I am currently using pROC in R to construct ROC curves and extract optimal coordinates for binary classification.

I want to set a manual sensitivity of 90% however I receive NA for this threshold when I run the following code yet performance metrics (sensitivity, specificity, etc.) are still being returned:

pROC::coords(roc_curve, x = 0.9, transpose = FALSE, 
      ret = 'all', input = "sensitivity")

Is there a reason the threshold for this sensitivity is not returned while other metrics are? Unfortunately, the next nearest sensitivity is either 95% or 82%. I was contemplating whether interpolation for these sensitivities is possible or if another method would be more appropriate?



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.