Bayesian Belief Network Model Research Articles

A reproductive resilience index for pelagic fish in the southern Humboldt Current Large Marine Ecosystem

Efforts to manage small and medium-sized pelagic fishes (SMPF) using traditional stock assessment methods are hampered by the elusive relationship between spawning stock biomass and recruitment. We propose to compute a reproductive resilience index (RRI) in three steps: ( i) we selected 16 biological traits related to distinct aspects of SMPF reproductive biology and characterized by a continuous (e.g., growth rate) or ordinal variable (e.g., spawning site fidelity) scored from 1 to 5 (1 representing a low contribution of the trait to reproductive resilience); ( ii) an expert panel assigned the traits’ scores to five exploited species in the southern Humboldt Current Large Marine Ecosystem; ( iii) a Bayesian Belief Network (BBN) model was used to estimate an RRI based on the combination of the traits’ scores. The BBN was used to explore environmental effects on the species’ RRI, as some reproductive traits can show intraspecies variability under external forces (e.g., fishing pressure). Through proving the RRI application to detect variability in species’ resilience at local time series, we show how the resulting RRI can be interpreted by fishery managers to improve the current management of SMPF.

Towards Sustainable Road Safety in Saudi Arabia: Exploring Traffic Accident Causes Associated with Driving Behavior Using a Bayesian Belief Network

Understanding the causes and effects of road accidents is critical for developing road and action plans in a country. The causation hypothesis elucidates how accidents occur and may be applied to accident analysis to more precisely anticipate, prevent, and manage road safety programs. Driving behavior is a critical factor to consider when determining the causes of traffic accidents. Inappropriate driving behaviors are a set of acts taken on the roadway that can result in aberrant conditions that may result in road accidents. In this study, using Al-Ahsa city in Saudi Arabia’s Eastern Province as a case study, a Bayesian belief network (BBN) model was established by incorporating an expectation–maximization algorithm. The model examines the relationships between indicator variables with a special focus on driving behavior to measure the uncertainty associated with accident outcomes. The BBN was devised to analyze intentional and unintentional driving behaviors that cause different types of accidents and accident severities. The results showed when considering speeding alone, there is a 26% likelihood that collision will occur; this is a 63% increase over the initial estimate. When brake failure was considered in addition to speeding, the likelihood of a collision jumps from 26% to 33%, more than doubling the chance of a collision when compared to the initial value. These findings demonstrated that the BBN model was capable of efficiently investigating the complex linkages between driver behavior and the accident causes that are inherent in road accidents.

Behavioral and socio-economic factors controlling irrigation adoption in Maharashtra, India

ABSTRACT Psychological frameworks are rarely used to understand irrigation adoption behaviour in developing countries. A Bayesian belief network (BBN) model was developed that integrated socio-economic characteristics and psychological factors to understand farmer behaviours with respect to irrigation practices in four districts of Maharashtra, India. Strong norms, risk perceptions of water scarcity, and attitude play roles in the adoption of irrigation technology and practices. Critically, it was found that no one factor can explain adoption behaviour; rather, an ensemble of factors is needed to understand farmer behaviour. A farmer who is highly educated, middle-aged, and moderately wealthy with a significant level of family help and an open well as their main water source, while receiving low promotional information related to water scarcity and irrigation adoption, is most likely to adopt irrigation technology. The application of the BBN in this study enables stakeholders and policymakers to better understand the linkages between different factors and behaviour.

Ecological Network on Benthic Diatom in Estuary Environment by Bayesian Belief Network Modelling

Ecological Network on Benthic Diatom in Estuary Environment by Bayesian Belief Network Modelling

Nexus between drivers of COVID-19 and country risks

COVID-19 has disrupted all spheres of life, including country risk regarding the exposure of economies to multi-dimensional risk drivers. However, it remains unexplored how COVID-19 has impacted different drivers of country risk in a probabilistic network setting. This paper uses two datasets on country-level COVID-19 and country risks to explore dependencies among associated drivers using a Bayesian Belief Network model. The drivers of COVID-19 risk, considered in this paper, are hazard and exposure, vulnerability and lack of coping capacity, whereas country risk drivers are economic, financing, political, business environment and commercial risks. The results show that business environment risk is significantly influenced by COVID-19 risk, whereas commercial risk (demand disruptions) is the least important factor driving COVID-19 and country risks. Further, country risk is mainly influenced by financing, political and economic risks. The contribution of this study is to explore the impact of various drivers associated with the country-level COVID-19 and country risks in a unified probabilistic network setting, which can help policy-makers prioritize drivers for managing the two risks.

Adoption of a probabilistic network model investigating country risk drivers that influence logistics performance indicators

There is a two-way interaction between the logistics performance and economic development of a country. The World Bank Group regularly publishes the Logistics Performance Index (LPI), which is a composite measure of the country-level logistics performance comprising six indicators, namely customs, infrastructure, international shipments, logistics competence, timeliness, and tracking and tracing. This study aims to explore dependencies among different drivers of country risk, including business environment, corruption, economic, environmental, financial, health and safety, and political risks, and the LPI indicators using a data-driven Bayesian Belief Network model. This study has made two unique contributions to the literature on environmental impact assessment. First, it explores dependencies among various country risk drivers and LPI indicators in a probabilistic network setting while mapping cause-effect relations between the input and outcome LPI indicators. Second, this study operationalizes a new data-driven methodology to help researchers and practitioners identify critical risk drivers influencing the LPI indicators. The results indicate a moderate to a strong correlation between individual risks and LPI indicators. Logistics competence is the most critical indicator impacting international shipments, timeliness, and tracking and tracing. On the other hand, economic and financial risks significantly impact all six LPI indicators.

Integrating a mixed‐cell cellular automata model and Bayesian belief network for ecosystem services optimization to guide ecological restoration and conservation

AbstractAn ecosystem is a complex system with a large number of dynamic variables, which poses challenges to the optimization of ecosystem services. However, traditional ecosystem services optimization methods do not take into account the complexity and uncertainty of variables. To address this complexity and uncertainty, we propose an innovative approach using a mixed‐cell cellular automata (MCCA) model and a Bayesian belief network (BBN) model for ecosystem service optimization. This approach was applied to the southern region of Sichuan Province, China, using an existing dataset to simulate land use patterns and predict ecosystem services in 2035 under different development scenarios. To achieve ecological restoration and conservation, we also determined the key factor combinations and key ecological regions at various ecosystem service levels. Results showed that ecological protection scenario design has important significance as a reference for maintaining and ameliorating regional ecosystem services and functions. We also identified that the highest level of ecosystem services was mainly located in the areas with the highest net primary productivity (NPP), the highest slope, the highest forestland area, and low ET. According to these findings, some suggestions for ecological restoration and conservation in key regions were put forward. This approach fully considers the uncertainty of factors; therefore, it can be used as an effective tool for designing ecosystem management strategies.

Exploring ecosystem services trade-offs using the Bayesian belief network model for ecological restoration decision-making: A case study in Guizhou Province, China

• A Bayesian belief network (BBN) model is developed to determine the driving factors regarding trade-off changes. • The multi-scenario scheme of ES optimization is obtained by visualizing the key state of key variables. • Suggestions for ecological restoration were given, and a good way is to adjust the combination of the key state of key variables. The relationships between trade-off and synergy in ecosystem services (ESs) is essential for eco-environmental management and restoration in Karst regions, and many researchers have focused heavily on this. However, to date, and to the best of our knowledge, we still lack an adequate understanding of the driving factors of relationship changes and the optimization of the spatial pattern of ESs in this region. In this study, a Bayesian belief network (BBN) model was developed to link three major ESs—net primary productivity (NPP), water yield (WY), and soil conservation (SC)—and determine the driving factors regarding trade-off changes. Furthermore, the multi-scenario scheme of ES optimization was obtained and suggestions are provided in this paper for ecological restoration and decision-making in Guizhou Province, China. The results revealed that the relationships of regional ESs were mainly synergistic, but the synergistic relationship was weakening, accompanied by a trade-off trend during the study period. The area ratios with an enhancement of trade-offs in NPP and SC, SC and WY, and NPP and WY were 17.9%, 28.6%, and 26.6%, respectively, which indicates a significant uncoordinated changes between WY and the two other ESs. The ES relationships in this region are underpinned by multiple factors. Among them, water resources are the primary natural factor that restricts the concerted development of ESs, and the afforestation area (AA) is the main human factor that lowers the trade-offs among ESs. The superiority and distribution scope of ecological restoration were recognized through a multi-scenario analysis of the BBN model. In terms of the superiority of ecological restoration, it was concluded that scenario 3 > scenario 2 > scenario 1; in terms of the restoration area, it was scenario 1 > scenario 2 > scenario 3. Duyun City, in the south of Guizhou Province, as well as surrounding counties, were regions with a high probability of trade-off occurrences, and they could be regions for future ES optimization. Based on the driving factors and the occurrence probability of trade-offs, the ecological restoration area was identified, and it was found the ESs trade-off in this region can be decreased by adjusting a combination of the key variables’ key states. This study provides insight towards achieving a mechanistic understanding and spatial expression of the driving forces underpinning trade-offs/synergies, in turn ensuring the accurate implementation of ecological restoration and the effective management of karst ecosystems.

Data-driven financial and operational risk management: Empirical evidence from the global tramp shipping industry

The global shipping industry has long suffered from high volatilities in freight rates and bunker fuel prices that lead to significant earnings risks. This paper aims to investigate the effectiveness of financial hedging and operational risk management strategies of 31 world leading tramp shipping companies through a Bayesian Belief Network (BBN) model using various data sources. Operational risk management strategies are categorized into long-term (e.g., fleet diversity and fleet age) and short-to-medium-term (e.g., relative trip distance, fleet repositioning flexibility, and trading diversity) strategies. We innovatively quantify the short-to-medium-term operational risk management strategies using Automatic Identification System (AIS) data. The results show that financial hedging can effectively reduce bunker fuel price risk exposure but cannot reduce freight rate risk exposure. Meanwhile, companies can use operational risk management strategies to effectively reduce both risk exposures. This study provides significant implications for shipping risk management.

Impact assessment of country risk on logistics performance using a Bayesian Belief Network model

PurposeThis paper aims to assess the impact of different drivers of country risk, including business environment, corruption, economic, environmental, financial, health and safety and political risks, on the country-level logistics performance.Design/methodology/approachThis study utilizes three datasets published by reputed international organizations, including the World Bank Group, AM Best and Global Risk Profile, to explore interactions among country risk drivers and the Logistics Performance Index (LPI) in a network setting. The LPI, published by the World Bank Group, is a composite measure of the country-level logistics performance. Using the three datasets, a Bayesian Belief Network (BBN) model is developed to investigate the relative importance of country risk drivers that influence logistics performance.FindingsThe results indicate a moderate to a strong correlation among individual risks and between individual risks and the LPI score. The financial risk significantly varies relative to the extreme states of the LPI score, whereas corruption risk and political risk are the most critical factors influencing the LPI score relative to their resilience and vulnerability potential, respectively.Originality/valueThis study has made two unique contributions to the literature on logistics performance assessment. First, to the best of the authors’ knowledge, this is the first study to establish associations between country risk drivers and country-level logistics performance in a probabilistic network setting. Second, a new BBN-based process has been proposed for logistics performance assessment and operationalized to help researchers and practitioners establish the relative importance of risk drivers influencing logistics performance. The key feature of the proposed process is adapting the BBN methodology to logistics performance assessment through the lens of risk analysis.

Development and comparison of 1-year survival models in patients with primary bone sarcomas: External validation of a Bayesian belief network model and creation and external validation of a new gradient boosting machine model.

Background:Bone sarcomas often present late with advanced stage at diagnosis and an according, varying short-term survival. In 2016, Nandra et al. generated a Bayesian belief network model for 1-year survival in patients with bone sarcomas. The purpose of this study is: (1) to externally validate the prior 1-year Bayesian belief network prediction model for survival in patients with bone sarcomas and (2) to develop a gradient boosting machine model using Nandra et al.’s cohort and evaluate whether the gradient boosting machine model outperforms the Bayesian belief network model when externally validated in an independent Danish population cohort.Material and Methods:The training cohort comprised 3493 patients newly diagnosed with bone sarcoma from the institutional prospectively maintained database at the Royal Orthopaedic Hospital, Birmingham, UK. The validation cohort comprised 771 patients with newly diagnosed bone sarcoma included from the Danish Sarcoma Registry during January 1, 2000–June 22, 2016. We performed area under receiver operator characteristic curve analysis, Brier score and decision curve analysis to evaluate the predictive performance of the models.Results:External validation of the Bayesian belief network 1-year prediction model demonstrated an area under receiver operator characteristic curve of 68% (95% confidence interval, 62%-73%). Area under receiver operator characteristic curve of the gradient boosting machine model demonstrated: 75% (95% confidence interval: 70%-80%), overall model performance by the Brier score was 0.09 (95% confidence interval: 0.077–0.11) and decision curve analysis demonstrated a positive net benefit for threshold probabilities above 0.5. External validation of the developed gradient boosting machine model demonstrated an area under receiver operator characteristic curve of 63% (95% confidence interval: 57%-68%), and the Brier score was 0.14 (95% confidence interval: 0.12–0.16).Conclusion:External validation of the 1-year Bayesian belief network survival model yielded a poor outcome based on a Danish population cohort validation. We successfully developed a gradient boosting machine 1-year survival model. The gradient boosting machine did not outperform the Bayesian belief network model based on external validation in a Danish population-based cohort.

Prioritizing Risks in Last Mile Delivery: A Bayesian Belief Network Approach

The remarkable explosion of e-commerce has marked the latest years of different industries and put forward a higher requirement for the last mile delivery. The last mile delivery is one of the most complex, costly, and inefficient processes along the entire logistics fulfillment chain in an e-commerce context. Its corresponding risks are major contributors to delivery failure. This work proposes a comprehensive framework on risk identification and analysis in the last mile delivery to support delivery planning. Risks were deduced from available literature, and others were induced through semi-structured interviews with experts in the field. Risks are categorized and the relative probability and severity of individual risks are determined. This study adopts a Bayesian Belief Network (BBN) model to identify the interdependency among risks and rank them, as the conventional ranking methods fail to take interdependency into account. The results indicate that privacy concerns, IT, and natural disasters are the most critical risks. This study will aid logistics service providers to ultimately decide the solutions of last mile delivery that need to be utilized by prioritizing last mile delivery possible risks to increase their competitiveness and market share, and minimize delivery costs.

A Bayesian Belief Network learning tool integrates multi-scale effects of riparian buffers on stream invertebrates

Riparian forest buffers have multiple benefits for biodiversity and ecosystem services in both freshwater and terrestrial habitats but are rarely implemented in water ecosystem management, partly reflecting the lack of information on the effectiveness of this measure. In this context, social learning is valuable to inform stakeholders of the efficacy of riparian vegetation in mitigating stream degradation. We aim to develop a Bayesian belief network (BBN) model for application as a learning tool to simulate and assess the reach- and segment-scale effects of riparian vegetation properties and land use on instream invertebrates. We surveyed reach-scale riparian conditions, extracted segment-scale riparian and subcatchment land use information from geographic information system data, and collected macroinvertebrate samples from four catchments in Europe (Belgium, Norway, Romania, and Sweden). We modelled the ecological condition based on the Average Score Per Taxon (ASPT) index, a macroinvertebrate-based index widely used in European bioassessment, as a function of different riparian variables using the BBN modelling approach. The results of the model simulations provided insights into the usefulness of riparian vegetation attributes in enhancing the ecological condition, with reach-scale riparian vegetation quality associated with the strongest improvements in ecological status. Specifically, reach-scale buffer vegetation of score 3 (i.e. moderate quality) generally results in the highest probability of a good ASPT score (99–100%). In contrast, a site with a narrow width of riparian trees and a small area of trees with reach-scale buffer vegetation of score 1 (i.e. low quality) predicts a high probability of a bad ASPT score (74%). The strengths of the BBN model are the ease of interpretation, fast simulation, ability to explicitly indicate uncertainty in model outcomes, and interactivity. These merits point to the potential use of the BBN model in workshop activities to stimulate key learning processes that help inform the management of riparian zones.

Applying Bayesian Belief Networks to Assess Alpine Grassland Degradation Risks: A Case Study in Northwest Sichuan, China.

Grasslands are crucial components of ecosystems. In recent years, owing to certain natural and socio-economic factors, alpine grassland ecosystems have experienced significant degradation. This study integrated the frequency ratio model (FR) and Bayesian belief networks (BBN) for grassland degradation risk assessment to mitigate several issues found in previous studies. Firstly, the identification of non-encroached degraded grasslands and shrub-encroached grasslands could help stakeholders more accurately understand the status of different types of alpine grassland degradation. In addition, the index discretization method based on the FR model can more accurately ascertain the relationship between grassland degradation and driving factors to improve the accuracy of results. On this basis, the application of BBN not only effectively expresses the complex causal relationships among various variables in the process of grassland degradation, but also solves the problem of identifying key factors and assessing grassland degradation risks under uncertain conditions caused by a lack of information. The obtained result showed that the accuracies based on the confusion matrix of the slope of NDVI change (NDVIs), shrub-encroached grasslands, and grassland degradation indicators in the BBN model were 85.27, 88.99, and 74.37%, respectively. The areas under the curve based on the ROC curve of NDVIs, shrub-encroached grasslands, and grassland degradation were 75.39% (P < 0.05), 66.57% (P < 0.05), and 66.11% (P < 0.05), respectively. Therefore, this model could be used to infer the probability of grassland degradation risk. The results obtained using the model showed that the area with a higher probability of degradation (P > 30%) was 2.22 million ha (15.94%), with 1.742 million ha (78.46%) based on NDVIs and 0.478 million ha (21.54%) based on shrub-encroached grasslands. Moreover, the higher probability of grassland degradation risk was mainly distributed in regions with lower vegetation coverage, lower temperatures, less potential evapotranspiration, and higher soil sand content. Our research can provide guidance for decision-makers when formulating scientific measures for alpine grassland restoration.

A Risk-Based Approach to Mine-Site Rehabilitation: Use of Bayesian Belief Network Modelling to Manage Dispersive Soil and Spoil

Dispersive spoil/soil management is a major environmental and economic challenge for active coal mines as well as sustainable mine closure across the globe. To explore and design a framework for managing dispersive spoil, considering the complexities as well as data availability, this paper has developed a Bayesian Belief Network (BBN)-a probabilistic predictive framework to support practical and cost-effective decisions for the management of dispersive spoil. This approach enabled incorporation of expert knowledge where data were insufficient for modelling purposes. The performance of the model was validated using field data from actively managed mine sites and found to be consistent in the prediction of soil erosion and ground cover. Agreement between predicted soil erosion probability and field observations was greater than 74%, and greater than 70% for ground cover protection. The model performance was further noticeably improved by calibration of Conditional Probability Tables (CPTs). This demonstrates the value of the BBN modelling approach, whereby the use of currently best-available data can provide a practical result, with the capacity for significant model improvement over time as more (targeted) data come to hand.

A Bayesian Belief Network Model for the Risk Assessment and Management of Premature Screen-Out during Hydraulic Fracturing

Hydraulic fracturing is a well completion technique for Oil and Gas production enhancement in both conventional and unconventional reservoirs. However, it can result in the unfavorable consequence of the premature screen-out, which occurs due to the proppant bridging across the perforations or similar restricted flow areas.The objective of this work is to propose a novel framework of analysis that enables to quantify the risk of screen-out occurrence, to identify the riskiest scenarios and to determine the best risk mitigation strategies. The premature screen-out problem is addressed within a Risk Management and Control Process, wherein the qualitative and quantitative assessments of the early screen-out risk are performed by a Features, Events and Processes Analysis structured with a Bayesian Belief Network. The BBN probabilities are subject to a thorough uncertainty and sensitivity analysis. Sensitivity analysis is performed by the Sobol's variance decomposition method and the identified most influential probabilities of the BBN are re-estimated in order to reduce the output uncertainty.Finally, risk mitigation plans are formulated using risk importance measures to identify the riskiest scenarios and cost-benefit analysis to determine the optimal risk reduction actionsThe developed framework has been applied to a case study of vertical wells.

Scalability constraints in Disruptive Agricultural Technologies (DATs) along Value Chain on agricultural production in South Sudan

Scalability known as the capacity of input variables along the Value Chain (VC) to effect transformative changes on agricultural production was evaluated for a farming system in Juba County of Central Equatoria State (CES), South Sudan. These transformative input variables commonly referred to as, Disruptive Agricultural Technologies (DATs) in the form of advisory, material as well as technological variables were shown to positively influence agricultural production from a default state. The objective of this study was to find out how a probability-based Bayesian Belief Network (BBN) software NETICA could be applied to assess as well as upscale the level of agricultural production P(Prodlevel | ) from a data input domain D. Simulation using a 700 kg ha-1 of cowpea yield at 50% Cumulative Probability Distribution (CPD) as a calibrant, the backcasting method showed that, scaling up of marginal probabilities in agrotechnology and financial resources from 0.025 to 0.1 (25% increment) and from 0.015 to 0.03 (50% increment) respectively, while keeping other input variables unchanged, increased cowpea yield from 692.9 to 783.1 kg ha-1 (about 12% increment). Conversely, where no DATs were introduced as in the worst-case scenario, production level was comparatively lower. The BBN model is thus, an indispensable tool that can provide useful information on scaling up agricultural production and hence improve livelihood opportunities in Juba County. However, for sustainable agricultural production, scalability may be constrained by spatial-temporal, environmental and socio-economic imperatives as well as on availability, accessibility, affordability of all input variables.

General corrosion vulnerability assessment using a Bayesian belief network model incorporating experimental corrosion data for X60 pipe steel

External corrosion is one of the leading causes of pipe failure in the oil and gas industry. In this study, a Bayesian belief network (BBN) model has been developed using corrosion rate (CR) data obtained from experimental test results and analytical burst failure models. The BBN model for CR was coupled with a time marching simulation to obtain corrosion defects and quantify failure pressure capacity. Finally, in a reliability framework, the failure pressure capacity was coupled with operating pressure to obtain the probability of failure. Furthermore, the developed BBN model was used to perform a parametric study to identify the critical parameters for the CR. The outcome of the study indicated that the proposed BBN model can be useful to integrate experimental and analytical models to derive reliability of a pipeline operating under various conditions.

Bayesian network based decision support for predicting and mitigating delay risk in TBM tunnel projects

Tunnel projects involve high levels of uncertainty stemming from the vagueness of geological conditions and the complexity of the mechanized tunnel boring process. Delay risk assessment is carried out by project managers to identify critical risk factors leading to time and cost overruns and formulate strategies to meet the project targets under different scenarios. In this study, a Bayesian Belief Network (BBN) based risk assessment method was developed for Tunnel Boring Machine (TBM) tunnel projects to predict delay. Based on the BBN model, a decision-support tool, BBN Tunnel, was developed to assess delay considering the impacts of implementing alternative risk mitigation strategies. The tool developed in collaboration with a company was utilized in a tunnel project to test its usability in practice. The results demonstrated that BBN Tunnel and risk assessment method could be used to model interrelations between risk factors, construct a risk network, predict delay and help decision-makers formulate cost-effective risk mitigation strategies.

Bayesian belief network modelling of household food security in rural South Africa

BackgroundAchieving food security remains a key challenge for public policy throughout the world. As such, understanding the determinants of food insecurity and the causal relationships between them is an important scientific question. We aim to construct a Bayesian belief network model of food security in rural South Africa to act as a tool for decision support in the design of interventions.MethodsHere, we use data from the Agincourt Health and Socio-demographic Surveillance System (HDSS) study area, which is close to the Mozambique border in a low-income region of South Africa, together with Bayesian belief network (BBN) methodology to address this question.ResultsWe find that a combination of expert elicitation and learning from data produces the most credible set of causal relationships, as well as the greatest predictive performance with 10-fold cross validation resulting in a Briers score 0.0846, information reward of 0.5590, and Bayesian information reward of 0.0057. We report the resulting model as a directed acyclic graph (DAG) that can be used to model the expected effects of complex interventions to improve food security. Applications to sensitivity analyses and interventional simulations show ways the model can be applied as tool for decision support for human experts in deciding on interventions.ConclusionsThe resulting models can form the basis of the iterative generation of a robust causal model of household food security in the Agincourt HDSS study area and in other similar populations.