1

If I have 1 independent variable (continues) and 1 dependent variable (binary), I can conduct logistic regression and ROC analysis, and I can get a cut-off point of independent variable using ROC analysis.

But, If I have 2 independent variable, how can I get two cut-off point in multivariate ROC analysis.

In STATA, I can get just maximal AUC score but I can't get the cut-off points each independent variables.

leetaey
  • 33
  • 1
  • 1
  • 5

1 Answers1

0

You can't, because with logistic regression you have an infinity of cut-off points that will correspond to each cut-off of the ROC curve. See this answer for more details.

Community
  • 1
Calimo
  • 2,829
  • 17
  • 26