Ranking method for mediators in social network

Abstract

This paper proposes a method for ranking mediators where a mediator is defined as node having an important role in a social network. To precisely rank the mediators in order of their importance, a method is used based on changes in the average shortest path length. However, the computational complexity for this method is O(N5), so an unreasonable amount of time it is required to determine complexity for a massive network. Our ranking method, whose complexity is no more than O(N2), is based on the relationships among adjacency nodes. Although the method does not provide a precise but an approximate rank, we found that there is a strong correlation between the ranks generated using the strict and the developed methods. Results on a variety of generated networks confirmed the feasibility of our method for a massive network. © Springer-Verlag Berlin Heidelberg 2007.