par Parsons, Simon;Saffiotti, Alessandro
Référence International journal of approximate reasoning, 14, 2-3 SPEC. ISS., page (187-216)
Publication Publié, 1996-02
Article révisé par les pairs
Résumé : Quantitative methods for reasoning under uncertainty have become well established, and many alternative formalisms have been suggested. In recent years there has been a growing interest in qualitative methods as helpful in situations in which the use of precise numerical methods is not appropriate. In this paper we demonstrate another use for qualitative models. The qualitative analysis of a quantitative model of uncertainty will reveal the qualitative behavior of that model when new evidence is obtained. This qualitative behavior may be studied to identify those situations in which the model does not behave as expected, and which quantitative values must be altered to correct this behavior. The demonstration is set within the context of the diagnosis of faults in an electricity network, and reports the results of the verification of a model representing a small fragment of a real application. The model was built using Pulcinella, a tool based on Shenoy and Shafer's valuation systems.