ROC curve analysis: sensitivity at fixed specificity

LgWagon — Sat, 04 Jan 2020 11:33:46 GMT

Dear all,

I'm a novice of JMP and I'm stunned with its great potential. I ask your help to solve my problem. I made a ROC analysis on a data set, and I need to calculate, based on ROC curve, sensitivity at fixed 80%, 95% and 99% specificity. The software doesn't seem to provide this option. I know I can approximate these values looking at the cut-offs in the ROC table, however it's only an approximation. Other software, for example MatLab or MedCalc give this option. For example MedCalc has the option to build "a table with estimation of sensitivity and specificity, with a BCa bootstrapped 95% confidence interval (Efron, 1987; Efron & Tibshirani, 1993), for a fixed and prespecified specificity and sensitivity of 80%, 90%, 95% and 97.5% (Zhou et al., 2002)." How can I solve?

Thank you very much, you are a great community!

Re: ROC curve analysis: sensitivity at fixed specificity

Thierry_S — Sun, 05 Jan 2020 19:40:36 GMT

Hi,
You may find what you need by expanding the "ROC Table" at the bottom the ROC plot.
Best regards,
TS

topic Re: ROC curve analysis: sensitivity at fixed specificity in Discussions

ROC curve analysis: sensitivity at fixed specificity

Re: ROC curve analysis: sensitivity at fixed specificity