Research On Social Network Data Publishing Method Satisfying Differential Privac

With the widespread use of intelligent devices,a large amount of valuable information has been generated in social networks and resides in social networks,which plays an important role in promoting the development of big data and artificial intelligence.However,publishing social network data may leak sensitive privacy information of users,which conflicts with privacy protection requirements,as protecting privacy requires restricting data publishing.Therefore,how to protect user privacy when publishing social network data has become an important issue.The existing methods for publishing social network data cannot preserve the attributes of social networks(such as clustering coefficients)due to excessive noise injection,or cannot resist attackers with background knowledge,thereby affecting the use of social network data.In order to overcome these shortcomings,this paper proposes two social network data publishing methods that meet Differential privacy,aiming at the problems of privacy protection and data availability in social network data publishing.(1)A d K sequence social network data publishing method that satisfies differential privacy.Aiming at the privacy and utility issues in social network data publishing,this paper proposes a social network data publishing method that satisfies differential privacy by combining social network d K statistical information and differential privacy technology.Firstly,statistical information on d K sequences in social networks is generated based on the distribution characteristics of nodes and edges of social networks.Secondly,the generated d K sequences are constructed using a Laplace perturbation algorithm using the perturbed d K sequences.Finally,in order to reduce the introduction of noise and improve the availability of data,a disturbance optimization algorithm is proposed.This optimization algorithm introduces the idea of grouping and groups the d K sequence before adding noise.From the level of each group after division,disturbances are added according to the sensitivity of different groups,thereby reducing the introduction of noise and improving the availability of data.It achieves the goal of social network data while meeting differential privacy,High data availability.(2)A structure attribute social network graph data publishing method that satisfies differential privacy.A structure attribute social network data publishing model under privacy protection is proposed to address the issues of single data availability caused by single structure protection and low data availability caused by multiple disturbances to structure protection in social network privacy protection.The model consists of two stages.The first stage is to protect the user attributes.The binary discrete method is used for the user attribute data to improve the operation efficiency.The discretization attribute data is flipped with probability to protect the user node attribute privacy.The second stage is to protect the structure of social networks by using community partitioning methods to reduce data sensitivity and reduce the introduction of noise.We also optimized the community partitioning method to improve operational efficiency.Next,in order to preserve the structural information of social networks to a greater extent,we introduce uncertainty graphs,and finally perform differential privacy protection on the generated edge information.The structure attribute social network data publishing method not only protects node and edge information,but also preserves the community structure to the maximum extent,ensuring data availability while protecting privacy.