Research On Privacy Preserving Techniques Based On K-Anonymity For Data Publishing In The Social Network

With the development of information technology,the number of social network users is increasing.Users can easily communicate,interact and share information on social networks.At the same time,the information of social network users is recorded and collected by major social network platforms.In order to better provide people with personalized and targeted services,social network data is usually released to third parties for data mining and analysis.However,the published data contains the user's private information.A malicious attacker can re-identify and resolve the published social network data through the acquired background knowledge,thereby causing the user's privacy leakage.Therefore,before the social network data is released,it needs to be anonymized,so that the published data meets the requirements of the privacy protection intensity,and the data availability is maintained for data analysis.Therefore,this paper studies the privacy protection technology based on k-anonymity social network data release.The main work includes:(1)This paper introduces the research background and significance of social network data release,analyzes the current situation of social network data release privacy disclosure and the development of social network privacy protection technology,and summarizes the domestic and international social network data release privacy protection technology.Research status.The theory and privacy protection technologies related to social networks are summarized and introduced,and the privacy protection methods and techniques commonly used in social networks are introduced.(2)For the randomization technology,the social network map is modified more,which is easy to cause the data utility of the anonymous data to be drastically reduced.The community partition is introduced into the randomization technology,and an improved kdegree anonymous privacy protection method is proposed.-subgraph.The community partitioning step is used to constrain the random disturbance range and protect the graph feature structure of the original data.Experiments show that compared with KDLD algorithm and PK(partial k-anonymity)algorithm,k-subgraph algorithm has smaller information loss rate,average path length and node average degree.(3)For the existing social network attribute privacy protection algorithm,the same strength privacy protection is applied to all attributes,which is easy to cause excessive anonymity.Entering the influence analysis of the node,a social network attribute personalized privacy protection method based on node division is proposed.D-KDLD.According to the influence of the node,the method divides the node into key nodes and edge nodes,and uses the data generalization and node segmentation methods to conceal the attribute information.The experimental results show that the D-KDLD method has less information loss than other detection methods.The innovations in this article include:(1)Propose an improved k-degree anonymous privacy method.The method divides the scope of random disturbances by community division,and protects the graph feature structure of the data,thereby improving the data utility after anonymity.(2)A personalized privacy protection method based on node partitioning for social network attributes is proposed.According to the influence of the node,the method divides the node into key nodes and edge nodes,and uses anonymous attribute information in different ways to reduce the loss of data information.