Research On Differential Privacy Protection Algorithms For Social Networks

In the Internet era,a lot of data in social networks are users' private information.In the process of data mining and analysis of social networks,these private information is easy to be leaked,causing unnecessary losses to users.Therefore,it is necessary to study the privacy protection of social networks.Considering the social network with static and dynamic properties,social network privacy protection,this thesis studies main work is as follows:1.Aiming at the privacy protection problem of static social network,a differential privacy protection algorithm for static social network is designed.Firstly,in view of the problem that the traditional differential privacy protection algorithm directly adds noise to the social network,the social network is grouped into different clusters by spectral clustering algorithm,and the noise is added randomly to different clusters to reduce the amount of noise and improve the utility of data.Secondly,in order to achieve a more balanced differential privacy protection,a new privacy budget parameter is designed,and the noise addition amount is determined according to the edge weight of the social network to achieve a more balanced privacy protection and further improve the data utility.Finally,the simulation results show that the proposed algorithm can effectively improve the utility of data under the premise of satisfying the differential privacy protection.2.Aiming at the problem of privacy protection in dynamic social networks,a differential privacy protection algorithm for dynamic social networks is designed.Firstly,a dynamic community discovery algorithm is proposed to solve the problem of dynamic social network data preprocessing.Through dynamic community discovery of the dynamic social network,the algorithm adapts to the changes of the dynamic social network,efficiently preprocesses the data and effectively improves the efficiency of the algorithm.Secondly,a differential privacy protection algorithm is proposed to solve the differential privacy protection problem of dynamic social networks.The algorithm to perform initialization difference social network privacy first,and then the results of dynamic community found in similarity calculation,through similarity to determine whether a community change,change for the community to add a new noise disturbance,and the change of the community the previous moment disturbance unchanged,thereby reducing network dynamic added noise in the iterative process,Further enhance the utility of data;Then,the social network graph after differential privacy protection is generated by the community connection algorithm.Finally,simulation experiments show that the proposed algorithm can effectively improve the efficiency of differential privacy protection algorithm for dynamic social networks,while retaining higher data utility.