• A dynamic risk assessment method involving system dynamics simulation and Bayesian network is proposed • The system dynamics model enables dynamic probability prediction under the coupling mechanism of faults • A Bayesian network-based consequence model is established to assess the effect of safety barriers • A case study demonstrates the effectiveness of the proposed methodology for methanol bunkering under SIMOPs • The societal risks associated with methanol leakage are assessed across varying population densities This paper proposes a dynamic risk assessment method based on system dynamics (SD) simulation to deal with both the complexity of systems involving simultaneous operations (SIMOPs) and their dynamic evolution over time. A fault tree model is constructed and transformed into an SD model comprising four feedback loops, enabling dynamic probability prediction under coupling mechanisms while accounting for each loss of containment (LOC). Critical hazard factors are identified and ranked based on mutual information (MI), allowing prioritization of safety interventions. Consequence-probability modelling is carried out by establishing a Bayesian network (BN)-based event tree considering safety barriers (SBs). The severity of potential consequences is evaluated using fire modelling in ALOHA. Individual and societal risk values are subsequently calculated, followed by the development of risk mitigation measures at a bunkering station in the Port of Shanghai, China, as a practical case study. The proposed mitigation strategies effectively reduce leakage probability and associated risks. The proposed framework effectively characterizes the temporal evolution of risk, provides enhanced representation of time-dependent risk dynamics compared with traditional static quantitative risk assessment models, and offers a transferable basis for alternative marine fuels and port operations. To the best of our knowledge, the combination of bow-tie analysis, dynamic simulation, and Bayesian networks for temporal methanol bunkering risk assessment remains unexplored in the current academic literature.

Last-layer committee machines for uncertainty estimations of benthic imagery.

Constructing Bayesian networks from knowledge graphs for risk assessment and causal inference of urban rail transit equipment.

Data-driven prediction of oil spill risk in ship-to-ship bunkering by reinforcement learning and Bayesian belief networks.

• Proposes an integrated ISM-PIS-RABN framework for full-cycle risk assessment of multimodal transportation systems. • Introduces personalized individual semantics to quantify expert heterogeneous, mitigating group decision bias. • Employs rule-augmented Bayesian networks with factor decomposition to overcome the curse of dimensionality. • Validated via a real-world corridor case, providing actionable insights from sensitivity analysis. Multimodal transportation plays a pivotal role in international trade and logistics due to its unparalleled advantages in cost, efficiency, sustainability, and accessibility. However, its complex structure, which is characterized by multiple transfer points, long transit times, and extended routes, makes it more prone to risks than unimodal transport. Conducting scientific risk assessment and identifying critical factors are essential for improving safety. Therefore, this study proposes a risk assessment model integrating Interpretive Structural Modeling (ISM), Personalized Individual Semantics (PIS), and a Rule-Augmented Bayesian Network (RABN), forming a systematic framework covering risk factors identification, correlation analysis, and comprehensive evaluation. Potential risk factors affecting multimodal transport safety are first identified from a 4M1E (Man, Machine, Material, Management, Environment) perspective. ISM is then used to classify these factors hierarchically and construct the Bayesian network structure. By combining PIS, similarity measures, and Inverse Distance Weighting (IDW), the model processes expert judgments to determine root node probabilities. Finally, RABN performs probabilistic inference for risk assessment. The proposed framework is validated through a case study on a route in the New International Land-Sea Trade Corridor. Sensitivity analysis further identifies key risk factors and verifies parameter robustness, offering decision-making support for multimodal transportation risk management.

A novel risk assessment framework for piracy threats: mapping system theoretic process analysis and complex network into Bayesian networks

Multimodal graph fusion-based GCN for Alzheimer's disease diagnosis using fMRI and T1-weighted MRI.

Data from psychophysiological measures can offer new insight into control room operators’ behaviour, cognition, and mental workload status. This is particularly helpful when assessing capacity to respond to critical plant conditions such as alarm response scenarios. However, wearable tools such as eye tracking and electroencephalography caps can be perceived as intrusive and unsuitable for daily operations. Therefore, this article examines the potential of using real-time data from process and operator-system interactions during abnormal scenarios, recorded and retrieved from the distributed control system’s historian or process log, to provide insight into operator behaviour and predict their response outcomes without intruding on daily tasks. Data for this study were obtained from a design of experiment using a formaldehyde production plant simulator and four human-in-the-loop support configurations. A comparison between configurations in terms of both behaviour and performance is presented. Then, a step-wise logistic regression and a Bayesian network model were used to predict operator error. The results identified predictive metrics, discussed in terms of their value as precursors of overall system performance in alarm response scenarios. Knowledge of relevant and predictive behavioural metrics accessible in real time can better equip decision-makers to predict outcomes and provide timely support measures for operators.

Anxiety, depression, and post-traumatic stress disorder among Palestinian refugees in Egypt: Gender-stratified item-level Bayesian network analysis.

Resilience performance analysis of interdependent infrastructures using Fault Tree and Dynamic Bayesian Network

Probabilistic assessment of heavy metal risks in sewage sludge using BCR speciation and Bayesian networks.

Ensemble-based uncertainty quantification and decomposition of probabilistic surrogate models using Bayesian neural networks

The issue of employee absenteeism presents a continual obstacle to the effectiveness of organizations, especially within sectors that rely heavily on labor. This research introduces an innovative probabilistic framework that employs Bayesian Belief Networks (BBNs) to forecast the risk of absenteeism as well as the absenteeism rate, tackling the intricate and uncertain relationships among various contributing factors. By employing thorough factor analysis and drawing on expert insights, we systematically identify and organize essential variables spanning personal, organizational, familial, health-related, and external domains. The model integrates the insights of experts and addresses the challenge of limited data by employing a functional interpolation technique to develop conditional probability tables. Two models based on Bayesian Belief Networks have been developed: one assesses the probability of an employee being absent on a specific day, while the other predicts absenteeism rates on a monthly basis. The Monte Carlo simulation serves as a valuable tool for addressing the uncertainty inherent in essential input variables. The validation process, which includes scenario, sensitivity, and diagnostic analyses, highlights the strength and relevance of the models. This study presents a practical and scalable resource for decision-makers to evaluate and address absenteeism, particularly in contexts where data may be limited. • Develops two models to predict absenteeism risk and absenteeism rate. •Applies Bayesian Belief Networks to model uncertainty. •Introduces a structured expert elicitation approach. •Enables customization using factor analysis. •Validated through real-world manufacturing case study.

Understanding how early mother-child interactions are linked to children's social-cognitive processes requires methods capable of capturing the temporal structure of naturalistic behavior. This study introduces a computational framework based on Bayesian Network modeling to identify sequential dependencies among nonverbal behaviors (smiles, gaze, and social touch) exchanged during free play in mother-child dyads (n = 38; age 3 years). From each network, we derived the Order of Sequential Interaction (OSI), a compact index of interaction complexity. We then examined its associations with behavioral, physiological, and neural measures relevant to cognitive development. Although OSI was not associated with language or executive-function scores, analyses revealed links between OSI and prosocial behavior, facial EMG, and neural responses (rTPJ, lIFG) during prosocial-scene viewing. These findings suggest that OSI may capture aspects of interaction structure specifically connected to children's social and affective responsiveness. Building on this, the present framework demonstrates how probabilistic graphical models can structure complex interaction data and support future investigations into multimodal processes in early socialcognition. SUMMARY: A Bayesian-network framework is proposed to model multivariate sequential dependencies in naturalistic mother-child interaction. The order of sequential interaction (OSI) quantifies interaction complexity from behavioral time-series data. Higher OSI is associated with greater prosocial behavior and with neural (rTPJ, lIFG) and physiological (facial EMG) responses during social processing. Interaction complexity is not associated with general cognitive or language measures, suggesting that it reflects a distinct dimension of social behavior. The proposed framework provides a basis for studying social-cognitive development from naturalistic interaction data.

Utilizing the Bayesian network algorithm for noninvasive prenatal paternity testing (NIPPT) and efficacy evaluation of NIPPT system.

Exercise training is an effective non-pharmacological strategy for reducing blood pressure; however, current guideline recommendations are largely derived from studies reporting resting clinic blood pressure rather than ambulatory blood pressure (ABP), which provides greater prognostic value. Comparative evidence on the effects of different exercise modalities on ABP remains limited. To compare the effects of major exercise modalities on 24-h systolic ABP using a Bayesian network meta-analysis, and to outline practical implications for antihypertensive exercise prescription. PubMed (MEDLINE) and the Cochrane Library were searched for randomized controlled trials published before January 2025 reporting changes in 24-h systolic ABP following an exercise intervention of at least 2 weeks' duration. Eligible studies included adults and compared aerobic exercise training (AET), resistance training (RT), high-intensity interval training (HIIT), combined training (CT), or isometric exercise training (IET), with a non-exercise control or another exercise modality. Bayesian network meta-analyses were conducted, with interventions ranked using surface under the cumulative ranking curve (SUCRA) values. Twenty-five studies (1096 participants) met inclusion criteria; 16 were included in the network meta-analysis. AET and HIIT reduced 24-h systolic ABP (-4.77 mmHg and -6.86 mmHg), with a smaller reduction following RT (-2.25 mmHg). Data were insufficient to analyse comparative effects for CT or IET. In the Bayesian network, rank order of effectiveness based on SUCRA values was HIIT (91.4%), AET (68.2%), and RT (35.4%). No statistically significant comparative differences were observed between exercise modalities. This Bayesian NMA demonstrates HIIT and AET as effective for reducing 24-h systolic ABP, supporting their role in hypertension management. However, the limited scale and connectivity of available trials restrict definitive comparisons between wider exercise modes. Larger, well-designed head-to-head trials incorporating ABP outcomes are needed to clarify the relative effectiveness of different exercise strategies and to better inform evidence-based exercise prescription for blood pressure management. Registration:PROSPERO: CRD420251177743.

This study aims to assess the adverse event profile of apremilast using FDA Adverse Event Reporting System (FAERS) data to identify potential safety risks and support clinical use. FAERS data from Q1 2014 to Q1 2024 were analysed. Adverse drug events (ADEs) related to apremilast were extracted and evaluated using four signal detection methods: Reporting Odds Ratio (ROR), Proportional Reporting Ratio (PRR), Bayesian Confidence Propagation Neural Network (BCPNN), and Empirical Bayesian Geometric Mean (EBGM). A total of 122 287 apremilast-related AE reports, encompassing 238 215 adverse reactions across 24 System Organ Classes (SOCs) and 60 apremilast-induced AE signals, were identified. Gastrointestinal disorders, skin and subcutaneous tissue disorders, and musculoskeletal and connective tissue disorders were the most frequently reported SOCs. Common Preferred Terms (PTs) included diarrhoea, nausea, and headache. Notably, some adverse effects not listed in the drug package insert were found, such as multiple allergic reactions and tumour signals. This study identified significant disproportionality signals for ADEs reported with apremilast, particularly concerning gastrointestinal reactions, psychiatric disorders, and infections. While these findings do not establish causality, they offer valuable real-world insights that underscore the importance of continued clinical monitoring and warrant further pharmacoepidemiologic investigation.

Hierarchical Bayesian physics-informed neural networks for fatigue crack growth prediction under multiaxial loading condition

Antibiotic safety signals in Saudi Arabia: A pharmacovigilance study using disproportionality analysis.

Incorporating uncertainty quantification into deep-learning-based travel mode choice modeling: A Bayesian Neural Network approach and an uncertainty-guided active survey framework