Coritivity-based influence maximization in social networks

Influence maximization problem is about finding a small set of nodes from the social network as seed set so as to maximize the range of information diffusion. In this paper, the theory of coritivity and method of finding core nodes in networks are introduced to deal with this problem. From the perspective of network structure, core nodes are the important ones to network connectivity and is a competitive measurement of node influence. By finding the core of the network through coritivity we can finally get the initial active nodes required in the influence maximization problem. We compare this method with other conventional node-selection approaches in USAir97 and HEPTH datasets. Experimental results demonstrate that: (a) the coritivity-based method achieves large influence spread in all the diffusion models we use, and (b) the proposed method converges fast in all cases we consider.