摘要：Several scientific works have modeled medical problems with assistance of Bayesian networks,
assisting doctors in the task of diagnosing a disease given the observed symptoms and evaluated
exams. This paper aims to present the execution time and convergence analyses for exact and approximate
algorithms for probabilistic inference, which allow to apply the Bayesian reasoning in the support
to the medical diagnosis. The results of the analyses supply a criterion for the choice of the algorithm to
be implemented depending on the resources that are wished to optimize.
关键词：Bayesian networks, inference algorithms, systems for medical diagnosis aid
