Research On Local Randomization Anonymous Method For Community Detection

In recent years,with the rapid development of society,especially the rise of social platform,more and more users join the online social platform,which makes the rapidly increase in the amount of social network data,not only to facilitate the individual and groups to better communicates,but also to facilitate the relevant scientific research institutions on the social network for more detailed researches and analysis.However,some data can not be directly available to the relevant researchers,because the data may contain some sensitive information,such as name,user relationship,wages,etc.If the original data is published directly,it will be likely to leak user privacy.Therefore,we must process the corresponding sensitive data to protect the privacy before publishing the data,but these privacy policies may largely modify the structure information in the original social network.Therefore,how to balance data privacy and data utility is a hot issue in social networks.The current social network data publishing method is mainly given a primitive social network map to do with the corresponding privacy anonymous protection before publishing.But it ignores the following questions:(a)only consider the privacy and security,did not take into account the social network map information changes;(b)the original social network map may contain communities,not the corresponding sub-community privacy and security in detail.The above two points make the practicality of the data lower.The more detailed the published social network map is,the more beneficial to the relevant social network analysts,this paper through the community detection algorithm on the sub-community structure(node degree)to do the corresponding privacy protection analysis,community-oriented social network privacy protection to do the appropriate research.The mainly work of this paper is as follows:Firstly,through analyzing the current social network privacy protection method,we found its shortcomings.This paper uses the social network structure(degree of node)as the background knowledge of the attacker,the traditional k degree anonymous method and the randomization method in the process of privacy protection does not fully consider the original social network structure diagram,and the original social network map exists multiple sub community,some of the edge of the connection can be divided into sub-community connections within the sub-community connection.In the process of privacy protection,it may destroy the structure of the original social network,such as the addition of some social network nodes to delete or add the deletion.In the process of k degree anonymous and randomizations,there may be a lot of uncertain graphs that disrupt the practicality of the data.Secondly,for the shortcomings of the above two traditional methods,this paper presents a new local randomized perturbation method for the protection of community structure information.In this method:first based on the community detection process,record the edge of the document.When the community finishes the completion of testing,we judge it according to the background knowledge of the attacker(degree of node)to determine whether there is privacy disclosure,if not,that the sub-community does not exist privacy disclosure problems,do not deal with;if there is privacy disclosure,we determine whether the privacy of the leaked node connected within the sub-community,if the community is inside the edge of the privacy of the community is leaked,then the probability of deletion or add edge,the use of randomization of the disturbance;If the edge of the leaked connection is connected to another community,it indicates that the node has a margin,indicating that there is a margin on the node,adjusting the probability that the edge is deleted,increasing the probability of being deleted,and then performing random processing operations in the subcategory.Through this method,the original appearance of the social network map is guaranteed to a great extent.For any sub-community,the social network structure of the sub-community has also been better guaranteed under the premise of guaranteeing the privacy requirement,so that the relevant scientific research personnel published social network map for related research and analysis.Finally,We use the real data set to verify the feasibility and utility of the proposed method,this method can guarantee the privacy and security at the same time and guarantee the structural characteristics of social networks better.