Link Prediction in Microblog Based on Exponential Random Graph Model

• 1、Pattern Recognition and Intelligent System (PRIS) Lab, Beijing University of Posts and Telecommunications, Beijing 100876

Abstract：Link prediction is to predict the possible links between nodes, according to the information existed in the network. Microblog, a new-style social networks developed rapidly, is gradually becoming a tie connecting peoples in the virtual world. For Microblog, link prediction and recommendation between people and people may enlarge the community of social interaction and highlight Microblog's effect in communication. Traditional methods of link prediction are appropriate for all kinds of networks, but ignore the interaction of people in social network. In this paper, the features of social network are fully considered. A sociological model: Exponential Random Graph Model (ERGM) has been introduced in link prediction for Microblog. Synthesizing the data of user info and network topology, a link prediction model has been established, training the model by some statistics of network, simulating the prediction network for link prediction.

Keywords： Link Prediction Microblog Exponential Random Graph Model (ERGM)