Entropy based flow transfer for influence dissemination in networks

This paper advances the use of entropy based centrality measure to locate authoritative spreaders in complex networks. The random Markov chain epidemiological aspect signifies to estimate nodes spreading power by summarizing the effect of influence transfer originated from the node in a completely susceptible network. The proposed method exploits entropy of influence path transfer from a node to its local network to determine its expected local scope of spread. It further facilitates community feature of neighboring nodes to a central node to appraise spreading outlook in global range. We exploit entropy estimation of spread and community feature of nodes to define a novel centrality measure for influence dissemination. We consider SI and SIR model for validating the performance of proposed method. The empirical experiments are implemented on real networks and results establish that influence disseminator detected by proposed method are dominantly more central than various benchmarks.