Social relationship hiding method based on attack node similarity

Abstract

The invention provides a social relationship hiding method based on attack node similarity. The method comprises the following steps: 1) dividing network structure data into a training set and a testset; 2) calculating similarity values of all node pairs in the training set according to resource allocation indicators; 3) traversing all target node pairs in a descending order of the similarity values, and if the target node pairs have connected edges in the training set, deleting the connected edges; if the target node pairs have connected edges in the training set, selecting a common neighbornode with the smallest similarity value of the target node pairs, and deleting the connected edge of the node and the target node; if the target node pairs have no connected edges, selecting neighbornodes of a non common neighbor with the smallest median value of the target node pairs, and adding the connection edge of the node and the target node; and if the total number of adding and deletingreaches to the limitation, stopping the execution, and outputting a result. According to the social relationship hiding method provided by the invention, the connected edge disturbance of the networkis performed by using the node similarity to achieve a better hiding effect than random disturbance.

Description

technical field [0001] The invention relates to the fields of network science and link prediction, in particular to a social relationship concealment method based on the similarity of attacking nodes. Background technique [0002] With the development of the mobile Internet and the rise of various social APPs, people are becoming more and more accustomed to interacting with friends in virtual networks. In the era of big data, data contains infinite value. At the same time, users' personal privacy data is also exposed on the Internet, and these sensitive data are not only important to users, but also to enterprises. Enterprises have the responsibility and obligation to protect users' privacy information. The method of the present invention looks at the user's social privacy from the perspective of link prediction. For example, some key links in social networks need to be hidden, but if a third party can predict the hidden links through the network structure, it means that t...

AI Technical Summary

Problems solved by technology

For example, some key links in social networks need to be hidden, but if a third party can predict the hidden links through the network structure, it means that the hiding is invalid, and the user's social privacy is also leaked.

Therefore, the problem can be transformed into the security problem of the link prediction algorithm, that is, how to reduce the accuracy of the link prediction algorithm by adding an appropriate amount of noise links to the social network, thereby reducing the recognition rate of key links and protecting the privacy of users to the greatest extent possible. Safety

Images