How do I plot a ROC graph of True Positive vs. False Negative?

I have a vector of length 1,000 with mostly 1s (True Positives) and some 0s (False Negatives).

Should I be using

• plotroc()
• perfcurv()
• or just plot() ?

and what should I be plotting my vector against?

Thanks for any help

Answer 1

[Thorsten]

The idea of ROC analysis is to have find TP and FP for different, say, 100 thresholds, and then

plot(FP, TP)

where FP and TP are 1x100 vectors with the number of TP's and FP's for each threshold. Just one vector does not allow you to plot a ROC curve, as far as I can see.