April 1, 2023
Conference Paper

Disruption-Robust Community Detection Using Consensus Clustering in Complex Networks


Topological (graph-theoretic) analysis of critical infrastructure networks provides insight on several aspects of resilience. Graph clustering or community detection, which identifies densely connected components in a graph, has been employed for analysis. In this paper, we propose employing consensus clustering, which is a technique to determine consensus from a collection of different clusters on an input, such that the resulting clustering is robust to disruptions, where a disruption is represented as loss of one or more vertices or edges in the graph. Using two critical infrastructure networks as case studies, we empirically demonstrate the need to compute consensus clustering in order to address the drastic changes in the topology due to disruptions in the network.

Published: April 1, 2023


Hussain M., M.H. Khan, A. Azad, S. Chatterjee, R.T. Brigantic, and M. Halappanavar. 2022. Disruption-Robust Community Detection Using Consensus Clustering in Complex Networks. In IEEE International Symposium on Technologies for Homeland Security (HST 2022), November 14-15, 2022, Virtual, Online, 1-6. Piscataway, New Jersey:IEEE. PNNL-SA-174273. doi:10.1109/HST56032.2022.10024983

Research topics