df,
> +  xlim=c(0,1),xlab="1-specificiteit",
> +  ylab="sensitiviteit",
> +  panel=function(x,y,subscripts,...){
> +   DD <- table(-x,y)
> +   sens <- cumsum(DD[,2])/sum(DD[,2])
> +   mspec <- cumsum(DD[,1])/sum(DD[,1])
> +   panel.xyplot(mspec,sens,type="l",...)
> +   panel.abline(0,1)
> +  })
>  > print(plot.a)
> Error in panel(x = c(0.000265710002003069, 0.000345712857778025,
> 0.000265710002003069,  :
>         subscript out of bounds
>
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> and provide commented, minimal, self-contained, reproducible code.

Since you haven't bothered to give us a reproducible example, all we
can say is that somewhere in your code, a subscript is out of bounds.
Since the only place where you use subscripts is as DD[,2] etc, I
would start by checking if DD indeed has two columns and two rows as
you seem to expect.

Deepayan

```