Exploiting Unlabeled Data for Improving Accuracy of Predictive Data Mining

doi 10.1109/icdm.2003.1250929