In binary classification, the precision-recall curve (PRC) has become a widespread conceptual tool for assessing classification performance. The curve relates the positive predictive value of a classifier to its true positive rate and provides a useful alternative to the well-known receiver operating characteristic (ROC). A smooth estimate of the PRC can be computed on the basis of a simple distributional assumption about the underlying decision values.
This archive contains an easy-to-use MATLAB implementation of this approach.
For full details, see:
K.H. Brodersen, C.S. Ong, K.E. Stephan, J.M. Buhmann (2010)
The binormal assumption on precision-recall curves.
Proceedings of the 20th International Conference on Pattern Recognition (ICPR), 4263-4266.