Forecasting low-probability high-risk accidents
We use influence diagrams to describe event trees used in safety analyses of low-probability high-risk incidents. This paper shows how the branch parameters used in the event tree models can be updated by a bayesian method based on the observed counts o