I would like to know what are the differences between somers’d value and the Gini coefficient. I have the following values for factors:

Factor     Somers’ D

Factor 1   -0.1
Factor 2   0.24
Factor 3   -0.01

In terms of Gini, I should change signs from Somers’D column:

Factor     Gini

Factor 1   0.1
Factor 2   -0.24
Factor 3   0.01

Is it correct? Could you please tell me what changes between the two representations?

  • $\begingroup$ I think this article explains the connection pretty well: medium.com/james-blogs/… $\endgroup$
    – 1muflon1
    May 28, 2020 at 14:57
  • $\begingroup$ Yes, i already had a look at it. However it is still not clear if it is ok to write the values just changing the signs (from negative to positive; from positive to negative). What would it mean? How could I explain it? $\endgroup$
    – Math
    May 28, 2020 at 14:58
  • $\begingroup$ that is because if you would plot the functions that Somers' D uses are reflections of the ones you would use for direct gini calculation across the 45 degree line (see the first picture in that source. Hence the sign flips in a same way as sign of slope of a line would if you would make reflection of it parallel to x axis. $\endgroup$
    – 1muflon1
    May 28, 2020 at 15:06
  • $\begingroup$ Thank you 1muflon1. Now it is clear. So it is not an error as I wrote in the post, is it not? $\endgroup$
    – Math
    May 28, 2020 at 15:10
  • 1
    $\begingroup$ Thank you. Yes, I got these results from a simulation to determine predictive power. It was said that it should be enough changing the signs to get Gini values, but it was not clear to me the logic behind it $\endgroup$
    – Math
    May 28, 2020 at 15:21


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.