Network reconstruction of social networks based on the public information.

Abstract

The structure of a social network plays a crucial role for dynamic analysis, which is invisible in most scenes. In this paper, we present a model for reconstructing the social network by taking into account the public opinion diffusion dynamic model for specific agenda. First, the initial polarity attitude of users i for the agenda, oi, is set in the range [-1,1], where negative and positive attitudes are set as -1 and 1, respectively, while 0 means that user i's attitude is uncertain. Second, we present an optimization model for detecting the relationship among each pair of users based on the generated public observable information. The experimental results for four synthetic networks and three real-world social networks show that the reconstruction accuracy depends on the uncertainty of the initial attitudes greatly. This work is helpful for revealing the structure of social networks in terms of public information.