Social Capital: The Power of Influencers in Networks (Extended Abstract)

12th International Conference on Autonomous Agents and Multiagent Systems

Karthik Subbian ; Dhruv Sharma ; Zhen Wen ; Jaideep Srivastava

PID  :  123885

来源: CEUR

【 摘 要 】

The problem of finding the influencers in social networks has been traditionally dealt in an optimization setting by finding the topk nodes that has the maximum information spread in the network. These methods aim to find the influencers in a network through the process of information diffusion. However, none of these ap proaches model the individual social value generated by collabora tions in these networks. Such social value is often the real moti vation for which the nodes connect to each other. In this work, we propose a framework to compute this network social value using the concept of social capital, namely the amount of bonding and bridging connections in the network. We first compute the social capital value of the network and then allocate this network value to the nodes of the network. We establish the fairness of our allocation using several axioms of fairness. Our experiments on the real data sets show that the computed social capital is an excellent proxy for finding influencers and our approach outperforms several popular