Re: st: Question on ROC analysis

[Steven Samuels <sjsamuels@gmail.com>]

Megan,


I suspect an incomplete download of -somersd-.  But there could also be a mistake in your code.  As the FAQ request, please show us exactly what you typed, as well as what Stata responded.

I misread your original question and missed the implication that lower values of x are associated with lower values of y. To get a large AUC, do the analysis with -x. Or indicate that Y is greater than the cutpoint. Either -somersd- or -roctab-  will work (ylow vs -x or yhigh vs x), though -somersd- has the advantage, among others, that it takes probability weights.
***********************
sysuse auto, clear
gen y = 1/mpg
gen ylow = y<.05
gen yhigh = 1-ylow
gen x = weight
gen nx = -x

scatter y x
roctab ylow nx, graph   //graphs the curve for -x

roctab ylow nx   // produce  the AUC for -x
roctab yhigh x   // same

roctab ylow x    // AUC for x
di .50 + (.50 -r(area))  // AUC for -x


/* somersd */
somersd  ylow nx, tr(c)   // also produces the AUC for -x
somersd ylow   x, tr(c)   // AUC for  x
di .50 + (.50 - _b[x])    // also AUC for  -x

**********************

Steve
sjsamuels@gmail.com





On May 23, 2011, at 5:29 PM, Megan Deitchler wrote:

Thanks - this helped.  I can now graph what I want but am still having
trouble calculating the AUC.

I am trying to use Roger Newson's somersd package for this, using the
c transformation option.  I receive the following error:

      tidotforsomersd():  3499  tidottree() not found
                <istmt>:     -  function returned error

Any suggestions?

On Thu, May 19, 2011 at 3:19 PM, Steven Samuels <sjsamuels@gmail.com> wrote:
> 
> Roger -senspec- from SSC should do what you want.
> 
> Steve
> sjsamuels@gmail.com
> 
> On May 19, 2011, at 2:46 PM, Megan Deitchler wrote:
> 
> I am interested in carrying out a simple two variable ROC analysis.
> 
> I want to assess how well low values of my x variable predict a low value of
> my y variable (e.g. y variable with cutoff less than 200).
> 
> However, if I understand correctly, the conventional ROC analysis in Stata
> creates the ROC by using incrementally increasing values of x to predict y.
> 
> How do I adapt the analysis so that the AUC result I obtain will
> be consistent with the relationship I am interested in quantifying?
> 
> *
> *   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/

*
*   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/