Research On Dynamic Social Network Anonymity Technology For Protecting Community Structure

With the continuous progress of Internet technology and the continuous improvement of smart phone R&D and innovation capability,Social media popularizing rate has soared.Social network applications becomes an important part of daily social interaction and provides users with a full range of service models.In the era of big data,massive social network data with great research value attracts researchers.Because the social network data that needs to be published and shared inevitably carries the privacy information of users,it is imperative to protect the privacy of the data before publishing the social network.In addition,the social network community structure can reflect the relationship between its structure and function well,so it has certain practical significance to protect the social network community structure while protecting the privacy of social network data.The current social network privacy protection methods have some shortcomings,such as serious destruction of community structure and low data processing ability of single workstation.The paper proposes a social network anonymity model for protecting community structure and a distributed social network degree anonymity algorithm to protect community structure.Social network community detection uses divide and agglomerate algorithm.The method obtains the grouping and anonymous results based on compressed binary tree constructed by aggregate vector.The purpose of adding virtual vertices to the original graph is to obtain anonymous social network.In order to improve the data availability of published graph,the paper designs a distributed virtual vertex removal-addition algorithm according to the community to which the vertex belongs.Aiming at the problem of vertex identity re-recognition in dynamic social networks,the paper defines a vertex degree sequence attack model of dynamic social networks.On this basis,the paper proposes an anonymity model of dynamic social network degree sequence,and then proposes a dynamic social network degree sequence anonymity algorithm to protect the community structure.A divide and agglomerate algorithm is expanded for community detection in dynamic social networks.Different grouping anonymous methods are proposed for different types of dynamic social network vertices.Virtual vertices are added in parallel to construct the initial anonymous graph,and the initial anonymous graph is reconstructed to reduce the loss of published graph information.Two privacy protection algorithms are tested and analyzed based on real social network data sets.Algorithms are implemented based on GraphX,a large-scale parallel graph processing system.The experimental results shows that the proposed social network anonymity algorithms can overcome the defects of the traditional algorithm in the community structure protection area and ensure the availability of data while meeting the requirement of anonymity.