Bayesian Belief Network Structure Research Articles

Enhancing risk assessment: an improved Bayesian network approach for analyzing interactions among risks

PurposeIn traditional Bayesian belief networks (BBNs), a large amount of data are required to complete network parameters, which makes it impractical. In addition, no systematic method has been used to create the structure of the BBN. The aims of this study are to: (1) decrease the number of questions and time and effort required for completing the parameters of the BBN and (2) present a simple and apprehensible method for creating the BBN structure based on the expert knowledge.Design/methodology/approachIn this study, by combining the decision-making trial and evaluation laboratory (DEMATEL), interpretive structural modeling (ISM) and BBN, a model is introduced that can form the project risk network and analyze the impact of risk factors on project cost quantitatively based on the expert knowledge. The ranked node method (RNM) is then used to complete the parametric part of the BBN using the same data obtained from the experts to analyze DEMATEL.FindingsCompared to the traditional BBN, the proposed method will significantly reduce the time and effort required to elicit network parameters and makes it easy to create a BBN structure. The results obtained from the implementation of the model on a mass housing project showed that considering the identified risk factors, the cost overruns relating to material, equipment, workforce and overhead cost were 37.6, 39.5, 42 and 40.1%, respectively.Research limitations/implicationsCompared to the traditional BBN, the proposed method will significantly reduce the time and effort required to elicit network parameters and makes it easy to create a BBN structure. The results obtained from the implementation of the model on a mass housing project showed that considering the identified risk factors, the cost overruns relating to material, equipment, workforce and overhead cost were 37.6, 39.5, 42 and 40.1%, respectively. The obtained results are based on a single case study project and may not be readily generalizable.Originality/valueThe presented framework makes the BBN more practical for quantitatively assessing the impact of risk on project costs. This helps to manage financial issues, which is one of the main reasons for project bankruptcy.

Control Action Demand for Major Cities Pollutants Using Bayesian Belief Networks

Air pollutants in large cities are an overwhelming problem and have been responsible for many premature deaths all around the world. Risk perception maps how people evaluate a hazard in a subjective manner using different statistical tools. In this paper, we use of Bayesian belief network (BBN) to estimate the likelihood of control action demand from people towards authorities based on a proposed framework relating risk perception, risk judgment, and demand. The results showed that it is possible to model control action demand based on BBN structure, given an observed scenario for risk perception and judgment. Different pollutants were compared and the method distinguished the most feared from the lesser feared.

Score-based Bayesian belief network structure learning in damage risk modelling of mining areas building development

The article describes the results of using the application of Bayesian approach, to create a decision support system to evaluate the risk of the damages to prefabricated reinforced concrete (RC) buildings, that have been exposed to the impacts from the industrial environment of mines in the form of spatial subsidence and tremors. This problem is important from the point of view of both structural safety and utility aspects. It concerns many buildings, in which the damage is often caused by the excessive consumption of thermal energy used for heating interiors or dehumidifying damp walls. The above issues make the subject extremely important from a technical, socio-economical, and energy point of view, therefore it can be categorised as sustainable development problem. The Bayesian belief network (BBN) is used to build a risk model that encompasses all the issues and variables that describe the damage process associated with the entire scope.The use of BBNs is an innovative approach in this interdisciplinary area, that encompasses civil engineering and mining environmental protection. There are two types of score-based methods for learning the structure of a BBN from data which will be applied and compared. The first is an iterative method for searching the space of potential solutions, based on the Tabu-search algorithm. The second approach used a method of stochastic searching for a space of structures (node orders) based on the global optimisation algorithm Simulated Annealing (SA). The analyses were performed in R, using the catnet and bnlearn packages. The basis for the study was a database relating to 129 prefabricated reinforced concrete buildings with a wall bearing structure, located in a copper mining area in Poland. Finally, there were eight damage intensity indexes (wui) obtained from using the above models when analysing the risk of damage to structural and finishing elements. The most effective method turned out to be using the Tabu-search algorithm with the assumed K2 criterion. As a result, the following damage indexes were obtained – 91.43%–97.14% correctly classified patterns for the training set, while for the test set – 79.17%–100%. The conducted research is interdisciplinary in the field of civil engineering and mining and environmental engineering, and the very issue of damage is also important from the socio-economical and energy point of view, especially for a large number of buildings.

A Bayesian Belief Network method for bridge deterioration detection

Bridges are one of the most important assets of transportation networks. A closure of a bridge can increase the vulnerability of the geographic area served by such networks, as it reduces the number of available routes. Condition monitoring and deterioration detection methods can be used to monitor the health state of a bridge and enable detection of early signs of deterioration. In this paper, a novel Bayesian Belief Network (BBN) methodology for bridge deterioration detection is proposed. A method to build a BBN structure and to define the Conditional Probability Tables (CPTs) is presented first. Then evidence of the bridge behaviour (such as bridge displacement or acceleration due to traffic) is used as an input to the BBN model, the probability of the health state of whole bridge and its elements is updated and the levels of deterioration are detected. The methodology is illustrated using a Finite Element Model (FEM) of a steel truss bridge, and for an in-field post-tensioned concrete bridge.

Evaluating the Probability of Power Loss in Ship Electric Propulsion Systems Based on Bayesian Belief Networks

AbstractThe development of smart and green ships has led to the wide use of electric propulsion systems, creating many safety problems. The ship electric propulsion system (SEPS) loses power frequently, creating significant risk in maritime transportation. This article proposed a probability evaluation method to assess SEPS power loss, based on Bayesian Belief Networks (BBNs). The BBN structure of power loss in a SEPS is considered the main component causing SEPS disruption. The prior probability and conditional probability tables of the model were generated using statistical data and expert judgment. A case study was conducted to illustrate the proposed model. Finally, a sensitivity analysis was used to identify several key indicators, including human factors, navigation conditions, winding failure, aging, and parameter setting. Compared with related research, the proposed method is rational and effective. As such, the method presented in this article provides a practical and flexible way to evaluate the probability of power loss in a SEPS. The model enables the safety manager to discover key influencing factors and components during ship operations, providing effective ways to reduce ship power loss.

Reprint of “Validation approaches of an expert-based Bayesian Belief Network in northern Ghana, West Africa”

Model validation is a precondition for credibility and acceptance of a model. However, it appears that there is no scientific standard for validation of Bayesian Belief Networks (BBNs). In this paper, we present a novel combination of BBN validation approaches. A set of qualitative and quantitative validation approaches for the BBN structure, the Conditional Probability Tables and the BBN output is presented and discussed. The validation approaches were tested for a BBN on food provision under land use and land cover changes and different weather scenarios in rural northern Ghana. Experts played an important role in developing and validating the BBN due to data scarcity. Furthermore, selected nodes and the BBN output were compared to existing data. A sensitivity analysis was conducted. Validation approaches show that structural model uncertainties are still high and reliability of input data is low. However, the extreme-condition test shows that the BBN works according to the assumed system understanding that food provision decreases under floods, droughts, land pressure and poverty. Therefore, the BBN can provide general trends for output nodes but lacks reliability if detailed results of single system components are required.

Validation approaches of an expert-based Bayesian Belief Network in Northern Ghana, West Africa

Model validation is a precondition for credibility and acceptance of a model. However, it appears that there is no scientific standard for validation of Bayesian Belief Networks (BBNs). In this paper, we present a novel combination of BBN validation approaches. A set of qualitative and quantitative validation approaches for the BBN structure, the Conditional Probability Tables and the BBN output is presented and discussed. The validation approaches were tested for a BBN on food provision under land use and land cover changes and different weather scenarios in rural northern Ghana. Experts played an important role in developing and validating the BBN due to data scarcity. Furthermore, selected nodes and the BBN output were compared to existing data. A sensitivity analysis was conducted. Validation approaches show that structural model uncertainties are still high and reliability of input data is low. However, the extreme-condition test shows that the BBN works according to the assumed system understanding that food provision decreases under floods, droughts, land pressure and poverty. Therefore, the BBN can provide general trends for output nodes but lacks reliability if detailed results of single system components are required.

A Parameter Free BBN Discriminant Function for Optimum Model Complexity versus Goodness of Data Fitting

Bayesian Belief Network (BBN) is an appealing classification model for learning causal and noncausal dependencies among a set of query variables. It is a challenging task to learning BBN structure from observational data because of pool of large number of candidate network structures. In this study, we have addressed the issue of goodness of data fitting versus model complexity. While doing so, we have proposed discriminant function which is non-parametric, free of implicit assumptions but delivering better classification accuracy in structure learning. The contribution in this study is twofold, first contribution (discriminant function) is in BBN structure learning and second contribution is for Decision Stump classifier. While designing the novel discriminant function, we analyzed the underlying relationship between the characteristics of data and accuracy of decision stump classifier. We introduced a meta characteristic measure AMfDS (herein known as Affinity Metric for Decision Stump) which is quite useful in prediction of classification accuracy of Decision Stump. AMfDS requires a single scan of the dataset.

Application of Bayesian Belief Networks to quantify and map areas at risk to soil threats: Using soil compaction as an example

The assessment of areas at risk from various soil threats is a key task within the proposed EU Soil Framework Directive. Such assessment is, however, hampered by the complex nature of the soil threats, which result from the sometimes poorly understood interaction of various soil physical properties, climatic factors and land management practices. Methodologies for risk assessment of soil threats are needed to protect the soil quality for future generations and to target resources to the areas at greatest risk. We present here a generic risk framework for the development of Bayesian Belief Networks (BBNs) to estimate the risk from soil threats. The generic BBN structure follows a standard risk assessment approach, where the risk is quantified by combining assessments of vulnerability and exposure. The soil's vulnerability to a given threat is determined from inherent soil and site characteristics as well as from climatic factors influencing soil characteristics, while the exposure estimate is based on an evaluation of the stresses inflicted by land management and climate. The generic framework is demonstrated by taking soil compaction as an example. Soil compaction is a major threat to soil function particularly in highly managed agricultural systems and is known to have many adverse effects on farming systems including decreased crop yield and soil productivity, increased management costs, increased emissions of greenhouse gases, and decreased water infiltration into the soil leading to accelerated run-off and risk of soil erosion. Existing modelling approaches to predict soil compaction risk either require data on soil mechanical behaviour that are difficult and expensive to collect, or are expert-based systems that are highly subjective and sometimes cannot accommodate the myriad of processes underlying compaction risk. Using the generic framework, a detailed BBN for assessing the risk of soil compaction is developed. The BBN allows for combining available data from standard soil surveys and land use databases with qualitative expert knowledge and explicitly accounts for uncertainties in the assessment of the risk. The BBN is applied to identify the distribution of the compaction risk across Scotland using data from the National Soils Inventory of Scotland.

An evolutionary Bayesian belief network methodology for participatory decision making under uncertainty: An application to groundwater management

An integrated participatory approach based on Bayesian belief network (BBN) and evolutionary multiobjective optimization is proposed as an efficient decision-making tool in complex management problems. The proposed methodology incorporates all the available evidence and conflicting objectives to evaluate implications of alternative actions in the decision-making process and suggests best decision pathways under uncertainty. A BBN provides a framework within which the contributions of stakeholders can be taken into account. It allows a range of different factors and their probabilistic relationship to be considered simultaneously. It takes into account uncertainty by assigning probability to those variables whose states are not certain. The integration of BBN with evolutionary multiobjective optimization allows the analysis of tradeoff between different objectives and incorporation and acknowledgement of a broader set of decision goals into the search and decision-making process. The proposed methodology can be used as a decision support tool to model decision-making processes for complex problems. It deals with uncertainties in decision making pertaining to human behavior and checks for consistency of the developed BBN structure and the parameters of the probabilistic relationship by uncovering discrepancies in the decision analysis process (e.g., bias in completeness or redundancy of the model based on a utility function). It generates a set of efficient management options (appropriate combinations of interventions) that balances conflicting objectives. The effectiveness of the proposed methodology is discussed through application to a real case study. It is shown that it successfully identifies any inconsistencies in the developed BBN models and generates large numbers of management options that achieve an optimal tradeoff between different objectives.

Bayesian belief network learning algorithms for modeling contextual relationships in natural imagery: a comparative study

This paper presents a comparative study of Bayesian belief network structure learning algorithms with a view to identify a suitable algorithm for modeling the contextual relations among objects typically found in natural imagery. Four popular structure learning algorithms are compared: two constraint-based algorithms (PC proposed by Spirtes and Glymour and Fast Incremental Association Markov Blanket proposed by Yaramakala and Margaritis), a score-based algorithm (Hill Climbing as implemented by Daly), and a hybrid algorithm (Max-Min Hill Climbing proposed by Tsamardinos et al.). Contrary to the belief regarding the superiority of constraint-based approaches, our empirical results show that a score-based approach performs better on our context dataset in terms of prediction power and learning time. The hybrid algorithm could achieve similar prediction performance as the score-based approach, but requires longer time to learn the desired network. Another interesting fact the study has revealed is the existence of strong correspondence between the linear correlation pattern within the dataset and the edges found in the learned networks.

Modeling Land-Use Decision Behavior with Bayesian Belief Networks

The ability to incorporate and manage the different drivers of land-use change in a modeling process is one of the key challenges because they are complex and are both quantitative and qualitative in nature. This paper uses Bayesian belief networks (BBN) to incorporate characteristics of land managers in the modeling process and to enhance our understanding of land-use change based on the limited and disparate sources of information. One of the two models based on spatial data represented land managers in the form of a quantitative variable, the area of individual holdings, whereas the other model included qualitative data from a survey of land managers. Random samples from the spatial data provided evidence of the relationship between the different variables, which I used to develop the BBN structure. The model was tested for four different posterior probability distributions, and results showed that the trained and learned models are better at predicting land use than the uniform and random models. The inference from the model demonstrated the constraints that biophysical characteristics impose on land managers; for older land managers without heirs, there is a higher probability of the land use being arable agriculture. The results show the benefits of incorporating a more complex notion of land managers in land-use models, and of using different empirical data sources in the modeling process. Future research should focus on incorporating more complex social processes into the modeling structure, as well as incorporating spatio- temporal dynamics in a BBN.

A Bayesian method for learning belief networks that contain hidden variables

This paper presents a Bayesian method for computing the probability of a Bayesian belief-network structure from a database. In particular, the paper focuses on computing the probability of a belief-network structure that contains a hidden (latent) variable. A hidden variable represents a postulated entity that has not been directly measured. After reviewing related techniques, which previously were reported, this paper presents a new, more efficient method for handling hidden variables in belief networks.