Network Structural Feature Preservation Method Based On K-anonymity

As the development of information technology,social networking websites(e.g.Facebook,Twitter,Weibo)are collecting a huge amount of user data with an unprecedented speed.Tremen-dous value underlies in these user data,which is a big treasure for the companies.It will bring more benefits to society if these data are released and shared while more researches take part in the analysis of it.However,this can also lead to serious privacy leaks.Several protection methods for the privacy of network data have been proposed to solve this problem.Most of traditional researches about protection of network data privacy focused on entirely protection of the data,in which all data nodes are protected at a same pre-set protection level.Although these works can solve the problem to some extent,their protection methods lack of per-tinence.In the network data,not all the nodes are facing a same level of risk of being attacked,while attackers choose target based their needs and experience.Therefore,users need to set differ-ent protection levels for different nodes based on the valuable or importance of different part of the data.In response to such user needs,the present paper proposed a social network structural feature protection method based on k-anonymity.Our method consists of three part of algorithms,respec-tively for the protection of three structure features includes degree,center fingerprint and subgraph.Users can choose the target to be protected and the level of the protection.Our approach is capable for reducing the risk of privacy exposure of the target through anonymization of the network.