Search

Nowadays, airlines are facing increasingly fierce market competition while ushering in development opportunities. Many scholars researched on airline passenger value using data mining approaches, but the evaluation index of air passenger value in the existing research is based on internal data sources. It is of great importance to blend the external data from third-party under the premise of safe and legal data privacy disclosure to extend the characteristic dimension of their customers. Therefore, this research proposes a novel model that can blend multi-source big data to enrich airline passengers’ feature dimensions under the premise of ensuring passengers’ information privacy security, and establish the user profile of passengers for accurately identifying the high-value passengers. It is proved that our proposed novel model has better performance compared with the results of the traditional model that only use one party data in terms of Area Under Curve (AUC) and Kolmogorov-Smirnov (KS) value.

Automobile insurance fraud detection has become critically important for reducing the costs of insurance companies. The majority of insurance companies use expert knowledge to detect fraud. Experience-based knowledge are interpretable and re-usable but the simplistic way that this knowledge is used in practice, often leads to some degree of misjudgment. This paper aims to establish a unique Evidential Reasoning (ER) rule that combines independent evidence from both experience based indicators and probabilities of fraud obtained from historical data. Each piece of evidence is weighted and then combined conjunctively with the weights optimised using a maximum likelihood evidential reasoning (MAKER) framework for datadriven inferential modelling. Based on a real-world insurance claim dataset, our experimental results reveal that the proposed approach preserves the interpretability and usability of expert detection system, and anticipates the changes in fraud practices by tracking the trend of the weights of experience-based indicators. Furthermore, the experimental results show that the proposed approach outperforms a number of widely used machine learning models, such as logistic regression and random forests.

Traditional vital signs are an essential part of triage assessment in emergency departments (ED), and have been widely used in trauma prediction models. Previous researchers have studied the effect of vital signs scores on predicting traumatic injury outcomes and have found it to be significant. Based on the vital signs’ scores, an Interpretable Machine Learning (IML) method is proposed to predict patient outcomes and is compared with various ML algorithms. Results indicate that the IML method has a comparable performance with a mean AUC of 0.683, and its interpretability would help in the early identification of trauma patients at risk of mortality.

The Evidential Reasoning (ER) approach is an inference method to analyse multi-criteria decision-making problems with uncertainties. However, due to the pressure of time, limit of knowledge and lack of data, such problems need to be assessed using both human judgments and limited data. In this case, it is suitable to provide a belief structure by means of linguistic terms, which we refer to as linguistic belief structure. Considering that different decision-makers may have different perceptions, the semantics of human judgments can also be different. In this paper, we introduce three linguistic scale functions to capture the semantics of linguistic belief degrees for constructing a complete linguistic information fusion method. We then extend the framework of the ER approach to multi-expert multi-criteria decision making. The method proposed in this study can fill the research gap of the ER approach for multi-expert multi-criteria decision-making problems with linguistic belief structures.

Multiple criteria decision analysis (MCDA) methods have become increasingly popular in the performance assessment and decision making of renewable energy systems due to the multi-dimensionality and complexity of its technical, environmental, economic and social impacts. The aim of this paper is to develop a preliminary performance assessment model for the feasibility analysis of constructing different multi-vector decentralized renewable energy systems in an industrial park. Four micro-grid energy systems are identified as alternative solutions in terms of the local environmental conditions and energy profile of the industrial park, and these alternatives are evaluated against a set of performance criteria. The evidential reasoning approach with the implementation of Intelligent Decision System (IDS) is then applied to aggregate assessment information in order to obtain the overall ranking of alternative solutions. sensitivity and trade-off analysis is conducted to validate the robustness of the decision making process.

Data mining requires a pre-processing task where data are prepared, cleaned, integrated, transformed, reduced and discretized to ensure data quality. Incomplete data are commonly encountered during data cleaning, which can have major impact on the conclusions that will be drawn from the data. In order to effectively carry out inferential modelling or decision making from incomplete independent variables or explanatory variables and consider different types of uncertainties, this paper adopts a data-driven inferential modelling approach, Maximum Likelihood Evidential Reasoning (MAKER) framework, which takes advantage of incomplete datasets without any imputation that may be required by other conventional machine learning methods. The MAKER framework reflects the plausibility of different values of missing data and expresses data-driven support for different values of missing data.

Online collaboration is currently emerging as one of the predominant technology trends in the collaborative learning atmosphere. However, the scarcity of collaborative devices combined with task-specific groupware tools may limit its full potential. In addition, there is a deficiency of investigations in the form of integrated settings between social media and mobile groupware. Therefore, this paper proposes the approach of mobile groupware (MobileMeeting) that can be integrated (i.e., embedded) into the popular social media posts that students currently use as a supplementary tool for collaborative learning. An investigation was conducted in real project-based scenarios with students, followed by post-session questionnaires including a log-report analysis from Google Analytics as supporting evidence. This study confirms that the proposed approach using MobileMeeting is capable of improving collective learning productivity and activities, including their satisfaction with the required groupware tools. The log-report analysis of students' case projects (social media marketing campaigns) also displays an increment in user engagement and retention by bringing the new user and a large number of returning user. In conclusion, these results have several practical implications: as collaborative learning activities become more complex, the setup of integrated social media and mobile groupware (MobileMeeting) can support a range of collective actions.

In this paper, we introduce the main concepts of a new maximum livelihood evidential reasoning (MAKER) framework for data-driven inferential modelling and decision making under different types of uncertainty. It consists of two types of model: state space model (SSM) and evidence space model (ESM), driven by the data that reflects the relationships between system inputs and output. SSM is constructed to describe different system states and changes. ESM is established by mapping data to a set of evidence that is partitioned into evidential elements each pointing to a system state set and together represents system behaviours in a probabilistic and distributed manner. The reliability of evidence and interdependence between a pair of evidence are explicitly measured. It is in the joint evidence-state space that multiple pieces of evidence with different degrees of interdependence and reliability are acquired from system inputs and combined to inference system output. A general optimal learning model is constructed, where evidence reliability can be learnt from historical data by maximising the likelihood of true state. In the MAKER framework, different types of uncertainty can be taken into account for inferential modelling, probabilistic prediction and decision making.

The modelling for renewable energy system (RES) performance evaluation and impact analysis can be a challenging task, where there are multiple criteria with both quantitative and qualitative forms under uncertainty. Multi-criteria decision analysis (MCDA) methods have become increasingly popular in the decision-making for renewable energy systems because of the multi-dimensionality of the sustainability goal and the complexity of the technical, environmental, economic and social perspectives. This paper aims to review the applications of MCDA methods to the performance modelling and impact analysis of RESs, primarily in four relevant areas, including renewable energy planning and policy, renewable energy evaluation and assessment, renewable energy project selection and allocation and RES environmental impact assessment. Further research can be conducted to study the feasibility of different RESs and the selection of an appropriate MCDA methodology in alternative energy decision-making.

A study was carried out to compare an integrated HTML5-based real-time video conferencing solution against a separately installed video conferencing solution in the same web-based meeting support application. To carry out this comparison, WebRTC and Skype have been integrated into an existing groupware application. Our study shows that comparing to a separately installed, best of breed video conferencing solution; the web browser-based solution would be an adequate and more convenient one. Our work indicates that the cost to implement WebRTC is not large in comparison to the number of benefits that can be gained. The quality of video conferencing has been continuously improving, and the implementation cost is getting lower as better WebRTC libraries emerge and evolve.

The UK is one of these countries in Europe, which have the highest death rate from asthma, and the rules to identify asthma control steps are vague in the current official guideline of asthma management. In this research, diagnosis rules on asthma control steps are developed to supplement the current guideline and to assist patients to monitor and manage their asthma on daily basis. The main challenge of developing the rules is missing values. Although the data examined have prodigious volumes of records for patients, no one have all and different patients have different information recorded. The large proportion of missing values lead to comparatively limited powers of some techniques like Decision Tree Analysis, Logistic Regression, ANN, Bayes' Rule and SVM. This research explores the Evidential Reasoning (ER) rule to develop prognostic rules for asthma control steps. ER is prior-free probabilistic inference and has not been applied to disease diagnosing and monitoring. The results are represented as probability distributions on asthma control steps given any combination of evidence, even if some combinations are not recorded in the current database. In practice, it could help clinicians to identify asthma control steps of patients, prescribe corresponding treatments, and monitor the effectiveness of the treatment and the progress of patients in asthma control management.

Track irregularity fault commonly leads to the abnormal vibration of train which is the main cause of poor ride and even derailment. Based on Evidence Reasoning (ER) rule, this paper presents a fault identification method for determining the dynamic management levels of track irregularity using sample data measured from the accelerometers mounted in the axle-box and car-body of in-service train. Firstly, statistical approach of the sample casting is given to generate diagnosis evidence. Secondly, the ER rule is used to combine the diagnosis evidence coming from two accelerometers. The estimated irregularity displacement can obtained from the combined results. Thirdly, the dynamic levels of track irregularity can be identified according to the estimated displacement and the corresponding management standard. Finally, a representative experiment in Chinese railway line shows the superior identification accuracy of the proposed method by comparing with classical neural network-based approximating methodology.

In this paper, a Genetic Algorithm (GA) search heuristic is proposed for validating the model-structure of Belief Rule-Based (BRB) methodologies. In order to ensure the balance between the model fit/ accuracy and the model complexity, the Akaike Information Criterion (AIC) is used in conjunction with the mentioned heuristic. The resulting framework is tested, using a model consisting of 3 inputs and one output, each of the 4 variables being allocated up to 5 referential values. The presented results illustrate the time-efficiency of the GA heuristic, as well as the penalty imposed by AIC on the number of parameters. The simplest model structure is indicated by AIC to be the optimal one. However, three additional model structures have been found to have AIC values which are moderately close to this optimum. An analysis of their coefficients of determination indicates a higher fit (than AIC optimum) on both testing sets and overall.

In this paper, we present a generic expert system software tool and its application in expert knowledge modelling and inference. The system is able to model expert rule-based knowledge and provide predictions for unseen data by the application of inductive inference. The theoretical framework for modelling expert knowledge is described independently of its implementation as a generic expert system. The framework is based on a review of the literature on belief rule-based inference methodology using the evidential reasoning approach (RIMER) and its applications. The application of the generic expert system is demonstrated by a case study which uses a guideline model for clinical risk assessment of acute upper gastrointestinal bleeding.

The supply represents the available number of products of a specific variable. This paper presents a new approach for studying the supply through the use of aggregation systems. Its main advantage is that it permits to represent the information in a more complete way. Thus, it is possible to develop forecasting methods that consider optimistic and pessimistic scenarios and those that are more expected to occur. A wide range of aggregation operators are used giving special focus on the ordered weighted average (OWA). Several generalizations that use it with the weighted average and the probability are presented including the OWA weighted average (OWAWA) and the probabilistic OWAWA (POWAWA) operator. A numerical example is also presented.

Increases in usage of websites as communication tools has lead to reductions in costs of product and service delivery for companies. Incorrect design of websites leads to redesigning them to fit customer needs. Redesign costs are particularly relevant for resource constrained SMEs. Modeling highly subjective, uncertain and repetitive decisions requires methods that include the benefits of both expert knowledge as well as machine learning. This paper reports a comparison on the application of two such hybrid approaches: Belief Rule-Base Inference Methodology Using the Evidential Reasoning Approach (RIMER) and Artificial Neural Fuzzy Inference System (ANFIS). A numerical example and a case study are reported in order to analyze the potential of the two methods. The results show that both methods can be used to accurately represent and predict the complicated and nonlinear relationships between website characteristics and overall satisfaction. The RIMER approach is found to outperform ANFIS when considering explanatory power.

Coordination between two different business entities is an important way to gain competitive advantage. The collaboration and coordination between upstream and downstream in the supply chain is the key to gain for the whole supply chain system. This paper analyzes the relationship of the cost and the profit between supplier and buyer in supply chain, establishes general cost-profit model in a two-period supply chain, and obtains non-coordinated model without outsourcing under shortage and coordinated model with outsourcing under shortage. The paper also illustrates the feasibility of the model through a numerical example.

The evidential reasoning (ER) approach was developed to support multiple criteria decision analysis (MCDA). It is based on the Dampster's combination rule for criteria aggregation and belief function for treating ignorance. In the original ER approach, however, alternative ranking depends on the accurate estimation of a value function, which may be difficult in certain decision environments. In this paper, the link and difference between the ER algorithm and Dampster's combination rule are analysed first. A new alternative ranking method is then investigated as an integrated part of the enhanced ER approach.

This paper illustrates how efficiency, effectiveness, consistency and reliability of nuclear safety culture assessment can be improved by the Intelligent Decision System (IDS) tool. It describes how IDS can be used as both a knowledge management and assessment supporting tool for nuclear safety culture assessment using the IAEA Safety Culture Assessment Review Team (SCART) guidelines. Once the assessment knowledge is built into IDS, the tool provides support at the data gathering stage by retrieving the right guidance at when and where it is needed so that the assessment can be made more efficiently, consistently and objectively. At the Discussion and closing stage of the assessment, the tool generates aggregated assessment results for both a team of and individual reviewers in text and graphical formats. The text report highlights the strengths and the areas for improvement, including advice on how to improve and examples of best practices in the areas. The various graphical interfaces and charts provide side by side comparisons, panoramic view of the strengths and weaknesses, trends and profiles of assessment outcomes made by different reviewers, for different organisations or for the same organisation at different time. Those outcomes give a speedy and accurate account of the review so that discussions and feedback can be made more productively and constructively.

In this paper, it is described how multiple criteria decision analysis (MCDA) methods, in particular the Evidential Reasoning (ER) approach, is applied to help Tesco, the largest UK retailer, to prioritise product groups for its carbon labelling program. The main objectives of the program are to maximise the positive impact of the program to the environment in terms of carbon footprint reduction, while not to introduce unintentionally non-carbon related risks such as resource depletion, pollution and ethical risks. The application is focused on comparing both the positive and negative impacts of labelling different product groups so that the ones with the relatively higher positive impacts are recommended for early participation in the program. The main challenges of the application are uncertainties in data and judgements, such as lack of data, inaccuracy of data estimates and weights of different criteria. It is demonstrated with examples how those challenges can be dealt with by applying the ER approach for MCDA.

This paper aims to contribute to offshore risk analysis by proposing a methodology to model causal relationships focusing on human and organisational errors. The proposed methodology uses James Reason's "Swiss cheese" model to form a generic offshore risk analysis framework and tailors Bayesian Network (BN) to fit into the framework to construct a causal relationship model. A case study of the collision risk between a Floating Production, Storage and Offloading (FPSO) unit and authorised vessels caused by human and organisational factors (HOFs) during operations is used to illustrate the application of the proposed methodology. This paper shows that James Reason's "Swiss cheese" model and BN can be jointly used in offshore risk analysis