Privacy-protected Social Network Data Release For Avoiding Degree Attacks And Attribute Disclosure

With the rapid development of computer technology,more and more social platforms are born,serving people and bringing convenience.People communicate with each other through social networks and understand each other.Social networks provide an attractive environment for people not only to Provide users with convenient services and bring people closer together.However,while social networks bring convenience to users and provide services,there is a need for a large amount of user data support.With the improvement of social platforms,the number of users is increasing,and the user data in the social platform database is also explosive.increase.In today's big data era,more and more data mining algorithms and data mining tools have been developed to apply knowledge discovery in massive data.The same is true for data collected by social platforms.It is used to mine hidden behind data.The rules,knowledge,and more convenient services for users,these rules,knowledge can often help commercial companies to make more reasonable,scientific decisions and predict future trends,or provide research directions for scholars,research social Potential knowledge in the development of the network.But data released social network for government,academic and commercial research companies in the social platform,if not pay attention to data security,attackers with malicious access to the data,is likely to cause social network user privacy leakage.In recent years,with the increasing awareness of the privacy of users,more and more users are reluctant to provide personal data to these social platforms.In this regard,we wish to propose a balance between user privacy and data privacy protection program can research and publish data in the social platform,both to protect the user's privacy is not compromised,while also maintaining a social network can research data.This paper deeply studied in simple social network model,focus only on the user's own identity-based K of anonymous social network privacy protection program,and analyzed the deficiency,also studied in the Properties-social network model,improved K of anonymity the scheme and proposed a privacy protection scheme based on node splitting.main tasks as follows:1.We propose a social network privacy protection scheme based on K-degree anonymity.The program is designed to address the specific location of the attacker to attack social network users by the degree of social network user node,after the success of the series K anonymity,modified by modifying the original network strategy map,making the original network reaches K anonymous,and do It may reduce the loss of information on social networks.In this scenario,we built a simple social network model and analyzed how the attacker attacked the user's specific location in the social network by the user's degree.Then we proposed the unity of the anonymity between the degree sequence and the graph.If the degree sequence satisfies K anonymity,the graph satisfies K anonymity,which leads to the K anonymity mode of the degree sequence,and the graph modification strategy finally makes the graph achieve the effect of K anonymity and ensures the researchability of the data.We validated the experiment using real data sources published by Facebook and compared it with other programs.The experimental results show that the scheme can further reduce the information loss while protecting the user's privacy,and the information loss is lower than the comparison scheme.2.We propose a social network privacy protection scheme based on node splitting.The program aims to deal with the problem of the user's privacy attribute based on the attribute of the user's neighbor set or the user's own attributes under the attribute-social network model.In this scheme,firstly,we demonstrate the existence and rationality of the two attacks through relevant literatures;then,the specific ways of node splitting and the calculation method of the correlation between non-privacy and privacy attributes are proposed.Finally,We propose the overall algorithm of the scheme and introduce some of them in detail.We have adopted Last.FM's open source dataset to systematically experiment with the scheme,and verified the privacy of the scheme through analysis,and verified the scheme by analyzing some evaluation criteria of the social network anonymous algorithm,while protecting the privacy of the user while maintaining the data.Availability.