A personalized trajectory privacy protection method

Abstract

Trajectory data of sports or activities are usually collected and shared into social apps like Wechat moments, Sina weibo in public to provide health services and recommendation, while a large number of friends with weak ties in social circle will cause privacy leakage of users’ locations and life habits. To solve the problem, a personalized trajectory privacy protection scheme based on relationship strength called PTPP is proposed, the location obfuscation algorithm based on noise radius limiting geo-indistinguishability and location clustering is explored. Not only is privacy protected, but privacy budgets are controlled in fine grain according to relationship strength between users. Meanwhile, a hybrid calculation model of social relationship strength called HCM is proposed, which combines clustering and BP neural network and improve the reasonableness of social relationship strength. Finally, the availability and security of the PTPP algorithm are analyzed in the application scenarios of social networks. Analysis and the experimental results show that the method proposed could evaluate the strength of the relationship between users effectively and achieve personalized trajectory privacy protection.