Research On Privacy Protection For Social Network Application

As the wave of the new round of productivity growth in the information revolution,the Internet has received more and more attention from people.What comes with it was the blowout development of the Internet application.As an important part of Internet applications,social networks have been accepted by more and more people and become a communication platform for people to make friends,publish information,and disseminate information.However,as the number of users of social networking application platforms increases,more and more users provide their own personal information to social networking application platforms through passive or active methods,such as personality,school,and date of birth.Information,such information usually contains the user's sensitive information.With the increase of user geometric times,there are more and more data information in the social network application platform,which makes experts and scholars in many fields hope to obtain a useful knowledge by mining and analyzing a large amount of data in the social network application platform.In addition,the openness of social network application platforms and other characteristics make users' sensitive information face the risk of leakage.Therefore,the paper analyzes the privacy information,attack methods,and commonly used privacy protection technologies in social networks,and conducts research on privacy leakage issues on social network edge weight.The main tasks include:(1)Aiming at the existing edge weight privacy protection technology has the problem that the attacker has limited background knowledge or the shortest path cannot be analyzed,this paper proposes an edge weight protection strategy based on differential privacy protection model.The strategy first constructs a query function based on the differential privacy protection model and the characteristics of social networks to form a differential privacy protection algorithm;and then constructs different subgraphs for the social network,and according to different subgraphs propose a perturbation scheme;finally different subgraphs are perturbed by different perturbation schemes to implement the differential privacy protection algorithm.It can defend against attackers with the greatest background knowledge,the shortest path of the perturbed social network remains unchanged,and its length is about the same.Ensure the security of the social network edge weight and the analyzability of the shortest path.(2)The edge weight protection strategy based on the differential privacy protection model is mainly aimed at the situation where the edge weight represent the specific sensitive information and the relationship between the nodes.When the edge weight represents the similarity of attributes between user nodes,the edge weight protection strategy based on the differential privacy protection model is proposed to protect the edge weight,and there exists the leakage of sensitive attributes of some target nodes.Therefore,an improved edge weight privacy protection strategy based on the differential privacy protection model is proposed to protect edge weight that represent the similarity of attributes among user nodes.This strategy improves the query function,subgraph construction,and perturbation scheme to satisfy the protection when the edge weight indicate the similarity of attributes and ensure the security of the sensitive attributes of the target node.