Finding Top-k Central Nodes in a Diffusion Network Using Various Methods

Seojun Yang
Concord Academy

Abstract

We created a synthetic undirected graph of disease diffusion network that expresses the disease infectee as a node and their relation to other infectees as an edge. To figure out the infectee who is influential the most in spreading the disease, we used various methods to compare each infectee›s influence across the network: degree centrality, betweenness centrality, closeness centrality, eigenvector centrality, PageRank, and Katz centrality. After calculating each infectee›s centralities and PageRank in the diffusion network, we concluded that betweenness centrality is the ideal method for diffusion network since the similarity between the infectees with high betweenness centrality and the infectees whose substantial influence is intuitively noticed is high. Also, we discussed future work to get the most central nodes in a graph more accurately.

Please click to read the paper.