Bipartite ranking refers to the problem of learning a ranking function from a training set of positively and negatively labeled examples. Applied to a set of unlabeled instances, a ranking function is
