Security Breach in Anonymized Social networks

Abstract

The purpose of the study was to show a threat of privacy leak for members of social networks. The general way to protect the privacy of these networks is using anonymization methods. On the other hand, deanynimization methods exist against these anonymization algorithms. Those methods were introduced and one of them was explained in details. Improvements of the algorithm were suggested by changing different parameters of it based on the social networks characteristics. These parameters were limiting similarity score, eccentricity, sorting, degree limit, ordering, and degree comparison and running the algorithm in two stages. The result of the experiment showed that by choosing proper values for similarity score, eccentricity, degree and also sorting the nodes, we can improve the performance of the algorithm by preventing it from propagating in the wrong direction. Sorting the nodes will also improve the result with a similar reason.