Abstrakt

Certain Investigations on Approaches for Protecting Graph Privacy in Data Anonymization

S.Charanyaa T.Shanmugapriya

Developing privacy preserving mechanisms for data sharing across network for research purposes and business decisions has become one of the issues of the days research interest. L.Sweeney et.al., (2002) [26] developed the concept of k-anonymity, a model for protecting privacy which poses the condition that a database to be kanonymous, then each record is indistinguishable from at least k-1 other records with respect to their quasi-identifiers. Despite the k-anonymity model, an intruder may gain access the sensitive information if a set of nodes share similar attributes. In this paper we systematically analyze the pure structure anonymization mechanisms and models proposed in the literature. Also we make a detailed study on k-degree-l-diversity anonymity model, which takes into consideration the structural information and sensitive labels of individuals as well. Also the study the algorithmic impact of adding noise nodes to original graph and the rigorous analyses on the theoretical limitations of the appended noise nodes and its impact.