抽象的

Sanitations To Prevent Inference Attack On Social Network Data

Chandra D,Antony Rosewelt.L

Social Networks, like Facebook, are used by lots of people. These networks allow users to share specific of them and hook up with their friends. Many of the information revealed inside these network is designated to be private. Yet it’s possible that corporations can use learning algorithms on released data to calculate undisclosed private data. This paper explore how you can launch inference attacks using released online community data to calculate undisclosed private information about individuals, including their political affiliation or sexual orientation. Then devise three possible sanitation techniques that may be found in various situations. Then, the potency of these methods by implementing them with a dataset from a certain nation-state in facebook online community application and looking to use ways of collective inference to learn sensitive features of the information set. And also this paper shows that have possibilities where the strength of both local and relational classification algorithm may be minimized with the sanitation methods as described.