ProfitLeader: identifying leaders in networks with profit capacity

摘要

Identifying Leaders' in a network is a significant step to optimize the use of available resources, accelerate or hinder spreading information. In this paper, we propose a new measure to characterize the importance of a node, called ProfitLeader, which considers the problem of critical node identification from a novel perspective: profit capacity. Unlike established approaches to identify influential nodes in networks based on its local topology or information diffusion capacity, ProfitLeader, ranks the influence of each individual node via its profit capacity. The basic idea is to view each node as a person, and measure how much profit the person can offer for others. Namely, the more profit the person provides for others, the more important the node is. To characterize the profit capacity of each node, its available resource and sharing probability to its neighbors are investigated. Building upon the profit capacity, ProfitLeader has several remarkable advantages: (a) It provides a simple, intuitive yet effective way to rank the nodes in networks from a new viewpoint; (b) ProfitLeader allows finding the most important nodes (with high quality). (c) ProfitLeader is fully automatic, and no parameters are required to be specified by users. Extensive experiments on both small networks and large-scale networks have demonstrated its superiority over several state-of-the-art algorithms.