A Bayesian Belief Network-based probabilistic mechanism to determine patient no-show risk categories

•Accurate prediction of patient no-shows is a critical task.•We formulate a probabilistic methodology to determine patient no-show risk scores.•Tree-Augmented Bayesian Belief Network (TAN) model provides superior results.•We build a parsimonious model by employing effective feature selection procedures.•The study uncovers the conditional inter-dependencies between the predictors.•It provides stakeholders with a web-based decision support system tool.