Research On Algorithms Of Utility-based Privacy-preserving For Social Network Users

Mobile Internet,social networks have been integrated into people's daily lives increasingly,especially with the development of the Internet of things,the diversity of personalized service,more and more privacy information of the social users is active or unintentional exposed to the network.In addition,the social network data bring more benefits for network application providers,but also provide the motivation to malicious attackers.Therefore,the research on the privacy protection of user information has great theoretical and practical significance.In recent years,researchers have proposed a variety of anonymous models and algorithms for privacy protection in social networks.However,the anonymity algorithms have to disturb social network data to achieve privacy protection,while reducing the accuracy of data mining analysis.Therefor the privacy security and the data availability are two contradictory goals of privacy protection.It is an important and challenging problem to study the trade-off between privacy and usability.In addition,compared to the study of information privacy security,the study on data utility is not yet mature enough,and there is no standard data utility loss measurement.Therefore,we take the user attribute in the social network as the research object,and study the anonymous model and algorithm which can effectively reduce the data utility loss.The research contents mainly include the following three parts:(1)In view of the privacy protection problem in the social network scenario,the existing anonymous algorithms generally use the modified number of the edge or node as the only criterion for assessing the loss of anonymous data utility.These algorithms have a higher degree of privacy protection,but may lead to excessive loss of data utility and affect the use value of anonymous data which because of the different modifications of the edge or node have different effects on the social network structure.Considering this problem,this paper designs a more comprehensive data utility measurement method UL(G,G').It is from the perspective of structural similarity and information integrity of data utility.The method comprehensively evaluates the anonymous operations on the network structure and data content.It reduces more utility loss of anonymous data than those in the past,only by the number of man-made changes.(2)In order to protect social user information,we improve the anonymous model of k-degree-l-diversity,and propose a differentiated privacy protection model(d,k,l)-u.Based on this model,we design a kind of attribute differentiation anonymity algorithm.According to the sensitivity function,the algorithm divides the attribute value of the sensitive attribute into high,medium and low three privacy anonymous groups,and different anonymous groups will use different anonymous rules.The algorithm corrects the privacy protection object from the attribute class to the specific attribute value,and reduces the utility loss of the anonymous data.And we verify the validity of the algorithm through the simulation experiment.(3)Considering that the user nodes in the social network have different influence.The split operation of the key nodes such as "bridge nodes" may change the overall characteristics of the network structure.Therefore,we propose an anonymous algorithm of node differentiation(DKDLD-U).The algorithm introduces the key node analysis in the social network analysis,and divides the nodes into important nodes and ordinary nodes.The attribute value generalization and node splitting two kinds of anonymous operation are used respectively in order to reduce the disturbance to the network structure and improve the data utility.The simulation results show that the proposed algorithm can effectively reduce the utility loss of anonymous data while ensuring privacy and security.