Component importance and sensitivity analysis in Bayesian networks

In this paper, component importance and sensitivity analysis in Bayesian networks are reviewed and a strategy for the case of non-deterministic structures is proposed. A complex system example is analyzed by the methods of Birnbaum´s derivative, variance-based sensitivity analysis in deterministic structures and non-deterministic structures respectively. The results are compared and demonstrate the feasibility of the proposed strategy for non-deterministic structures in Bayesian networks.