Sammendrag:
Fault and event trees have for several decades been the most commonly used modeling tools in risk analysis for structuring and modeling large complex systems. In this presentation I will give an introduction to Bayesian networks and I will explain why I find that Bayesian networks represent a superior modeling tool to both fault and event trees when structuring and modeling large complex systems. Note that any fault or event tree relatively easy may be converted into a Bayesian Network, whereas the opposite rarely is the case.
Some of the primary issues that must be kept in mind, when modeling large complex systems, are transparency, uniformity in modeling complexity, and verifiability of probabilistic modeling. Bayesian networks helps to assure that these issues can be satisfactorily handled.
A Bayesian network is a graphical model or representation of uncertain quantities (and decisions) that explicitly reveals the probabilistic dependence between the set of variables and the flow of information in the model. A Bayesian network is designed as a knowledge representation of the considered problem and may therefore be considered as the proper vehicle to bridge the gap between formulation and analysis.
In this seminar I will give an introduction to the theory behind Bayesian network and illustrate the application of Bayesian network to a few selected risk analysis problems. Further, because of the transparency of Bayesian network, Bayesian statistical analysis becomes much easier to understand and follow. I will illustrate this through a few Bayesian statistical analysis examples.