Research On Social Network Personalized Privacy Protection Technology

Now, with the rapid development of network technology, plenty of virtual socialnetworking site quickly popular in various fields of people’s life, obtains populace’saffection and the application, for example: Facebook, Myspace, Twitter, Linkedin and soon is the present socialization network and the web2.0windvane. Since a large number ofthe establishment and use of online social networks, more and more people participate inthe social network. Due to the needs of scientific research and practical applications, thedata owner can carry on the social network data the public issue, which produced a hugeamount of personal data information in social networks, in which some informations is theuser is not willing to disclose, namely in social network privacy. A variety of socialnetwork analysis and mining technologies are used, some social network data is alsopublicized, privacy began to be eroded in a social network. Therefore, the social networkprivacy protection is a priority.The existing social network privacy protection algorithms are usually based on allusers had the same privacy protection needs, ignored that different users had differentpreferences and he attacker has the diversity of background knowledge. In order to solvethis problem, personal privacy protection framework has been put forward, according todifferent user privacy protection needs to extract some subset, set up three kinds of privacyprotection level: First of all, we should remove the node label simply; Second, in order toprotect information of the node degree, we put forword k-d_sub (k-degree_subset)algorithm based on dynamic programming ideas, the algorithm ensures that there are atleast k-1of the other nodes which their degree are equals to this sub-set of each node’degree; Finally, in order to prevent the sensitive attributes which are identified, combinedthe thought of l-diversity with the k-d_sub algorithm, we proposed k-d_l_sub(k-degree_l_subset) algorithms, the algorithm ensures that each node in the sub-set whichmet k-degree anonymous and ensure that there are at least l nodes’ sensitive attributestypes in the same anonymous group. At last, Based on the above personal privacy protection framework, Focusing on the issue that the current related research about socialnetwork do not consider subsets of neighborhood’s privacy preserving, and the specificproperties of subsets of neighborhood also lead individual privacy disclosure, a new (θ, k)-anonymous model was proposed. According to the k-isomorphism ideology, the modelremoved labels of subsets of neighborhood which needed to be protected in social network,made use of neighborhood component coding technique and the method of node refiningprocessed node candidate set and its neighborhood information, then completed theoperation of specific subsets isomorphism, which considered the issues of the sensitiveattribute distribution. Ultimately, the model satisfied that each node in neighborhoodsubset met neighborhood isomorphism with at least k-1nodes, as well the model requiredthe difference between the attribute distribution of each node in the neighborhood subsetand throughout the subsets is not bigger than θ.This article tests algorithm performance through the massive different experiments,the theoretical analysis and the experimental testing the proved, the experiment proved, themehods of proposed new personalized framework and the (θ, k)-anonymous model hasbeen realized through the increase least quantity side, reduces the anonymization cost andmaximize the utility of data, has the high anonymous quality, can protect in effectively thesocial network user’s privacy.