Quantification and Semi-supervised Classification Methods for Handling Changes in Class Distribution

In realistic settings the prevalence of a class may change after a classifier is induced and this will degrade the performance of the classifier. Further complicating this scenario is the fact that la