Enhancing De-anonymization Attacks on Graph Data

Please enable javascript in your browser.

Aliakbari, Javad | 2019

480 Viewed

Type of Document: M.Sc. Thesis
Language: Farsi
Document No: 51644 (05)
University: Sharif University of Technology
Department: Electrical Engineering
Advisor(s): Mohajeri, Javad; Salmasizadeh, Mahmoud
Abstract:
Social networks and the shared data in these networks are always considered as good opportunities in hands of the attackers. To evaluate the privacy risks in these networks and challenge the anonymization techniques, several de-anonymization attacks have been introduced so far. In this thesis, we propose a technique to improve the success rate of passive seed based de-anonymization attacks. Our proposed technique is simple and can be applied in combination with different types of de-anonymization attacks. We show that it can achieve high success rates with low number of seeds compared to similar attacks. Our technique can also be used for applying partial attacks on graphs which results in high success rate. We show the result of applying our technique on one of the best passive seed-based de-anonymization attacks introduced by Ji et al. The results prove our claims
Keywords:
De-anonymization Attack ; Social Networks ; Graph Data Structure ; Extended De-anonymization Attacks ; Privacy ; Network Security

No TOC