Common quandaries and their practical solutions in Bayesian network modeling

摘要: Abstract Use and popularity of Bayesian network (BN) modeling has greatly expanded in recent years, but many common problems remain. Here, I summarize key BN model construction interpretation, along with suggested practical solutions. Problems include parameterizing probability values, variable definition, complex structures, latent confounding variables, outlier expert judgments, correlation, peer review, tests calibration validation, overfitting, wicked problems. interpretation objective creep, misconstruing influence, conflating correlation causation, proportion expectation probability, using opinion. Solutions are offered for each problem researchers urged to innovate share further