People and entity retrieval in implicit social networks

Online social networks can be viewed as implicit real world networks, that manage to capture a wealth of information about heterogeneous nodes and edges, which are highly interconnected. Such abundant data can be beneficial in finding and retrieving relevant people and entities within these networks. Effective methods of achieving this can be useful in systems ranging from recommender systems to people and entity discovery systems. Our main contribution in this paper is the proposal of a novel localized algorithm that operates on the sub graph of the social graph and retrieves relevant people or entities. We also demonstrate how such an algorithm can be used in large real world social networks and graphs to efficiently retrieve relevant people/entities.