Node role analysis algorithm based on directed topological potential

The majority of real-world networks are directed, weighted and dynamic. Aiming at the problem of node role analysis in directed weighted network, in this paper, we introduce topological potential to network analysis and propose a novel node role analysis method based on directed topological potential, which can divide nodes into four kinds of roles based on their behavior pattern and local influence. By utilizing the proposed node role analysis method, the experiment on a typical dataset is done, and the result shows that our approach can quickly evaluate the node role and further analysis, such as temporal networks evolution, could be done effectively.