Traditional Bayesian Network Research Articles

A Maximum Entropy Multisource Information Fusion Method to Evaluate the MTBF of Low-Voltage Switchgear

When analyzing the reliability of low-voltage switchgear by Bayesian method, the maximum entropy multisource information fusion method was proposed to obtain the prior information of low-voltage switchgear and then evaluate the reliability. The historical data of low-voltage switchgear was collected and organized from a manufacturer. According to the expert experience and the data, the creditability analysis and the compatibility test were presented by the Smirnov test method. Based on the high creditability and compatibility, the result of the maximum entropy multisource information fusion method is the determination of prior information. Therefore, the distribution type of the prior information was confirmed by using the maximum entropy method, and the parameter of the prior information was received by bootstrap method with MATLAB. Then the posterior distribution was obtained to evaluate the MTBF of low-voltage switchgear. Finally, the historical data of years from 2007 to 2010 was taken as prior information to illustrate the maximum entropy multisource information fusion method and to get the MTBF of low-voltage switchgear. The evaluation result reduces the experimental period and test cost, which is an improvement for the reliability evaluation and management of low-voltage switchgear and also an improvement for other systems with simple sample data. Compared with traditional Bayesian networks, the proposed method can fuse experts experience and historical data and has advantages for the use of prior information effectively.

Process Monitoring Based on Multivariate Causality Analysis and Probability Inference

System security is one of the key challenges of the cyber-physical systems. Bayesian approach can estimate and predict the potentially harmful factors of the general system, but it has many limitations that can lead to undesirable effects in the complex systems. This paper presents a new modeling and monitoring framework to avoid the traditional Bayesian network disadvantage. A multivariate causal analysis method is proposed to establish a compact system structure. Combined with network parameter learning, we constructed a corresponding multivariate alarm predict graph model, in which the qualitative and quantitative relationships among the process variables are revealed distinctly. Then this model is used to accurately predict the future possible alarm events via the probability inference. Similarly, it also can be used to detect faults and find the source of the fault. The effectiveness of the proposed method is verified in public data sets and the Tenessee Eastman process. Simulation results show that the established causal relationship is completely consistent with the actual mechanism, and the alarm state of the critical variable is accurately predicted.

Addressing the epistemic uncertainty in maritime accidents modelling using Bayesian network with interval probabilities

Bayesian Network (BN) is often criticized for demanding a large number of crisp/exact/precise conditional probability numbers which, due to the lack of statistics, have to be obtained through experts’ judgment. These exact probability numbers provided by the experts often carry a high level of epistemic uncertainty due to the incompleteness of human knowledge, not to mention the hardness in obtaining them in the first place. The existence of uncertainty in risk modelling was well recognized but seldom discussed. This paper explores the extension of BN with interval probabilities to the modelling of maritime accidents, which allows for the quantification of the epistemic uncertainty. Ship collision is chosen for case study for the strategic importance of navigational safety. The user friendly linguistic terms defined with interval scales were used for elicitation of interval conditional probabilities from industry experts. Inferences were made directly with the interval probabilities with the GL2U algorithm. Meanwhile, the interval probabilities were converted into point probabilities and computed with the traditional BN method for comparison, which were all shown to be within the ranges of the calculated posterior intervals probability. Results with inputs from different experts reveal discrepancies, which in turn verify the existence of uncertainty in risk modelling. A discussion was also provided on how the uncertainty in risk assessment propagates to the decision making process and influences the ranking of potential risk control options.

Inference of time-delayed gene regulatory networks based on dynamic Bayesian network hybrid learning method.

Gene regulatory networks (GRNs) research reveals complex life phenomena from the perspective of gene interaction, which is an important research field in systems biology. Traditional Bayesian networks have a high computational complexity, and the network structure scoring model has a single feature. Information-based approaches cannot identify the direction of regulation. In order to make up for the shortcomings of the above methods, this paper presents a novel hybrid learning method (DBNCS) based on dynamic Bayesian network (DBN) to construct the multiple time-delayed GRNs for the first time, combining the comprehensive score (CS) with the DBN model. DBNCS algorithm first uses CMI2NI (conditional mutual inclusive information-based network inference) algorithm for network structure profiles learning, namely the construction of search space. Then the redundant regulations are removed by using the recursive optimization algorithm (RO), thereby reduce the false positive rate. Secondly, the network structure profiles are decomposed into a set of cliques without loss, which can significantly reduce the computational complexity. Finally, DBN model is used to identify the direction of gene regulation within the cliques and search for the optimal network structure. The performance of DBNCS algorithm is evaluated by the benchmark GRN datasets from DREAM challenge as well as the SOS DNA repair network in Escherichia coli, and compared with other state-of-the-art methods. The experimental results show the rationality of the algorithm design and the outstanding performance of the GRNs.

An open toolbox for the reduction, inference computation and sensitivity analysis of Credal Networks

Bayesian Networks are a flexible and intuitive tool associated with a robust mathematical background. They have attracted increasing interest in a large variety of applications in different fields. In spite of this, inference in traditional Bayesian Networks is generally limited to only discrete variables or to probabilistic distributions (adopting approximate inference algorithms) that cannot fully capture the epistemic imprecision of the data available. In order to overcome these limitations, Credal Networks have been proposed to integrate Bayesian Networks with imprecise probabilities which, adopting non-probabilistic or hybrid models, allow to fully represent the information available and its uncertainty.Here, a novel computational tool, implemented in the general purpose software OpenCossan, is proposed. The tool provides the reduction of Credal Networks through the use of structural reliability methods, in order to limit the cost associated with the inference computation without impoverishing the quality of the information initially introduced. Novel algorithms for the inference computation of networks involving probability bounds are provided. In addition, a novel sensitivity approach is proposed and implemented into the Toolbox in order to identify the maximum tolerable uncertainty associated with the inputs.

Predictive Abnormal Events Analysis Using Continuous Bayesian Network

The reliable prediction and diagnosis of abnormal events provide much needed guidance for risk management. The traditional Bayesian network (traditional BN) has been used to dynamically predict and diagnose abnormal events. However, its inherent limitation caused by discrete categorization of random variables degrades the assessment reliability. This paper applied a continuous Bayesian network (CBN)-based model to reduce the above-mentioned limitation. To compute complex posterior distributions of CBN, the Markov chain Monte Carlo method (MCMC) was used. A case study was conducted to demonstrate the application of CBN, based on which a comparative analysis of the traditional BN and CBN was presented. This work highlights that the use of CBN can overcome the drawbacks of traditional BN to make dynamic prediction and diagnosis analysis more reliable.

Lifecycle cost risk analysis for infrastructure projects with modified Bayesian networks

PurposePrevious research has little specific guidance on how to improve large infrastructures’ risk analysis. This paper aims to propose a practical risk analysis framework across the project lifecycle with Bayesian Networks (BNs).Design/methodology/approachThe framework includes three phases. In the qualitative phase, primary risks were identified by literature reviews and interviews; questionnaires were used to determine key risks at each project stage and causal relationships between stage-related risks. In the quantitation, brainstorming and questionnaires, and techniques of ranked nodes/paths, risk map and Bayesian truth serum were adopted. Then, a BN-based risk assessment model was developed, and risk analysis was conducted with AgenaRisk software.FindingsTwenty key risks across the lifecycle were determined: some risks were recurring and different risks emerged at various stages with the construction and feasibility most risky. Results showed that previous stages’ risks significantly amplified subsequent stages’ risks. Based on the causality of stage-related risks, a qualitative model was easily constructed. Ranked nodes/paths facilitated the quantification by requiring less statistical knowledge and fewer parameters than traditional BNs. As articulated by a case, this model yielded very simple and easy-to-understand representations of risks and risk propagation pathways.Originality/valueRare research has developed a BN risk assessment model from the perspective of project stages. A structured model, a propagation network among individual risks, stage-related risks, and the final adverse consequence, has been designed. This research provides practitioners with a realistic risk assessment approach and further understanding of dynamic and stage-related risks throughout large infrastructures’ lifecycle. The framework can be modified and used in other real-world risk analysis where risks are complex and develop in stages.

Research on Fault Diagnosis for Pumping Station Based on T-S Fuzzy Fault Tree and Bayesian Network

According to the characteristics of fault diagnosis for pumping station, such as the complex structure, multiple mappings, and numerous uncertainties, a new approach combining T-S fuzzy gate fault tree and Bayesian network (BN) is proposed. On the one hand, traditional fault tree method needs the logical relationship between events and probability value of events and can only represent the events with two states. T-S fuzzy gate fault tree method can solve these disadvantages but still has weaknesses in complex reasoning and only one-way reasoning. On the other hand, the BN is suitable for fault diagnosis of pumping station because of its powerful ability to deal with uncertain information. However, it is difficult to determine the structure and conditional probability tables of the BN. Therefore, the proposed method integrates the advantages of the two methods. Finally, the feasibility of the method is verified through a fault diagnosis model of the rotor in the pumping unit, the accuracy of the method is verified by comparing with the methods based on traditional Bayesian network and BP neural network, respectively, when the historical data is sufficient, and the results are more superior to the above two when the historical data is insufficient.

Multivariate probabilistic safety analysis of process facilities using the Copula Bayesian Network model

Integrated safety analysis of hazardous process facilities calls for an understanding of both stochastic and topological dependencies, going beyond traditional Bayesian Network (BN) analysis to study cause-effect relationships among major risk factors. This paper presents a novel model based on the Copula Bayesian Network (CBN) for multivariate safety analysis of process systems. The innovation of the proposed CBN model is in integrating the advantage of copula functions in modelling complex dependence structures with the cause-effect relationship reasoning of process variables using BNs. This offers a great flexibility in probabilistic analysis of individual risk factors while considering their uncertainty and stochastic dependence. Methods based on maximum likelihood evaluation and information theory are presented to learn the structure of CBN models. The superior performance of the CBN model and its advantages compared to traditional BN models are demonstrated by application to an offshore managed pressure drilling case study.

A novel privacy preserving based ensemble cross defect prediction model for decision making

Summary In recent years, defect prediction and severity assessment have been successfully applied in software defects and metrics prediction in business applications. Providing essential security to the software metrics and decision patterns are the two main issues in the traditional business models for inter and intra communication mechanisms. Traditional software prediction and decision pattern models are important to the business analysts for decision making and market analysis. But these models could not provide enough privacy to the business models for secure transmission of the decision patterns. In this proposed model, a new privacy preserving based defect prediction classification model was implemented on multiple associated products to predict metric relationship, along with defects patterns. Experimental results show that proposed model has high true positive rate compared to traditional Bayesian network and privacy preserving models. Also, this model generates a high privacy preserved decision patterns on various business software applications for secure communication.

A Survey of Bayesian Network Models for Decision Making System in Software Engineering

Defect prediction and assessment are the essential steps in large organizations and industries where the software complexity is growing exponentially. A large number of software metrics are discovered and used for metric prediction in the literature. Bayesian networks are applied to find the probabilistic relationships among the software metrics in different phases of software life cycle. Defects in a software project lead to minimize the quality which might be the impact on the overall defect correction. Traditional Bayesian networks are system dependable and their models are invariant towards the accurate computation. Bayesian network model is used to predict the defect correction at various levels of the software development. This model reveals the high potential software efforts and metrics required to minimize the overall cost of the organization for decision support.

A Machine Intelligence Designed Bayesian Network Applied to Alzheimer's Detection Using Demographics and Speech Data

Bayesian networks (BNs) have classically been designed by two methods: expert approach (ask an expert for nodes and links) and data driven approach (infer them from data). An unexpected by-product of previous Alzheimer's / dementia research (presented at CAS2015) was yet another approach where the results of a hybrid design were used to configure a BN. A complex adaptive systems approach, (e.g. GA-SVM-oracle hybrid) can sift through the combinatorics of feature subset selection, yielding a modest set of only the most influential features. Then using known likelihoods of demographics associated to dementia, and assuming direct and independent influence of dementia upon speech features, the BN is specified. The conditional probabilities needed can be estimated with far fewer data than the traditional BN data-driven approach. Although BNs have advantages (intuitive interpretation and graceful handling of missing data) they also have challenges. We report initial implementation results that suggest the need to reduce continuous variables to discrete categories, and the still-remaining need to estimate a substantial number of conditional probabilities, remain challenges for BNs. We suggest some ways forward in the application of BNs with the objective of improving / refining Alzheimer's / dementia detection using speech.

The role of local partial independence in learning of Bayesian networks

Bayesian networks are one of the most widely used tools for modeling multivariate systems. It has been demonstrated that more expressive models, which can capture additional structure in each conditional probability table (CPT), may enjoy improved predictive performance over traditional Bayesian networks despite having fewer parameters. Here we investigate this phenomenon for models of various degree of expressiveness on both extensive synthetic and real data. To characterize the regularities within CPTs in terms of independence relations, we introduce the notion of partial conditional independence (PCI) as a generalization of the well-known concept of context-specific independence (CSI). To model the structure of the CPTs, we use different graph-based representations which are convenient from a learning perspective. In addition to the previously studied decision trees and graphs, we introduce the concept of PCI-trees as a natural extension of the CSI-based trees. To identify plausible models we use the Bayesian score in combination with a greedy search algorithm. A comparison against ordinary Bayesian networks shows that models with local structures in general enjoy parametric sparsity and improved out-of-sample predictive performance, however, often it is necessary to regulate the model fit with an appropriate model structure prior to avoid overfitting in the learning process. The tree structures, in particular, lead to high quality models and suggest considerable potential for further exploration.

SLAM Bayesian Network Model For Robot Behavior

In this paper, a bayesian framework for fault detection and isolation (FDI) based on kalman filtering is developed. Furthermore, in order to detect the faults affecting on the covariance matrix of the kalman filter, a real-time approach is presented. This proposed framework extracts the proper behavior of a mobile robot besides simultaneously localization and map building (SLAM). Actually the framework is a combination of the kalman filter and bayesian networks. Learning the model of the world is difficult. In particular when the system dynamics become nondeterministic, all aspects of the system cannot be directly observed and the sensors are subjected to noise. In many situations, learning a complete model is not possible. Therefore, only probabilistic models which are capable of taking uncertainty of sensors and environment can be employed. In this paper, we describe a framework as a composition between model-free and model-based systems. Model learning is perfectly based on bayesian network (BN) and fault detection is done by kalman filter. Experimental results show that the learned model outperforms the traditional BN. We demonstrate how the resulting algorithm can be used to detect faults in a complex system. Proposed method is not very sensitive to changing the map of robot. However, the Bayesian network and dynamic Bayesian network are very sensitive to changing the map and in the presence of the fault. The proposed method is tested in a real home environment with a mobile robot. Keywords: Bayesian network, behavior, fault detection, kalman filter, mobile robot, SLAM.

Transfer learning for temporal nodes Bayesian networks

Traditional machine learning algorithms depend heavily on the assumption that there is sufficient data to learn a reliable model. This is not always the case, and in situations where data is limited, transfer learning can be applied to compensate for the lack of information by learning from several sources. In this work, we present a novel methodology for inducing a Temporal Nodes Bayesian Network (TNBN) when training data is scarce by applying a transfer learning strategy. A TNBN is a probabilistic graphical model that offers a compact representation for dynamic domains by defining multiple time intervals in which events can occur. Learning a TNBN poses additional challenges to learning traditional Bayesian networks due to the incorporation of time intervals. Our proposal incorporates novel approaches to transfer knowledge from several TNBNs to learn the structure, parameters and intervals of a target TNBN. To evaluate our algorithm, we performed experiments with a synthetic network, where we created auxiliary models by altering the structure, parameters and temporal intervals of the original model. Results show that the proposed algorithm is capable of retrieving a reliable model even when few records are available for the target domain. We also performed experiments with a real-world data set belonging to the medical domain of HIV, where we were able to learn some documented mutational pathways and their temporal relations by applying transfer learning.

A Bayesian network intrusion detection algorithm based on principal component analysis and sliding window

In traditional Bayesian network (BN) intrusion detection, it is not taken into account that the dataset has an excessive number of attributes, which leads to an excessive calculation in the process of BN structure and thus greatly affects the detection efficiency. In addition, traditional BN intrusion does not consider attacks in the detection process; instead, it just simply relies on fixed BN to test the new dataset, which has a certain impact on detection accuracy. To solve these two problems, a new BN intrusion detection technology based on principal component analysis (PCA) and sliding window is introduced in this paper. The new algorithm reduces data dimensionality and uses the detected data to update the training dataset, which relatively completely reflects the overall status of the system. The experiments show that the improved algorithm can greatly reduce the computation cost and improve the detection accuracy.

Learning Bayesian network structure: Towards the essential graph by integer linear programming tools

The basic idea of the geometric approach to learning a Bayesian network (BN) structure is to represent every BN structure by a certain vector. If the vector representative is chosen properly, it allows one to re-formulate the task of finding the global maximum of a score over BN structures as an integer linear programming (ILP) problem. Such a suitable zero-one vector representative is the characteristic imset, introduced by Studený, Hemmecke and Lindner in 2010, in the proceedings of the 5th PGM workshop. In this paper, extensions of characteristic imsets are considered which additionally encode chain graphs without flags equivalent to acyclic directed graphs. The main contribution is a polyhedral description of the respective domain of the ILP problem, that is, by means of a set of linear inequalities. This theoretical result opens the way to the application of ILP software packages. The advantage of our approach is that, as a by-product of the ILP optimization procedure, one may get the essential graph, which is a traditional graphical BN representative. We also describe some computational experiments based on this idea.

BAYESIAN RETRIEVAL USING A SIMILARITY-BASED LEMMATIZER

The present paper describes a Bayesian network approach to Information Retrieval (IR) from Web documents. The network structure provides an intuitive representation of uncertainty relationships and the embedded conditional probability table is used by inference algorithms in an attempt to identify documents that are relevant to the user's needs, expressed in the form of Boolean queries. Our research has been directed in constructing a probabilistic IR framework that focus on assisting users to perform Ad-hoc retrieval of documents from the various domains such as economics, news, sports, etc. Furthermore, users can integrate feedback regarding the relevance of the retrieved documents in an attempt to improve performance on upcoming requests. Towards these goals, we have expanded the traditional Bayesian network IR system and tested it on several Greek web corpora on different application domains. We have developed two different approaches with regards to the structure: a simple one, where the structure is manually provided, and an automated one, where data mining is used in order to extract the network's structure. Results have depicted competitive performance against successful IR models of different theoretical backgrounds, such as the vector space utilizing tf-idf and the probabilistic model of BM25 in terms of precision-recall curves. In order to further improve the performance of the IR system, we have implemented a novel similarity-based lemmatization framework, reducing thus the ambiguity posed by the plethora of morphological variations of the languages in question. The employed lemmatization framework comprises of 3 core components (i.e. the word segregation, the data cleansing and the lemmatization modules) and is language-independent (i.e. can be applied to other languages with morphological peculiarities and thus improve Ad-hoc retrieval) since it achieves the mapping of an input word to its normalized form by employing two state-of-the-art language independent distance metric models, meaning the Levenshtein Edit distance and the Dice coefficient similarity measure, combined with a language model describing the most frequent inflectional suffixes of the examined language. Experimental results support our claim on the significance of this incorporation to Greek texts web retrieval as results improve by a factor of 4% to 11%.

A hierarchical naive Bayesian network classifier embedded GMM for textural image

To address several problems in high-dimensional textures feature space and the deficiencies of the single Gaussian distribution for remote sensing data, this paper proposes a hierarchical naive Bayesian network classifier embedded in a Gaussian mixture model for high-dimensional textural image classification. High-dimensional features are grouped by the model on the basis of the correlations between them. In this way, the high-dimensional problem is decomposed into multiple problems of lower dimension. At the same time, for each group of features, a Gaussian mixture model is applied to simulate the data distribution in feature space for land covers, which fits the “original” data distribution better than a single Gaussian model. The Gaussian mixture model is embedded as a child node into a naive Bayesian network, and then the final classification result is obtained within the naive Bayesian network classifier framework. Experimental results for the classification of Landsat ETM+ and QuickBird image textures demonstrated that the classification accuracy of this method is better than that of a traditional Bayesian network classifier and some other classical classifiers. Comparing with the method dealing with original high-dimensional features, it is also more efficiency and effectiveness with fewer demand of sample size and lower time complexity.

BAYESIAN NETWORK REASONING WITH UNCERTAIN EVIDENCES

This paper investigates the problem of belief update in Bayesian networks (BN) with uncertain evidence. Two types of uncertain evidences are identified: virtual evidence (reflecting the uncertainty one has about a reported observation) and soft evidence (reflecting the uncertainty of an event one observes). Each of the two types of evidence has its own characteristics and obeys a belief update rule that is different from hard evidence, and different from each other. The particular emphasis is on belief update with multiple uncertain evidences. Efficient algorithms for BN reasoning with consistent and inconsistent uncertain evidences are developed, and their convergences analyzed. These algorithms can be seen as combining the techniques of traditional BN reasoning, Pearl's virtual evidence method, Jeffrey's rule, and the iterative proportional fitting procedure.