A Random Method Of Society Network Based On Certain Constraints

At present, there are lots of privacy information about individuals or collective in the social network, such as customer shopping social network, friends network, telephone network academic cooperation network, and various groups network. With the develement of computer technology and network communication technology, and the increasement of social network sites, more and more people participate into actions of the social network. The relationship of these social networks could be used for trend analysis, market forecast and disease forecast.These social networks contain individuals or collective privacy information, therefore, when we comunicate with each other and share some information, some privacy information inside them may be leaked. Therefore, privacy protection problem is a very importmant issue within the work of publishing in the social networks.Due to its own safty and effectiveness, random anonymous methods has been a research focus spot of social network anonymous field recently. Random anonymous methods by deleting relationship in the social network or adding relationship which is not exist in the original social network may prevent the attacker exactly identifying privacy information or sensitive information of individuals or collective. In this way, they could protect the privacy information of individuals or collective. In this paper, privacy protection problem of the social network are discussed with two aspects of random anonymous method and algorithm. The main works of this paper contains:There is a method with only four points perturbation for the randomization society network which based on spectrum constraint, and the privacy protection degree of it is not well enough after perturbation to the social network. To solve the problem, this paper proposes a method with multi-point disturbance to randomize the society network. In the randomization process, this method is based on the adjacency matrix and signless Laplace matrix. It controls the social network spectral radius in certain within constraints to ensure the usability and improve the privacy protection degree of the social network. The experimental results of disturbance algorithm prove that this method can better protect the social network privacy.To beeter protect the security of social network and to ensure the availability of social network after perturbation, this paper introduces non-randomness of the social network and then proposes a method of multi-point disturbance to the society network base on certain constraints. Because non-randomness of the social network could reflect its structure more than the other properties, so it is regarded as one of the constraint conditions in our method and could effectively protects the structural features of social network. This paper also analyzes the security of this method. At last, different random methods are compared each other in three aspects of harmonic mean of the shortest distance of the social network, subgraph centrality and the social network non-randomness. From the results of above experiments, the proposed method in this paper effectively protects the social network privacy and improves the availability of the social network.