Exploring influential nodes using multi-attribute information

In recent years there has been greatly increased interests in finding influential nodes in social networks. For a long time, PageRank has been widely used and has also been adopted as a solid baseline in the field of influence maximization. However, there are all kinds of interactions among social entities which contributes to varieties of relationships between nodes in information diffusion graphs of social networks. We refer to these relationships as multi attributed relationships and examples include follow, comment in Facebook, coauthor, citation in collaboration networks and so on. PageRank and its variants use just one kind of relationship and haven´t taken multi relationships and the differences between relationships into consideration. This results in a crude simplification of the real world and leads to poor performance. To address this problem, in this paper, we construct a multi-attribute information diffusion graph(MAID graph) by integrating different relationships. A variant of PageRank referred as multi-attribute Rank(MA-Rank) is proposed. The experimental results show that our proposed method achieves higher accuracy and efficiency in finding influential nodes comparing with PageRank and LeaderRank.