Date:         Tue, 21 May 2002 16:31:49 -0400
Reply-To:     Doc Muhlbaier <lawrence.muhlbaier@DUKE.EDU>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Doc Muhlbaier <lawrence.muhlbaier@DUKE.EDU>
Organization: Duke University, Durham, NC, USA
Subject:      Re: Area under the curve

If you are looking at ROC curve as in logistic regression, the c-index is
the area under that curve.

Doc Muhlbaier

"Jay Weedon" <jweedon@earthlink.net> wrote in message
news:0a8leug696vdc5ksroroe0mbclvvabtnkg@4ax.com...
> On 21 May 02 16:34:03 GMT, santoshkverma21@YAHOO.COM (Santosh Verma)
> wrote:
>
> >If I have x and y points or to be more specific 'senstivity' and '1-
> >specificity'. How can I get area under the curve?
>
> There's more than one way to do this, as you'll see by reading a text
> on Receiver Operating Characteristic analysis. The simplest method is
> probably to use the nonparametric "trapezoid rule":
>
> The area increment between consecutive points (x1,y1) & (x2,y2) is
> given by
>
> (x2-x1)*(y1+y2)/2.
>
> JW