Learning the structure of Bayesian networks with ancestral and/or heuristic partition

Tan, Xiangyuan; Gao, Xiaoguang; Wang, Zidong; Han, Ho Jae; Liu, Xiaohan; Chen, Daqing

doi:10.1016/j.ins.2021.10.052

Cited by 15 publications

(3 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The steps for BNA have been described previously [37]. Briefly, the analysis involves (i) structure learning using the score-based hill-climbing (HC) learning algorithm, which determines the topological relationships between the nodes in the network and the sample data [38]; (ii) bootstrapping using 10,000 bootstrap samples to prune connections observed in less than 50% of the bootstrap samples [10,37,39,40]; and (iii) parameter learning using the Bayesian method to estimate the conditional probabilities between connected nodes from the identified network and the observed sample data [41].…”

Section: Bayesian Network Analysismentioning

confidence: 99%

Antimicrobial Susceptibility in Respiratory Pathogens and Farm and Animal Variables in Weaned California Dairy Heifers: Logistic Regression and Bayesian Network Analyses

Morgan Bustamante,

Chigerwe,

Martínez-López

et al. 2024

Antibiotics

View full text Add to dashboard Cite

Weaned dairy heifers are a relatively understudied production group. Bovine respiratory disease (BRD) is the most common cause of antimicrobial drug (AMD) use, morbidity, and mortality in this production group. The study of antimicrobial resistance (AMR) is complicated because many variables that may affect AMR are related. This study generates hypotheses regarding the farm- and animal-level variables (e.g., vaccination, lane cleaning, and AMD use practices) that may be associated with AMR in respiratory isolates from weaned dairy heifers. A cross-sectional study was performed using survey data and respiratory isolates (Pasteurella multocida, Mannheimia haemolytica, and Histophilus somni) collected from 341 weaned dairy heifers on six farms in California. Logistic regression and Bayesian network analyses were used to evaluate the associations between farm- and animal-level variables with minimum inhibitory concentration (MIC) classification of respiratory isolates against 11 AMDs. Farm-level variables associated with MIC classification of respiratory isolates included the number of source farms of a calf-rearing facility, whether the farm practiced onsite milking, the use of lagoon water for flush lane cleaning, and respiratory and pinkeye vaccination practices. Animal-level variables associated with a MIC classification included whether the calf was BRD-score-positive and time since the last phenicol treatment.

show abstract

Section: Bayesian Network Analysismentioning

confidence: 99%

Antimicrobial Susceptibility in Respiratory Pathogens and Farm and Animal Variables in Weaned California Dairy Heifers: Logistic Regression and Bayesian Network Analyses

Morgan Bustamante,

Chigerwe,

Martínez-López

et al. 2024

Antibiotics

View full text Add to dashboard Cite

show abstract

“…Note that this time limit applies to the graph search only; it does not include the pre-computation of parent set scores. Tan et al (2022) propose two variable partitioning approaches which they demonstrate can improve learning times in algorithms such as A*, often by orders of magnitude. The first heuristic is ancestral partitioning which assumes a partial ordering as illustrated in Fig.…”

Section: Exact Search Of Node-ordering Spacementioning

confidence: 99%

A survey of Bayesian Network structure learning

et al. 2023

View full text Add to dashboard Cite

Bayesian Networks (BNs) have become increasingly popular over the last few decades as a tool for reasoning under uncertainty in fields as diverse as medicine, biology, epidemiology, economics and the social sciences. This is especially true in real-world areas where we seek to answer complex questions based on hypothetical evidence to determine actions for intervention. However, determining the graphical structure of a BN remains a major challenge, especially when modelling a problem under causal assumptions. Solutions to this problem include the automated discovery of BN graphs from data, constructing them based on expert knowledge, or a combination of the two. This paper provides a comprehensive review of combinatoric algorithms proposed for learning BN structure from data, describing 74 algorithms including prototypical, well-established and state-of-the-art approaches. The basic approach of each algorithm is described in consistent terms, and the similarities and differences between them highlighted. Methods of evaluating algorithms and their comparative performance are discussed including the consistency of claims made in the literature. Approaches for dealing with data noise in real-world datasets and incorporating expert knowledge into the learning process are also covered.

show abstract

“…Typically, the BN structure is determined by establishing causal relationships between nodes. For complex systems with unclear internal mechanisms, it can be challenging to clarify these causal relationships, requiring the use of optimization algorithms for BN structure learning [28]. However, in the case of chillers, their thermodynamic principles are relatively well-defined, and the influence relationship between typical faults and features is generally understood.…”

Section: Bn Driven By the Fusion Of Residuals And Datamentioning

confidence: 99%

Fault Diagnosis Based on Fusion of Residuals and Data for Chillers

et al. 2023

View full text Add to dashboard Cite

Feature data refer to direct measurements of specific features, while feature residuals represent the deviations between these measurements and their corresponding benchmark values. Both types of information offer unique insights into the system’s behavior. However, conventional diagnostic systems often struggle to effectively integrate and utilize both types of information concurrently. To address this limitation and improve diagnostic performance, a hybrid method based on the Bayesian network (BN) is proposed. This method enables the parallel fusion of feature residuals and feature data within a unified diagnostic model, and a comprehensive framework for developing this hybrid method is also given. In the hybrid BN, the symptom layer consists of residual nodes representing feature residuals and data nodes representing measured feature data. By applying the proposed method to two chillers and comparing it with state-of-the-art existing methods, we demonstrate its effectiveness and superiority. The results highlight that the proposed method not only accommodates the absence of either type of information but also leverages both of them to enhance diagnostic performance. Compared to using a single type of node, the hybrid method achieves a maximum improvement of 24.5% in diagnostic accuracy, with significant enhancements in F-measure observed for refrigerant leakage fault (34.5%) and excessive lubricant fault (32.8%), respectively.

show abstract

Learning the structure of Bayesian networks with ancestral and/or heuristic partition

Cited by 15 publications

References 30 publications

Antimicrobial Susceptibility in Respiratory Pathogens and Farm and Animal Variables in Weaned California Dairy Heifers: Logistic Regression and Bayesian Network Analyses

Antimicrobial Susceptibility in Respiratory Pathogens and Farm and Animal Variables in Weaned California Dairy Heifers: Logistic Regression and Bayesian Network Analyses

A survey of Bayesian Network structure learning

Fault Diagnosis Based on Fusion of Residuals and Data for Chillers

Contact Info

Product

Resources

About