MAXIMAL CONSISTENCY, THEORY OF EVIDENCE, AND BAYESIAN CONDITIONING IN THE INVESTIGATIVE DOMAIN
Aldo Franco Dragoni, SAMUELE ANIMALI · Cybernetics & Systems · 2003
We present the architecture of an Inquiry Support System whose aim is to help a detective or a judge in eliciting a maximally consistent set of beliefs as the most believable piece of knowledge to reason with. This is done by (1) finding incoherence inside and across the various depositions; (2) generating the alternate maximally consistent set of beliefs; (3) estimating the degree of reliability of the various informants; and (4) estimating the degree of credibility of the various evidences. The solution of the case is searched among the various possible plots compatible with the maximally consistent set of beliefs retained by the system as the most believable one.