N-triad link prediction model based on structural statistics in social network

Almost all social networks provide a mechanism for user to evaluate a subject or another user which they trust or distrust (like or dislike). These link data is an efficacious tool to help system understand user´s interests and other useful information. Therefore, a lot of research on social networks has focused on it. In this paper, we investigate how the positive and negative link reflect the emotion of node and how the basic triad structures influence the link prediction between any two nodes in the social network. And then we develop a model of link prediction based on structural statistics. Finally, discuss our prediction results in Epinions dataset.