INVESTIGATING THE USE OF BAYESIAN NETWORKS TO PROVIDE DECISION SUPPORT TO MILITARY INTELLIGENCE ANALYSTS

In this paper we consider a typical military scenario where the intention of an enemy force is unknown, but there are a number of plausible hypotheses. As time passes, information in the form of various sightings and reports become available. We employ a Bayesian network, a type of probabilistic graphical model, to process these reports and update the probabilities of the various hypotheses in the light of the latest information. We also demonstrate the beneficial effect of incorporating 'negative' or false evidence on the plausibility of the various hypotheses.