We counsel caution in the application of ROC analysis for prediction of classifier performance under varying class distributions. We argue that it is not reasonable to expect ROC analysis to provide accurate prediction of model performance under varying distributions if the classes contain causally relevant subclasses whose frequencies may vary at different rates or if there are attributes upon which the classes are causally dependent.
About this article
Cite this article
Webb, G., Ting, K. On the Application of ROC Analysis to Predict Classification Performance Under Varying Class Distributions. Mach Learn 58, 25–32 (2005). https://doi.org/10.1007/s10994-005-4257-7
- model evaluation
- ROC analysis