Social Network Footprints: Identifying the roles of nodes and edges in social-network graphs from topology information

Social network graphs contain a great deal of useful information that can be used for a wide range of applications including commerce, dissipation of critical information, influencing political outcomes, etc. In this talk, we will explore different approaches for analyzing the topology of social networks to identify the "importance" or "influence" of nodes on other nodes within the network graph. The talk will review traditional techniques based on different network-graph centrality measures, and in particular, eigenvector based graph centrality. Shortcomings of traditional techniques will be highlighted to motivate the need for a more general framework for analyzing social network graphs. This will lead to the discussion of a new spectral-analysis framework for identifying influential nodes and information hubs in massive social network graphs. Other network-graph models for identifying different roles of nodes, such as leaders and followers, will also be discussed. Few ongoing research problems that are being explored will also be highlighted.