Applications of Bayesian belief networks in water resource management: A systematic review

Bayesian belief networks (BBNs) are probabilistic graphical models that can capture and integrate both quantitative and qualitative data, thus accommodating data-limited conditions. This paper systematically reviews applications of BBNs with respect to spatial factors, water domains, and the consideration of climate change impacts. The methods used for constructing and validating BBN models, and their applications in different forms of decision-making support are examined. Most reviewed publications originate from developed countries (70%), in temperate climate zones (42%), and focus mainly on water quality (42%). In 60% of the reviewed applications model validation was based on the expert or stakeholder evaluation and sensitivity analysis, and whilst in 27% model performance was not discussed. Most reviewed articles applied BBNs in strategic decision-making contexts (52%). Integrated modelling tools for addressing challenges of dynamically complex systems were also reviewed by analysing the strengths and weaknesses of BBNs, and integration of BBNs with other modelling tools.     

CROSPAL,software that uses agronomic expert knowledge to assist modules selection for crop growth simulation

Crop growth models are used for a wide range of objectives. For each objective a specific model has to be developed, because the reusability of a model is often limited by the necessity of a fundamental restructuring to adapt it to a different objective. To overcome this limitation, we developed a method to facilitate model restructuring by a novel combination of software technology with expert knowledge.This resulted in the decision-making software application CROSPAL (CROp Simulator: Picking and Assembling Libraries). CROSPAL includes (1) a library of processes each containing different modelling approaches for each crop physiological process and (2) a procedure based on expert knowledge of how to combine the different processes for the objective of the simulation.A brief overview of the state of the art in crop modelling is presented, followed by an account of the developed concept to improve flexibility in crop modelling considering expert knowledge. We describe the design of the software and how expert knowledge is integrated. The use of CROSPAL is illustrated for the modelling of crop phenology. We conclude that CROSPAL is a helpful tool to improve flexibility in crop modelling considering expert knowledge but further development and evaluation is required to extend its range of application to more processes and issues crop modelling is presently addressing.     

Mining and visualising ordinal data with non-parametric continuous BBNs

Data mining is the process of extracting and analysing information from large databases. Graphical models are a suitable framework for probabilistic modelling. A Bayesian Belief Net (BBN) is a probabilistic graphical model, which represents joint distributions in an intuitive and efficient way. It encodes the probability density (or mass) function of a set of variables by specifying a number of conditional independence statements in the form of a directed acyclic graph. Specifying the structure of the model is one of the most important design choices in graphical modelling. Notwithstanding their potential, there are only a limited number of applications of graphical models on very complex and large databases. A method for mining ordinal multivariate data using non-parametric BBNs is presented. The main advantage of this method is that it can handle a large number of continuous variables, without making any assumptions about their marginal distributions, in a very fast manner. Once the BBN is learned from data, it can be further used for prediction. This approach allows for rapid conditionalisation, which is a very important feature of a BBN from a user’s standpoint.     

On the use of Bayesian belief networks for the prediction of software productivity

In spite of numerous methods proposed, software cost estimation remains an open issue and in most situations expert judgment is still being used. In this paper, we propose the use of Bayesian belief networks (BBNs), already applied in other software engineering areas, to support expert judgment in software cost estimation. We briefly present BBNs and their advantages for expert opinion support and we propose their use for productivity estimation. We illustrate our approach by giving two examples, one based on the COCOMO81 cost factors and a second one, dealing with productivity in ERP system localization.     

The event bush as a semantic-based numerical approach to natural hazard assessment (exemplified by volcanology)

The event bush is a new formalism for organizing knowledge in various fields of geoscience, particularly suitable for hazard assessment purposes. Acting as an intermediary between expert knowledge and the well-established field of Bayesian belief networks, the event bush allows at the same time a variety of other applications, linking geoscientific knowledge to the field of artificial intelligence and uniting probabilistic, deterministic, and fuzzy approaches. In this paper, we present basic principles, mathematical formulation, guidelines for application, and examples, including the connection with Bayesian belief networks. Further development of the method will include spatial and temporal modelling, implementation in mapping in GIS medium, formalization by means of predicate logic, definition of variable states in BBNs by membership functions based on the event bush semantics, and other applications.     

Expert support systems: an intermediate reality for management

The application of expert systems (ES) in several technical areas has proven to be rather successful. However, the utilization of ES in strategic management areas has encountered some difficulties. This new endeavour deals with managerial decision making for multi-disciplinary problems involving many behavioural variables. A pragmatic solution is to utilize expert support systems (ESS) as an intermediate measure. The intention is to use ESS as supporting tools rather than using them to replace human beings. This is a situation in which machines complement human beings in decision making. The ESS will provide some knowledge and reasoning procedures while the decision-maker will supplement it with the overall problem-solving direction. This co-operation between man and machine will better accommodate the deficiencies in understanding human behavioural variables by ES. This study analyses the two-pronged development of ES and the advantages of using ESS during the interim period to overcome the uncertainties of human behaviour.     

Optimization Models for Training Belief-Rule-Based Systems

A belief rule-base inference methodology using the evidential reasoning approach (RIMER) has been developed recently, where a new belief rule representation scheme is proposed to extend traditional IF-THEN rules. The belief rule expression matrix in RIMER provides a compact framework for representing expert knowledge. However, it is difficult to accurately determine the parameters of a belief rule base (BRB) entirely subjectively, particularly, for a large-scale BRB with hundreds or even thousands of rules. In addition, a change in rule weight or attribute weight may lead to changes in the performance of a BRB. As such, there is a need to develop a supporting mechanism that can be used to train, in a locally optimal way, a BRB that is initially built using expert knowledge. In this paper, several new optimization models for locally training a BRB are developed. The new models are either single- or multiple-objective nonlinear optimization problems. The main feature of these new models is that only partial input and output information is required, which can be either incomplete or vague, either numerical or judgmental, or mixed. The models can be used to fine tune a BRB whose internal structure is initially decided by experts' domain-specific knowledge or common sense judgments. As such, a wide range of knowledge representation schemes can be handled, thereby facilitating the construction of various types of BRB systems. Conclusions drawn from such a trained BRB with partially built-in expert knowledge can simulate real situations in a meaningful, consistent, and locally optimal way. A numerical study for a hierarchical rule base is examined to demonstrate how the new models can be implemented as well as their potential applications.     

Elicitator: An expert elicitation tool for regression in ecology

Expert elicitation is the process of retrieving and quantifying expert knowledge in a particular domain. Such information is of particular value when the empirical data is expensive, limited or unreliable. This paper describes a new software tool, called Elicitator, which assists in quantifying expert knowledge in a form suitable for use as a prior model in Bayesian regression. Potential environmental domains for applying this elicitation tool include habitat modelling, assessing detectability or eradication, ecological condition assessments, risk analysis and quantifying inputs to complex models of ecological processes. The tool has been developed to be user-friendly, extensible and facilitate consistent and repeatable elicitation of expert knowledge across these various domains. We demonstrate its application to elicitation for logistic regression in a geographically based ecological context. The underlying statistical methodology is also novel, utilizing an indirect elicitation approach to target expert knowledge on a case-by-case basis. For several elicitation sites (or cases), experts are asked simply to quantify their estimated ecological response (e.g. probability of presence), and its range of plausible values, after inspecting (habitat) covariates via GIS.     

Evaluation and comparison of data-driven and knowledge-supported Bayesian Belief Networks to assess the habitat suitability for alien macroinvertebrates

Defining habitats vulnerable to invasion is important to support the management of invasive alien species (IAS). We developed and applied data-driven and knowledge-supported data-driven Bayesian Belief Networks (BBNs) to assess the habitat suitability for alien gammarids. Data-driven model development using a Naive Bayes classifier and equal width discretization resulted in a habitat suitability model with a moderate technical performance (CCI = 68% K = 0.33). Although the structure of the knowledge-supported model yielded important ecological insight between environmental and biotic variables and the occurrence of alien gammarids, the performance was lower (CCI = 60% K = 0.19) compared to the purely data-driven model. The lower predictive performance of the knowledge-supported model may be attributed to its higher model complexity. Our study shows that BBNs can support the management of IAS as they are visually appealing, transparent models that facilitate integration of monitoring data and expert knowledge.     

Eliciting and modelling expert knowledge

This paper evaluates the usefulness of various psychological techniques that can be utilized to elicit and model expert knowledge for subsequent representation in rule-based expert systems. Interviewing, protocol analysis and multidimensional scaling are described and evaluated as complementary methods of knowledge elicitation. In addition ‘context-focusing’ and card-sorting are introduced as short-cut methods for the knowledge engineer's ‘tool box’.It is argued that expert knowledge about uncertainty can be represented as subjective probabilities and that these assessments can (and therefore should) be checked for consistency and coherence as a pre-condition for realism.Finally, the issue of whether it is possible to improve upon expert judgement is discussed and evidence is reviewed which shows that, in repetitive decision-making situations, statistical models of the expert can out-perform the expert on whom the models are based. Statistical modelling has a valid but limited application as a replacement for expert judgement.     

基于人工神经网络和专家系统Shel的城市发展水平综合评价专家系统的研究

基于人工神经网络（ＡＮＮ）和专家系统Ｓｈｅｌ（ＥＳＳ），提出一种城市发展水平综合评价专家系统（ＣＣＥＥＳ）的基本结构模式，并对ＣＣＥＥＳ中评价指标规范化方法、基于ＡＮＮ的知识获取方法、基于决策表的知识库自动生成方法以及推理机的工作原理作了描述。ＣＣＥＥＳ现已在ＩＢＭＰＣ３８６／４８６微机上用ＥＳＳ—ＶＰ－Ｅｘｐｅｒｔ和ＴｕｒｂｏＣ、ＦＯＸＢＡＳＥ＋语言实现，并取得了较好的应用效果     

CIMOSA modelling processes

Engineering, integrating and managing complex enterprises requires the understanding, and the ability to partition and simplify their operational complexity. Enterprise modelling supports these requirements by providing means for describing process oriented systems and decomposing those into manageable pieces. However, enterprise modelling requires both a common modelling language and a sufficient modelling methodology. The language provides for common understanding on enterprise models across the industrial community. Modelling methodologies will guide users through the rather complex enterprise modelling tasks. Depending on the skills and the tasks of the modelling person, different methodologies will be implemented in the supporting modelling tool. The paper presents both a methodology for the modelling expert and one for the business user. Whereas the modelling expert will be involved in creating new models, structuring the model contents and developing new modelling components, the business user will use process models for decision support. The latter therefore has a need to modify and adapt enterprise models to represent operational alternatives. A methodology for this type of work has to be based on menus. Menus which are created and maintained by the modelling expert. The business user will mostly work with existing process models. He will evaluate process alternatives and will implement the best solution as the new model of his tasks. This mode of operation will thereby provide for automatic update of the models and will keep the models in sync with the changing reality.     

Using a hybrid rule-based approach in developing an intelligent tutoring system with knowledge acquisition and update capabilities

In this paper, we present the architecture and describe the functionality of an Intelligent Tutoring System (ITS), which uses an expert system to make decisions during the teaching process. The expert system uses neurules for knowledge representation of the pedagogical knowledge. Neurules are a type of hybrid rules integrating symbolic rules with neurocomputing. The expert system consists of three components: the user modelling unit, the pedagogical unit and the inference system. The pedagogical knowledge is distributed in a number of neurule bases within the user modelling and the pedagogical unit. Another important component of the ITS, for both its development and maintenance, is its knowledge management unit, which provides knowledge acquisition and knowledge update capabilities to the system, that is, offers expert knowledge authoring capabilities to the system.     

Neurofuzzy identification of an autonomous underwater vehicle

Neurofuzzy modelling is ideally suited to many nonlinear system identification and data modelling applications. By combining the attractive attributes of fuzzy systems and neural networks transparent models of ill-defined systems can be identified. Available expert a priori knowledge is used to construct an initial model. Data modelling techniques from the neural network, statistical and conventional system identification communities are then used to adapt these models. As a result accurate parsimonious models which are transparent and easy to validate are identified. Recent advances in the datadriven identification algorithms have now made neurofuzzy modelling appropriate for high-dimensional problems for which the expert knowledge and data may be of a poor quality. In this paper neurofuzzy modelling techniques are presented. This powerful approach to system identification is demonstrated by its application to the identification of an Autonomous Underwater Vehicle (AUV).     

Model-based expert system to automatically adapt milling forces in Pareto optimal multi-objective working points

The objective of this paper is to present an open and modular expert rule-based system in order to automatically select cutting parameters in milling operations. The knowledge base of the system presents considerations of stability, machine drives efficiency and restrictions while adaptively controlling milling forces in suitable working points. Moreover, a novel classical cost function has been conceived and constructed to Pareto-optimise cutting parameters subjected to multi-objective purposes, namely: tool-life, surface roughness, material remove rate and stability rate parameter. Different Pareto optimal front solutions can be obtained modulating the weighting factors of the cost function. Additional rules have been added in order to manually and/or automatically modulate this cost function. Furthermore, a database which relates weighting factors, cutting conditions and cost function variables is produced for learning purposes. Chatter detection and suppression system automatically feedback to the system to take into account non-modelled disturbances. Finally, since the knowledge of the system is basically obtained from mathematical models, the possibility of combining experience and knowledge from expert engineers and operators is included. In this way, best practice from mathematical modelling and expert engineers and operators is joined in one system obtaining a full, automated system combining the best of each world.As a result, the expert rule-based system selects Pareto optimal cutting conditions for a broad range of milling processes, sorting out automatically different problems such as chatter vibrations, incorporating model reference adaptive control (MRAC) of forces. This procedure is intuitive, being executed in the same way as a human expert would do and it provides the possibility to interact with expert engineers and operators in order to take into account their experience and knowledge. Finally, the expert system is designed in modular form allowing incorporating new functionalities in rule based forms to them or just adding new modules to improve the performance of the milling system.     

Improving ecosystem services modelling: Insights from a Bayesian network tools review

Numerous studies attempt to unravel the role played by Biodiversity in ecosystems and ES reliance on Biodiversity. Achieving this aim is difficult given: the multi-layered Biodiversity-ES relationship; the temporal and spatial heterogeneity of ES; and, the interactions between biotic and abiotic components in ecosystems influencing processes and services. Bayesian networks have recently gained importance in ecological modelling. The integration of empirical data with expert knowledge and the explicit treatment of uncertainties, demonstrate their usefulness. Publications describing network-based Biodiversity-ES models, demonstrate their application is still limited. A watershed's environmental risk management network modelled from a Biodiversity-ES perspective is discussed. It demonstrates an improvement on conventional approaches, expressing risk in terms of the underlying causal relations between environmental risk events, triggers, controls and consequences. The model is developed in AgenaRisk and two other tools, Netica and Hugin. A comparison between them highlights the dependence on the tool of choice.     

The reuse of knowledge: a user-centred approach

The motivation for the work reported in this paper is the belief that not only is it beneficial to reuse knowledge but it is essential if we wish to build knowledge-based systems (KBS) that meet the needs of users. The focus of most KBS research is on complex modelling at the knowledge level which requires a knowledge engineer to act as the intermediary between the expert and the system. The type of reuse primarily considered is the reuse of ontologies or problem-solving methods so that improvements can be made in system quality and development time. However, there is little focus on the needs of users to access the knowledge in a variety of ways according to the individual's decision style or situation. The system described in this paper seeks to support the user in a number of different activities including knowledge acquisition, inferencing, maintenance, tutoring, critiquing, “what-if” analysis, explanation and modelling. The ability to ask different types of questions and to explore the knowledge in alternative ways is a different type of knowledge reuse. The knowledge acquisition and representation technique used as the foundation is known as ripple-down rules (RDR). To support the exploration activities, RDR have been combined with formal concept analysis which automatically generates an abstraction hierarchy from the low-level RDR assertions. The paper suggests that rapid and incremental KA together with retrospective modelling can be used to provide the user with a system that they can own, build and explore without the difficulties associated with capturing and validating the conceptual models of experts via the mediation of a knowledge engineer.     

Behavlets: a method for practical player modelling using psychology-based player traits and domain specific features

As player demographics broaden it has become important to understand variation in player types. Improved player models can help game designers create games that accommodate a range of playing styles, and may also facilitate the design of systems that detect the currently-expressed player type and adapt dynamically in real-time. Existing approaches can model players, but most focus on tracking and classifying behaviour based on simple functional metrics such as deaths, specific choices, player avatar attributes, and completion times. We describe a novel approach which seeks to leverage expert domain knowledge using a theoretical framework linking behaviour and game design patterns. The aim is to derive features of play from sequences of actions which are intrinsically informative about behaviour—which, because they are directly interpretable with respect to psychological theory of behaviour, we name ‘Behavlets’. We present the theoretical underpinning of this approach from research areas including psychology, temperament theory, player modelling, and game composition. The Behavlet creation process is described in detail; illustrated using a clone of the well-known game Pac-Man, with data gathered from 100 participants. A workshop-based evaluation study is also presented, where nine game design expert participants were briefed on the Behavlet concepts and requisite models, and then attempted to apply the method to games of the well-known first/third-person shooter genres, exemplified by ‘Gears of War’, (Microsoft). The participants found 139 Behavlet concepts mapping from behavioural preferences of the temperament types, to design patterns of the shooter genre games. We conclude that the Behavlet approach has significant promise, is complementary to existing methods and can improve theoretical validity of player models.     

Knowledge-based versus data-driven fuzzy habitat suitability models for river management

Aquatic habitat suitability models have increasingly received attention due to their wide management applications. Ecological expert knowledge has been frequently incorporated in such models to link environmental conditions to the quantitative habitat suitability of aquatic species. Since the formalisation of problem-specific human expert knowledge is often difficult and tedious, data-driven machine learning techniques may be helpful to extract knowledge from ecological datasets. In this paper, both expert knowledge-based and data-driven fuzzy habitat suitability models were developed and the performance of these models was compared. For the data-driven models, a hill-climbing optimisation algorithm was applied to derive ecological knowledge from the available data. Based on the available ecological expert knowledge and on biological samples from the Zwalm river basin (Belgium), habitat suitability models were generated for the mayfly Baetis rhodani (Pictet 1843). Data-driven models appeared to outperform expert knowledge-based models substantially, while a step-forward model selection procedure indicated that physical habitat variables adequately described the mayfly habitat suitability in the studied area. This study has important implications on the application of expert knowledge in ecological studies, especially if this knowledge is extrapolated to other areas. The results suggest that data-driven models can complement expert knowledge-based approaches and hence improve model reliability.