Re: st: ROC-curves

[Marta Garcia-Granero <mgarciagranero@gmail.com>]

Hi Ragnhild:

Since I did not see any replies,  I'm giving you some ideas (the ones I 
use):

a) You can get the 95%CI for the AUC using bootstrapping:

program bootroc, rclass
    quietly logit "your model goes here"
    quietly lroc, nograph
    return scalar area=r(area)
end

bootstrap roc=r(area), reps(1000) : bootroc
stat bootstrap, all

b) you can save the predicted probabilities using  "predict double prob, 
xb". Then you can use -rocreg- with "roc(1-spec value)" option. Also, 
the predicted probabilities can be used with other roc commands that 
give 95%CI with different methods (Hanley, deLong...)

Take a look at this page:

http://www.stata.com/features/overview/receiver-operating-characteristic/

Anyway, I'm sure that others can give more detailed explanations, and 
better ideas than these ones.

HTH,
Marta Garcia-Granero

El 14/10/2013 18:54, Ragnhild Bergene Skråstad escribió:
> Hi!
> I investigate how diffrent tests, in combination, can predict a given 
outcome.
>
> I have made a logistic model with the command "logistic" and plotted 
the ROC-curve with the command "lroc". This cave me the ROC-curve and 
the AUC. I wonder:
> - how can I get the 95 % CI for this AUC?
> and
> - I would like to get the sensitivity at a given fixed false-positive 
rate. Do I have to get all the coordinates on the ROC curve and identify 
the one at the FPR at interest- and if so, how do I do that, or is it a 
direct way to do this?
> best wishes
> Ragnhild B Skråstad

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/