Inference approaches to constructing covert social network topologies

Abstract

Social network analysis techniques are being increasingly employed in counter-terrorism and counter-insurgency operations to develop an understanding of the organisation, capabilities and vulnerabilities of adversary groups. However, the covert nature of these groups makes the construction of social network topologies very challenging. An additional constraint is that such constructions often have to be made on a fast time-scale using data that has a limited shelf-life. Consequently, developing effective processes for constructing network representations from incomplete and limited data of variable quality is a topic of much current interest. Here we show how Bayesian inference techniques can be used to construct candidate network topologies and predict missing links in two different analysis scenarios. The techniques are illustrated by application to data from open-source publications. © 2009 Springer Vienna.