Methods for Constructing Uncertainty Intervals for Queries of Bayesian Nets

Abstract

In this paper the issue of finding uncertainty intervals for queries in a Bayesian Network is reconsidered. The investigation focuses on Bayesian Nets with discrete nodes and finite populations. An earlier asymptotic approach is compared with a simulation-based approach, together with further alternatives, one based on a single sample of the Bayesian Net of a particular finite population size, and another which uses expected population sizes together with exact probabilities. We conclude that a query of a Bayesian Net should be expressed as a probability embedded in an uncertainty interval. Based on an investigation of two Bayesian Net structures, the preferred method is the simulation method. However, both the single sample method and the expected sample size methods may be useful and are simpler to compute. Any method at all is more useful than none, when assessing a Bayesian Net under development, or when drawing conclusions from an 'expert' system.

Rejoinder to ‘Environmental Decision‐making using Bayesian Networks: Creating an Environmental Report Card’

“…The sensitivity assessment then involved modification of the probability distribution in this node. The problem of sampling variation was addressed by constructing uncertainty intervals on the parameters in the network [2]. The assessments revealed that for our case study, although some of the parameters were indeed affected, the overall conclusions remained largely unchanged.…”

To Trade or Not to Trade? Using Bayesian Belief Networks to Assess How to Manage Commercial Wildlife Trade in a Complex World

“…Thus although BBNs can be used to better understand and model complex problems, identify knowledge gaps, prioritize future research and evaluate management options (Johnson et al, 2014), they would need to be complemented by other approaches if the uncertainties in the probabilities need to be captured for decision making. Strategies for estimating these uncertainties have been proposed (Van Allen et al, 2008;Donald and Mengersen, 2014), but they are not currently available in standard BBN software. This current limitation on BBNs does not, however, undercut the value of the general approach of constructing a BBN and exploring it to investigate the relationships between species productivity, different aspects of management and demand, in order to assess impacts of different policy decisions on the sustainability of species' populations.…”