A Sliding Window-Based Algorithm for Detecting Leaders from Social Network Action Streams

Influential users or leaders in a social network play important roles in viral marketing by spreading news quickly to a large number of people. Hence, various organizations aim to discover these leaders as campaign targets for advertisement so as to maximize customer reachability. Existing approaches detect leaders from a static social network. However, as social networks are evolving, detecting leaders from dynamic streams of social network data is in demand. In this paper, we propose a sliding window-based leader detection (SWLD) algorithm for discovering leaders from streams of user actions in social networks. Experimental results show that SWLD is accurate, requires short runtime and a small amount of memory space.