Exploring the intervention problem with the networked poisson process in a real heterogeneous social network

Abstract

We model the microblog as a heterogeneous network with individuals in different roles, and introduce "intervention" to describe the phenomenon that some tiny variations about a small subset of the individuals change the whole network's status. Our main contributions are: (1) proposing the Networked Poisson Process (NPP) to model the dynamic tweeting patterns for the microblog; (2) formalizing a NP-hard problem: the intervention impacts maximization (IIM); (3) proposing heuristic algorithms to solve the IIM; (4) providing sufficient experiments to test our methods. The experimental results show that NPP captured the real interaction patterns for the users in the microblog. © 2014 Springer International Publishing Switzerland.