On the Sensitivity of Probabilistic Networks to Reliability Characteristics

Diagnostic reasoning in essence amounts to reasoning about an unobservable condition, based on indirect observations from diagnostic tests. Probabilistic networks that are developed for diagnostic reasoning, typically take the reliability characteristics of the tests employed into consideration to avoid misdiagnosis. In this paper, we demonstrate the effects of inaccuracies in these characteristics by means of a sensitivity analysis of a real-life network in the medical domain.

[1]  Silja Renooij,et al.  The Effects of Disregarding Test Characteristics in Probabilistic Networks , 2001, AIME.

[2]  Linda C. van der Gaag,et al.  Properties of Sensitivity Analysis of Bayesian Belief Networks , 2002, Annals of Mathematics and Artificial Intelligence.

[3]  Finn V. Jensen,et al.  Bayesian Networks and Decision Graphs , 2001, Statistics for Engineering and Information Science.