no code implementations • 21 Jun 2013 • Tyler Sanderson, Clayton Scott
The first problem studied is that of class proportion estimation, which is the problem of estimating the class proportions in an unlabeled testing data set given labeled examples of each class.