A clustering based-approach to analyze centrality of terrorist networks using games with a priori unions

Algaba, Encarnación, Saavedra-Nieves, Alejandro, Saavedra-Nieves, Paula
OR Spectrum · Springer Nature · 2026-11-19 · DOI: 10.1007/s00291-026-00878-7

Аннотация

Intelligence services need to identify the key players in terrorist networks to effectively allocate the surveillance resources according to the risk analysis in the network. Unlike of classical transferable utility games, transferable utility games with a priori unions allow for ranking terrorists in a network incorporating the affinity information through a partition of their members. However, the selection of this (in principle, non-unique) partition is not usually trivial in practice. In this work, we propose a methodological solution to accomplish this task from well-known clustering-graph techniques. Apart from briefly analyzing classical graph clustering methods, a novel and different graph clustering algorithm is introduced to capture the most natural and relevant features of these networks such as communication and hierarchical properties, which play a significant role in its performance. Moreover, with the purpose of describing the influence of the partitions determined from different clustering algorithms, including our new proposal, we compute the Banzhaf–Owen value to provide and compare the final rankings obtained for the Zerkani network. In contrast to the other clustering methods analyzed, with our new proposal, Khalid Zerkani is ranked as the first in the top-ten of the ranking.

Читать у издателя →

Похожие статьи