Hello everyone, I have been using scikit for my dissertation and so far it's been great! However, I often need to compute confidence intervals for auc values and I have not been able to find a scikit method for this. I have done some research on the maths behind auc confidence interval computation and it would be great if scikit could offer the option of computing auc conf intervals using at least a few different methods (delong, bootstrap, cortes etc). It would probably be useful for researchers trying to compare classifier performance. Perhaps this could be a good project for the sprint?
Kind regards, Camilla
------------------------------------------------------------------------------
_______________________________________________ Scikit-learn-general mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
