Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: An index for determining cut points using ROC


From   Steve Samuels <[email protected]>
To   [email protected]
Subject   Re: st: An index for determining cut points using ROC
Date   Wed, 20 Jun 2012 21:38:03 -0400


You can compute both from first principles with Roger Newson's -senspec- (SSC)
or with any other command that outputs sensitivity and specificity. It's "Youden",
by the way, for WJ Youden.

http://www-history.mcs.st-and.ac.uk/Biographies/Youden.html 

Here's a code excerpt that you can adapt.

*******************************************************
senspec mets bmxwaist, sensitivity(sens) specificity(spec)
gen youden= sens-(1-spec)
egen youdenmax= max(youden), by(gender)

gen dist = sqrt((1-sens)^2 + (1-spec)^2)
egen distmin = min(dist), by(gender)

/* Identify cut-offs */
list sens spec  youdenmax   dist   ///
  bmxwaist if abs(youden -youdenmax)<0.0001
list sens spec  youden     distmin  ///
 bmxwaist if abs(dist - distmin)<0.0001
********************************************************
Steve
[email protected]

On Jun 20, 2012, at 9:13 PM, Alan Acock wrote:

I have a diagnostic test and want to predict which children will be school ready (0,1) based on the diagnostic test. ROC curves are great for this. There appear to be two indexes. Youdon's index and the point where the ROC curve is closest to 0,1.  Is there a way to compute either or both of these using Stata or a user written command? Findit youdon shows nothing.

Alan Acock

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


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


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index