Article
Mathematics, Applied
Duc-Lam Duong, Tapio Helin, Jose Rodrigo Rojo-Garcia
Summary: We study the stability properties of the expected utility function in Bayesian optimal experimental design. We provide a framework for this problem in a non-parametric setting and prove the convergence rate of the expected utility with respect to a likelihood perturbation. The assumptions set out for the general case are satisfied in the specific case of non-linear Bayesian inverse problems with Gaussian likelihood, and the stability of the expected utility with respect to perturbations to the observation map is regained. Numerical simulations are used to demonstrate the theoretical convergence rates in three different examples.
Article
Mathematics, Applied
Keyi Wu, Thomas O'Leary-Roseberry, Peng Chen, Omar Ghattas
Summary: This paper proposes a method for solving large-scale Bayesian optimal experimental design problems with partial differential equations as constraints and infinite-dimensional parameter fields. By replacing the PDE solve with a derivative-informed projected neural network, the problem is simplified to a process solved by a greedy algorithm, avoiding computationally intensive calculations.
JOURNAL OF SCIENTIFIC COMPUTING
(2023)
Article
Environmental Sciences
Marc Stutter, Samia Richards, Adekunle Ibiyemi, Helen Watson
Summary: This study aimed to evaluate the spatio-temporal processes within mesoscale catchment rivers to understand internal phosphorus loading mechanisms. The research found that heterogeneous sediment residence within the channel plays a vital role in sediment-water phosphorus exchange, influencing the overall pollution status.
SCIENCE OF THE TOTAL ENVIRONMENT
(2021)
Article
Environmental Sciences
Ken Aho, Dewayne Derryberry, Sarah E. Godsey, Rob Ramos, Sara R. Warix, Samuel Zipper
Summary: Non-perennial streams have gained increasing attention from researchers, but there is a lack of suitable methods for measuring their hydrologic connectivity. In this study, the authors developed Bayesian statistical approaches to measure average active stream length and a new metric called average communication distance. They applied these methods to Murphy Creek in Idaho, USA and found significant increases in effective stream lengths due to flow rarity, as well as seasonal differences in both average stream length and average communication distance. The study highlights the unique perspectives provided by communication distance and demonstrates the usefulness of Bayesian approaches in analyzing non-perennial streams.
WATER RESOURCES RESEARCH
(2023)
Article
Engineering, Marine
Jisu Lim, Minjoo Choi, Seungjae Lee
Summary: Dynamic analysis is powerful but time-consuming for mooring system design. In this study, we proposed a fast convergence Bayesian optimization algorithm (BOA) that updated the objective function as more data points were obtained. Compared with genetic algorithm (GA), which used a pre-trained surrogate model, BOA achieved a 50% reduction in maximum tension for an initial mooring system. However, GA required 20 times more computation time due to the training of the surrogate model.
JOURNAL OF MARINE SCIENCE AND ENGINEERING
(2023)
Article
Engineering, Environmental
Jip de Vries, Michiel H. S. Kraak, Richard A. Skeffington, Andrew J. Wade, Piet F. M. Verdonschot
Summary: Aquatic ecosystems are impacted by various environmental stressors, and it is crucial to understand how ecosystems respond to stressors and their combined effects on ecological status. Bayesian Networks are used to simulate stream macroinvertebrates' responses to multiple stressors, providing a promising avenue for scenario analyses in restoration management.
Article
Automation & Control Systems
Di Wu, G. Gary Wang
Summary: To reduce computational costs in engineering design, a novel metamodel called causal artificial neural network (causal-ANN) is developed in this paper, which leverages cause-effect relations and intermediate variables to train sub-networks and improve accuracy. By utilizing the structure of the causal-ANN and Bayesian Networks theory, attractive design subspaces can be identified.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2021)
Article
Ergonomics
Yueng-hsiang Huang, Yimin He, Jin Lee, Changya Hu
Summary: The study identifies key drivers of safety climate from the perspective of leader-member exchange (LMX) and uses Bayesian Network simulations to predict the most effective strategies for improving safety climate in the trucking industry. Results show that supervisory integrity and LMX have the strongest independent effects on safety climate, with joint strategies involving LMX and psychological ownership being the most effective for promoting organizational safety climate. Enhancing leaders' communication skills, encouraging commitment from employees/leaders, and providing more autonomy to employees are key strategies for improving safety climate in the trucking industry.
ACCIDENT ANALYSIS AND PREVENTION
(2021)
Article
Engineering, Multidisciplinary
Dustin Taylor, Steven E. Rigdon, Rong Pan, Douglas C. Montgomery
Summary: This paper explores the use of Bayesian approach in selecting appropriate D-optimal designs in various lifetime regression models, highlighting the importance of this method.
QUALITY AND RELIABILITY ENGINEERING INTERNATIONAL
(2022)
Article
Medicine, General & Internal
Yvan Jamilloux, Nicolas Romain-Scelle, Muriel Rabilloud, Coralie Morel, Laurent Kodjikian, Delphine Maucort-Boulch, Philip Bielefeld, Pascal Seve
Summary: The study aimed to implement and validate a Bayesian belief network algorithm for the differential diagnosis of the most relevant causes of uveitis. The algorithm, based on simple epidemiological characteristics and anatomoclinical features of uveitis, achieved high accuracy in both the training and test datasets.
JOURNAL OF CLINICAL MEDICINE
(2021)
Article
Computer Science, Artificial Intelligence
Auwal Haruna, Pingyu Jiang
Summary: The rapid development of Additive Manufacturing (AM) has brought many advantages to manufacturing end-use products and components, but it also faces limitations that need to be addressed through the concept of design for AM (DFAM) to reform it as a mainstream manufacturing method. This paper proposes a framework based on the Fuzzy Bayesian Network (FBN) for DFAM decision-making and investigates the potential adaptability of DFAM using 20 impact factors.
ADVANCED ENGINEERING INFORMATICS
(2022)
Article
Physics, Multidisciplinary
Chen Zhou, Haiyan Guo, Shujuan Cao
Summary: A gene network associated with Alzheimer's disease is constructed from multiple data sources, divided into modules and evaluated using different methods. Functional modules are identified through enrichment analysis, and essential genes are predicted using network topology properties and a logical regression algorithm under a Bayesian framework. Based on network pharmacology, potential herbs and herb compounds for AD are selected through visualization and enrichment analysis.
Article
Multidisciplinary Sciences
William C. Thompson
Summary: This article uses signal detection theory and Bayesian network model to analyze the impact of examiners' decision thresholds on the value of forensic pattern-matching evidence. The study found that small shifts in decision thresholds can dramatically affect the value and utility of such evidence in the legal system.
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA
(2023)
Article
Computer Science, Hardware & Architecture
Yexiao He, Xiaoning Zhang, Zixiang Xia, Yutao Liu, Keshav Sood, Shui Yu
Summary: NFV is a new approach to meet diverse demands of network services by decoupling network functions and devices. This study focuses on joint optimization of service chain graph design and mapping in NFV networks to improve network performance by minimizing maximum link load factor. The proposed algorithm effectively reduces the maximum link load factor according to extensive simulation results.
Article
Computer Science, Hardware & Architecture
Man Ding, Lingying Zhao, Mingyu Sun, Haocheng Qin
Summary: Most current studies on product emotional design (PED) lack consideration of the combined effect of all product features on emotion image and the relationship between the whole and the part. A new methodology based on ISM-BN-GA is proposed, which includes developing a cognitive model and constructing a hierarchical directed network of PED system elements. The methodology utilizes a genetic algorithm to search for optimal design schemes and has been validated through a handheld camera design, showing feasibility and practicality in the field of PED.
Article
Entomology
Daniel Cook, Boyd Tarlinton, James M. McGree, Alethea Blackler, Caroline Hauxwell
Summary: Strength auditing of European honey bee colonies is critical for colony health management. This study evaluates the use of temperature sensing technology in colony strength assessment and identifies key parameters linking temperature to colony strength. The presence of bees in hives significantly affects hive temperature and range, and sensor placement across the width of the hive is important when linking sensor data with colony strength. Statistical models can be used to predict colony strength from temperature sensor data.
JOURNAL OF ECONOMIC ENTOMOLOGY
(2022)
Article
Medicine, Research & Experimental
J. M. McGree, C. Hockham, S. Kotwal, A. Wilcox, A. Bassi, C. Pollock, L. M. Burrell, T. Snelling, V. Jha, M. Jardine, M. Jones
Summary: The CLARITY trial is a randomized controlled trial investigating the effectiveness of angiotensin receptor blockers in COVID-19 patients. The trial design is Bayesian adaptive, with the primary outcome being clinical status assessed using a 7-point ordinal scale. The trial will follow the intention-to-treat principle in data analysis.
Article
Engineering, Industrial
Harry Sisley, Guvenc Dik, James McGree, Paul Corry
Summary: This paper examines the challenges of managing the supply chain for seasonal crops in Australia and introduces a supply chain model that manages the production of multiple crops across different regions. By using deterministic mixed integer programming and a heuristic solution method, the proposed model can solve the problem faster and with less deviation in the planning process.
INTERNATIONAL JOURNAL OF PRODUCTION RESEARCH
(2023)
Article
Ecology
Pubudu Thilan Abeysiri Wickrama Liyanaarachchige, Rebecca Fisher, Helen Thompson, Patricia Menendez, James Gilmour, James M. McGree
Summary: This article describes the characteristics of time series data commonly observed in ecological monitoring and proposes methods for modeling and adaptive monitoring in such settings. Analyzing the monitoring data from Scott Reef, it is found that future monitoring designs do not need to prioritize specific locations and that sampling sites can be omitted based on observed disturbances without substantial loss in expected information gain. As the methods developed in this study are generic, this research has the potential to improve ecological monitoring in collecting complex data over time.
ECOLOGY AND EVOLUTION
(2022)
Article
Mathematical & Computational Biology
S. G. J. Senarathne, Werner G. Mueller, James M. McGree
Summary: Model-based geostatistical design involves selecting locations to collect data in order to minimize an expected loss function over all possible locations. The loss function reflects the goal of data collection, which in geostatistical studies is often to minimize prediction uncertainty at unobserved locations. This paper proposes a new approach to this design problem by considering the entropy of model predictions and parameters as part of the loss function. The approach extends to generalized linear spatial models, allowing for experiments with multiple responses.
BIOMETRICAL JOURNAL
(2023)
Article
Environmental Sciences
A. W. L. P. Thilan, P. Menendez, J. M. McGree
Summary: This study developed an approach to assess trends in hard coral cover and evaluate the effectiveness of adaptive designs for estimating such trends in coral reef communities. The findings show that adaptive designs can maintain trends over time with little to no loss in information, even with reduced sampling effort. This research serves to further promote adaptive design methods for efficient and effective ecological monitoring.
Article
Engineering, Civil
I. P. Gustave S. Pariartha, Shubham Aggarwal, Srinivas Rallapalli, Prasanna Egodawatta, James McGree, Ashantha Goonetilleke
Summary: Climate change and urbanization have adverse impacts on rainfall and sea level, contributing to future flood risk. This study presents an innovative flood damage and hazard prediction model that integrates MIKE FLOOD and GIS technology to assess flood scenarios for different time horizons. Results show that changes in rainfall patterns significantly affect the average annual damage caused by flooding. The proposed model can guide decision-makers in assessing future flood management.
JOURNAL OF HYDROLOGY
(2023)
Article
Agronomy
Mahmoud Masoud, Jeff Hsieh, Kate Helmstedt, James McGree, Paul Corry
Summary: Beef production plays a crucial role in Australia's agricultural economy, with an annual agricultural production value of AUD11 billion. The profitability of cattle farms is highly affected by weather conditions and the associated uncertainty in pasture growth, as well as the need to manage stocking rates to prevent overgrazing. Predictive modeling of pasture growth and cattle weight gain can effectively assist producers in managing this challenge.
FOOD AND ENERGY SECURITY
(2023)
Article
Health Policy & Services
Jack Powers, James M. McGree, David Grieve, Ratna Aseervatham, Suzanne Ryan, Paul Corry
Summary: The use of a dynamic priority scoring (DPS) system can prioritize elective surgery patients more equitably based on waiting time and clinical factors, reducing subjectivity and increasing transparency in the waiting list management.
HEALTH CARE MANAGEMENT SCIENCE
(2023)
Article
Computer Science, Artificial Intelligence
Sander J. J. Leemans, James M. McGree, Artem Polyvyanyy, Arthur H. M. ter Hofstede
Summary: Through process mining, organisations can improve business processes by utilizing recorded data. Despite advances in the field, a solid statistical foundation is still lacking. This article contributes statistical tests and measures for treating process behavior as a variable, providing a more objective assessment method.
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING
(2023)
Article
Mathematics, Interdisciplinary Applications
Antony Overstall, James McGree
Summary: In this paper, an extended framework is proposed to enhance robustness, ensure computational feasibility, and allow realistic prior specification. An asymptotic approximation to the expected loss under an alternative model is derived, and the properties of different loss functions are established. The framework is demonstrated in various experimental design scenarios.
Correction
Health Care Sciences & Services
Suzanne McDonald, Samuel X. Tan, Shamima Banu, Mieke van Driel, James M. McGree, Geoffrey Mitchell, Jane Nikles
PATIENT-PATIENT CENTERED OUTCOMES RESEARCH
(2022)
Article
Health Care Sciences & Services
Suzanne McDonald, Samuel X. Tan, Shamima Banu, Mieke van Driel, James M. McGree, Geoffrey Mitchell, Jane Nikles
Summary: ME/CFS is a chronic condition with unknown causes, characterized by a variety of disabling symptoms. The heterogeneity of symptom presentation makes management challenging, as treatments may not work for all individuals. This study aims to explore the feasibility and acceptability of using novel patient-centred N-of-1 observational designs to investigate symptom fluctuations and triggers in ME/CFS at the individual level.
PATIENT-PATIENT CENTERED OUTCOMES RESEARCH
(2022)
Article
Medicine, Research & Experimental
Jane Nikles, Patrick Onghena, Johan W. S. Vlaeyen, Rikard K. Wicksell, Laura E. Simons, James M. McGree, Suzanne McDonald
Summary: This article explores the unique features of Single-Case Designs, the establishment and activities of the International Collaborative Network for N-of-1 Trials and Single-Case Designs (ICN), and encourages healthcare professionals and researchers to join and learn more about SCDs for improved health outcomes.
CONTEMPORARY CLINICAL TRIALS COMMUNICATIONS
(2021)
Article
Medicine, Research & Experimental
Guy Bashford, Samuel X. Tan, James McGree, Veronica Murdoch, Jane Nikles
Summary: Management of neuropathic pain is often ineffective due to high variability in patient response. N-of-1 trials offer a potentially precise and cost-effective method for selecting treatments for individual patients with neuropathic pain.
CONTEMPORARY CLINICAL TRIALS COMMUNICATIONS
(2021)