MedPub: A new Clustering-based approach for grading methods for Publishing Medical Data

Abstract

Most medical data publishing studies does not partition the semantics of the values of the sensitive attribute. This leaves the data prone to attacks such as homogeneity. This problem has been discussed and resolved in this paper by presenting a model that uses clustering based approach to partition the attributes values that are disease sensitive by grading methods for publishing medical data. The effectiveness of the proposed graded medical publishing method has been experimentally shown. Results shows that the new proposed model is nearly same as L-diversity in released information quality