Re: [R] area under roc curve

From: Frank Harrell <>
Date: Wed, 13 Apr 2011 15:20:23 -0700 (PDT)

ROC area does not measure goodness of prediction but does measure pure predictive discrimination. The generalization of the ROC area is the C-index for continuous or censored Y. See for example the rcorr.cens function in the Hmisc package.

agent dunham wrote:
> Dear all,
> I want to measure the goodness of prediction of my linear model. That's
> why I was thinking about the area under roc curve.
> I'm trying the following, but I don't know how to avoid the error. Any
> help would be appreciated.
> library(ROCR)
> model.lm <- lm(log(outcome)~log(v1)+log(v2)+factor1)
> pred<-predict(model.lm)
> pred<-prediction(as.numeric(pred), as.numeric(log(outcome)))
> auc<-performance(pred,"auc")
> Error en prediction(as.numeric(pred), as.numeric(log(outcome))) :
> Number of classes is not equal to 2.
> ROCR currently supports only evaluation of binary classification tasks.

Frank Harrell
Department of Biostatistics, Vanderbilt University
View this message in context:
Sent from the R help mailing list archive at

______________________________________________ mailing list
PLEASE do read the posting guide
and provide commented, minimal, self-contained, reproducible code.
Received on Wed 13 Apr 2011 - 22:23:33 GMT

Archive maintained by Robert King, hosted by the discipline of statistics at the University of Newcastle, Australia.
Archive generated by hypermail 2.2.0, at Wed 13 Apr 2011 - 22:30:30 GMT.

Mailing list information is available at Please read the posting guide before posting to the list.

list of date sections of archive