Protection Scheme For Link Privacy In Social Network Analysis

With the rise of social networks,more and more Internet data applications,malicious data mining based on the useful information,which may contain the user's sensitive information,so it needs to be processed anonymously before the data is published.One way of anonymizing the social network data is by removing all the attributes,and leaving only the social network structure.However,if the attacker has a certain degree of awareness of the structure information around the user,it can re-identify the user according to the structure information,and then obtain the user's identity information.How to protect the privacy of users effectively in social networks more and more attention.The main research work of this paper is as follows:1.Network link privacy protection scheme based on neighbor randomization algorithm.Neighbor randomization algorithm with thefalse destination to replace the original destination node in social network,and the choice of false destination will affect the availability of the network.In this paper,we introduce the important parameters of the social network,such as degree centrality,betweenness centrality and clossness centrality,the false destination are selected according to the weight of the parameters,thereby changing the node selection mode.From the perspective of security parameters,the number of destination decoy sets,the radius,computational degree centrality,betweenness centrality and clossness centrality Spearman similarity,and get more conclusions about preserving chart usability.2.Important nodes Protection Scheme Based on k-shell Decomposition.Because important nodes have a significant impact on network,an attacker can take control of the network's core nodes to quickly affect the network,thereby imposing more protection on important nodes.In this paper,the importance of network nodes is judged according to k-shell decomposition,and the probability that the correct connections of nodes with larger ks values are replaced is increased according to certain rules in order to achieve more perturbed to such nodes.In addition,this paper calculates the probability of replacement of connections based on the degree.This not only gives you more protection for important nodes,but also more for the retention of graph usability.This is because,under the condition of satisfying a certain safety parameter,the total number of the replaced edges can reach a certain value.Due to more disruption of important nodes,the link of more nodes can be preserved,so the algorithm can effectively reduce the changes to the graph.3.Novel k-out anonymous privacy protection algorithm.Link perturbation with the false destination to replace the original destination node in social network,the out-degree did not change,therefore,attacker can still correctly identify the user identity information.For this problem,this paper improves the k-out anonymous privacy protection algorithm.which is different from the traditional k-degree anonymous model:grouping—by calculating the edge loss,determining whether the anonymous node is classified or starting a new group;the best degree —through the dichotomy of this idea,the group makes the least change on the edge of the change;the node selection—if the edges greater than the best degree,we randomly remove the destination,otherwise,we added edges from the node's neighbor node.Finally,the experimental results show that the proposed algorithm improves the usability of graph.