Research On Privacy Security Of Social Network Community Discovery Algorithm

With the continuous progress of human network technology and the change of social needs,the number of users of social platforms is increasing.In order to better carry out data mining to achieve accurate recommendation,the community detection algorithm is introduced to divide the social network community,so as to understand the preferences and interests of the user group more quickly and comprehensively.As a community division algorithm,community detection algorithm can divide the nodes with high tightness in large complex networks into a group.Moreover,the community structure of "people divided by groups" in social networks is objective,and its in-depth research has important value in many fields.However,in recent years,with the development and popularization of various social platforms,personal privacy and security issues have become increasingly prominent.In order to solve the problem of leakage of personal privacy information of users in the process of community detection,this paper introduces the localized differential privacy technology and the idea of blockchain to preprocess the data,so as to improve the privacy security of community division.The main work completed is as follows:1)The initial community detection algorithm is to directly put user data into the algorithm for analysis,ignoring the security of user sensitive information.Moreover,early privacy protection models need to assume possible attack models,and must require continuous improvement of their own models.The differential privacy protection model has strong background knowledge to make up for the shortcomings.In addition,it does not need to consider the size of the data set when adding noise,so it is very suitable for privacy protection in massive data analysis.In addition,in order to achieve accurate division more quickly,the process of central sequence is added to the process of community detection.In summary,this paper proposes a social network community detection algorithm based on differential privacy technology.The algorithm reduces the number of iterations and time complexity by adding a central sequence to the community discovery algorithm,and realizes privacy protection while ensuring the accuracy of social network community division by introducing differential privacy technology.2)Aiming at the security problem of social network data storage,a community discovery data storage mechanism based on blockchain is designed based on the decentralized characteristics of blockchain.The community discovery data storage mechanism proposed in this paper uses Ethereum as the underlying blockchain,and stores the encrypted sensitive key information of the community nodes in the system in Ethereum,which ensures the credibility and immutability of the data.In this mechanism,a list of trust relationships is made for each data owner.When a node requests,it is judged whether the node is in the trust relationship,if it is returning data directly.Otherwise,authentication needs to be submitted.In addition,a PBFT consensus algorithm based on multi-layer grouping is proposed,which classifies the social network user nodes according to node attributes and reaches a consensus hierarchically.The algorithm improves the consensus efficiency and reduces the communication cost and consensus delay in the social network with many nodes and a large amount of data.3)Combined with the research results of privacy and security of community division algorithm,a privacy and security research system of community detection based on Django framework is designed.The system combines Django framework and Bootstrap front-end framework,which mainly includes community nodes,community discovery,login and registration modules,and security modules.It aims to protect the security of users personal information,and help relevant researchers efficiently mine valuable information in social networks,so as to achieve more accurate recommendation.