ABSTRACT

Causal models are powerful tools for modeling complex cause-effect relationships among variables. In this context, Bayesian networks (Pearl (1988)) have emerged as one of the most useful formalisms for performing, for instance, qualitative and quantitative risk analysis (Firmino et al. (2006) and Droguett et al. (2006)), military vehicles performance measurement (Neil et al. (2001)), dynamic systems modeling (Boudali & Dugan (2005) and Bouissou & Pourret (2003)), automatic target identifications, image recognition, adaptive signal processing and digital communications (Heckerman et al. (1995)). Such widely application is due to the fact that BNs are useful for performing diagnostic and prognostic analysis in the domain of uncertainty.