Identification of an agent interaction network
Kiriakos Kiriakidis · 2010
Knowledge of some particular fact, subject, or event is conveyed among disparate agents affecting their respective activities and, in turn, the generation of new information. The author's main objective is to reconstruct the flow of information in the form of a weighted digraph, based on data that reflect agent activities. Embarking from the principles of maximum entropy, the author casts the digraph identification in the Bayesian estimation framework.