Mrs Mehri Rajaee is going to defense his Phd thesis on “Preserving Privacy in Social Network Data Publishing with Preserving of an Acceptable Level of Data Utility” on Saturday January 02, 2016. The session will be held in Phd defense hall, Computer Engineering Department, Iran University of Science and Technology, Tehran, Iran.
Nowadays, huge amounts of social network data are generated.These data have graph structure and contain useful information about interactions between members of society. So, their analysis has been interested by political, social, economical, geographical, management analysts. To protect the privacy of individuals, data owners could not share them with analysis, since social network data contain sensitive and private information about individuals. One solution to overcome this problem is to safely transform original network data by anonymization operations to anonymous release of data and publish them. But the published data should permit useful analysis while protecting privacy. This problem is known as privacy-preserving network data publication.
In this dissertation, we propose total framework for above problem for directed network data whose nodes contain attributes. In the proposed framework, we investigate the problem from base. First, we propose a privacy model special for network data. The purpose is to prevent disclosure of presence, sensitive attribute, degree and relationship (link). Second, an anonymization technique based on anatomization that specifies the format of published data is proposed. The data that stored based on our proposed anonymization technique have the ability to prevent above four disclosures. Third, we propose a greedy anonymization algorithm to transform the original data to the specified format of anonymization technique. This algorithm preserves privacy of all members of social network under specified thresholds, and also preserves data utility at an acceptable level.
We did some experiments on real and synthetic datasets to evaluate data utility of four kinds of queries (aggregate tabular query, aggregate network query, graph topological and spectrum properties). Experimental results show that our proposed approach and framework make good balance between privacy and data utility.
Keywords: privacy, network data, social network, anonymization, publish data, sensitive attribute, data utility, information loss
Phd cadidate:Mehri Rajaee
Supervisor: Professor Dr.Haghjoo, Dr. Khanjari .
Jury Committee: Professors Dr. Rasoul Rajaee,، Dr. Masoud Rahghozar, ، Dr. Saeed Parsa, Dr. Behrooz Minaee
Time: Location: 15:00 PM, Saturday January 02, 2016, Phd defense hall, Computer Engineering Department, Iran University of Science and Technology, Tehran, Iran