Guide query in social networks

Abstract

In this paper, we study a new problem of social guide query, which is to find the most informative neighboring nodes of the query node considering the given requirements. After the target nodes are identified, instead of directly providing information of the target nodes, we would like to find the querier's friends who are targets or who can introduce the querier to some targets. To answer the social guide query, we define InfScore based on an existing model of influence diffusion. In addition, we define DivScore that can be integrated in the ranking process to appreciate the diversity of possibly accessible target nodes. To evaluate the guide query, we develop a prototype system using the coauthor network of DBLP computer science bibliography. The results show that the guide query is practical, which can provide valuable answers efficiently. © 2013 Springer-Verlag Berlin Heidelberg.