Article
Ecology
Hanna M. McCaslin, Abigail B. Feuka, Mevin B. Hooten
Summary: Bayesian hierarchical models play a crucial role in ecology, but can be computationally intensive. Recursive Bayesian computing and transformation-assisted RB methods help improve the efficiency and interpretability of Bayesian models, reducing computation time for fitting complex ecological statistical models.
METHODS IN ECOLOGY AND EVOLUTION
(2021)
Article
Statistics & Probability
Alan Riva-Palacio, Fabrizio Leisen, Jim Griffin
Summary: A novel Bayesian nonparametric model for regression in survival analysis, which can efficiently model hazards and allow nonproportionality, is presented. The model, characterized by competing latent risks, utilizes an MCMC scheme for Bayesian inference of posterior means and credible intervals.
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION
(2022)
Article
Economics
Paul Ho
Summary: This paper develops a tool for global prior sensitivity analysis in large Bayesian models. The methodology provides bounds for posterior means or quantiles given any prior close to the original in relative entropy and reveals features of the prior that are important for the posterior statistics of interest. It finds that the prior tightness hyperparameters in the hierarchical vector autoregression model from Giannone et al. (2015) are relatively insensitive to their hyperpriors, but the error bands for the impulse response of output to a monetary policy shock in the New Keynesian model of Smets and Wouters (2007) depend heavily on the prior.
JOURNAL OF ECONOMETRICS
(2023)
Article
Physics, Multidisciplinary
Jakob Robnik, Uros Seljak
Summary: In hypothesis testing applications, mixed priors are common, with informative priors for some parameters but not for others. Bayesian methodology is helpful for informative priors, while frequentist hypothesis testing is better for cases where the prior is not completely known. Combining both methodologies by using the Bayes factor as a test statistic in frequentist analysis is recommended when only partial prior information is available.
Article
Economics
Jan Prueser
Summary: This paper proposes two data-based priors for vector error correction models, which require minimal user input and have automatic approaches. The first prior encourages shrinkage towards a low-rank, row-sparse, and column-sparse long-run matrix, while the second prior shrinks all elements of the long-run matrix towards zero using the horseshoe prior. Empirical investigations show that Bayesian vector error correction models equipped with these priors perform well in higher dimensions and forecasting. Compared to VARs in first differences, they effectively exploit the information in level variables and improve forecasts for some macroeconomic variables. A simulation study demonstrates that the BVEC with data-based priors has good frequentist estimation properties.
INTERNATIONAL JOURNAL OF FORECASTING
(2023)
Article
Computer Science, Artificial Intelligence
Georg Ralph Spinner, Christian Federau, Sebastian Kozerke
Summary: The study improved the IVIM model parameter estimation performance in brain cancer and acute stroke patients using Bayesian inference, showing better results compared to other methods in both simulation and real data. By combining hierarchical and spatial priors, the method effectively identified and analyzed the brain pathologies, demonstrating reduced errors and improved contrast-to-noise ratio compared to conventional approaches.
MEDICAL IMAGE ANALYSIS
(2021)
Article
Mathematics, Interdisciplinary Applications
Dimitris Fouskakis, Ioannis Ntzoufras
Summary: This paper explores the powerexpected-posterior (PEP) prior as a generalization to the expected-posterior prior (EPP) in objective Bayesian model selection under normal linear models. It is proven that PEP can be represented as a mixture of g-prior, providing closed-form posterior distributions and Bayes factors for computational tractability. Comparisons with other mixtures of g-prior are made and results are presented in both simulated and real-life datasets.
Article
Acoustics
Liang Yu, Yue Bai, Ran Wang, Kang Gao, Weikang Jiang
Summary: With the increasing bypass ratio, fan noise has become an important component of civil aviation engines. Mode identification plays a key role in in-duct fan noise testing, but it is challenging due to the large number of modes and limited number of microphones. This paper proposes a Sparse Bayesian Learning algorithm to solve the problem by describing the in-duct sound field with a statistical model and using a block coordinate descent algorithm for sparse solution.
JOURNAL OF SOUND AND VIBRATION
(2023)
Article
Construction & Building Technology
Matteo Favero, Antonio Luparelli, Salvatore Carlucci
Summary: Thermal comfort research aims to study the relationship between thermal environment and human sense of warmth, usually through subjective thermal response. However, using linear regression to analyze ordinal data may result in severe errors. This study establishes methodological foundations for analyzing subjective thermal comfort data and demonstrates the practical consequences of fallacious assumptions through a Bayesian approach. The findings highlight the importance of using ordinal models instead of metric models for analyzing ordinal data.
ENERGY AND BUILDINGS
(2023)
Article
Operations Research & Management Science
Peter Tea, Tim B. Swartz
Summary: Anticipating opponents' serve and being aware of one's own serve tendencies are essential skills in tennis. Using Bayesian hierarchical models, this paper investigates the intended serve direction of professional tennis players at Roland Garros and reveals discernible differences between men's and women's tennis, as well as individual players.
ANNALS OF OPERATIONS RESEARCH
(2023)
Article
Economics
Hedibert F. Lopes, Robert E. McCulloch, Ruey S. Tsay
Summary: State-space models are widely used in engineering, economics, and statistics, and have attracted much interest in Bayesian analysis. However, prior specification is a challenging issue in large scale models. This paper proposes a flexible prior for state-space models, which achieves parsimony in high-dimensional systems by mixing four commonly entertained models.
JOURNAL OF ECONOMETRICS
(2022)
Article
Engineering, Electrical & Electronic
Pakshal Bohra, Thanh-an Pham, Jonathan Dong, Michael Unser
Summary: This study presents a Bayesian reconstruction framework for nonlinear imaging models, where the prior knowledge on the image is specified through a deep generative model. The authors develop a tractable posterior-sampling scheme based on the Metropolis-adjusted Langevin algorithm for the class of nonlinear inverse problems. The advantages of this framework are illustrated through its application to various nonlinear imaging modalities, such as phase retrieval and optical diffraction tomography.
IEEE TRANSACTIONS ON COMPUTATIONAL IMAGING
(2022)
Article
Biochemical Research Methods
Jacob Williams, Shuangshuang Xu, Marco A. R. Ferreira
Summary: In this study, a novel Bayesian variable selection method based on nonlocal priors is proposed for genome-wide association studies. The method, called BGWAS, effectively reduces false positive rates while maintaining the ability to detect true positive SNPs. It achieves this through a two-step process of screening and model selection.
BMC BIOINFORMATICS
(2023)
Article
Meteorology & Atmospheric Sciences
Alvaro Ossandon, J. S. Nanditha, Pablo A. Mendoza, Balaji Rajagopalan, Vimal Mishra
Summary: In this study, a Bayesian hierarchical model (BHM) was developed and tested to improve the predictions of physically based hydrological models and generate ensemble forecasts. The results showed that the BHM model increased probabilistic skill and provided reliable ensemble forecasts for multiple sites.
JOURNAL OF HYDROMETEOROLOGY
(2022)
Article
Psychology, Multidisciplinary
Michael J. Zyphur, Ellen L. Hamaker, Louis Tay, Manuel Voelkle, Kristopher J. Preacher, Zhen Zhang, Paul D. Allison, Dean C. Pierides, Peter Koval, Edward F. Diener
Summary: This article discusses the potential uses of Bayesian estimation in time-series and panel data models by incorporating prior probabilities in addition to observed data. By using informative shrinkage or small variance priors, the article highlights the benefits of increased model parsimony, stability of estimates, and improved out-of-sample predictions and interpretability. The use of priors also allows for estimating otherwise under-identified models and higher-order lagged effects in a more trustworthy manner than under maximum likelihood estimation.
FRONTIERS IN PSYCHOLOGY
(2021)
Article
Ecology
Matthew Hovland, Ricardo Mata-Gonzalez, R. Paul Schreiner, Thomas J. Rodhouse
RANGELAND ECOLOGY & MANAGEMENT
(2019)
Article
Ecology
Daniel M. Esposito, Thomas J. Rodhouse, Ricardo Mata-Gonzalez, Matthew Hovland
RANGELAND ECOLOGY & MANAGEMENT
(2019)
Article
Ecology
Kathryn M. Irvine, Wilson J. Wright, Erin K. Shanahan, Thomas J. Rodhouse
METHODS IN ECOLOGY AND EVOLUTION
(2019)
Article
Ecology
Thomas J. Rodhouse, Rogelio M. Rodriguez, Katharine M. Banner, Patricia C. Ormsbee, Jenny Barnett, Kathryn M. Irvine
ECOLOGY AND EVOLUTION
(2019)
Article
Biodiversity Conservation
Katharine M. Banner, Kathryn M. Irvine, Thomas J. Rodhouse, Deahn Donner, Andrea R. Litt
ECOLOGICAL INDICATORS
(2019)
Article
Biodiversity Conservation
Melissa Nicolli, Thomas J. Rodhouse, Devin S. Stucki, Matthew Shinderman
WESTERN NORTH AMERICAN NATURALIST
(2020)
Article
Ecology
Thomas J. Rodhouse, Kathryn M. Irvine, Lisa Bowersock
FRONTIERS IN ECOLOGY AND EVOLUTION
(2020)
Article
Ecology
Jamie L. Brusa, Jay J. Rotella, Katharine M. Banner, Patrick R. Hutchins
Summary: Survival rates are an important aspect of life-history strategies in large vertebrate species and exhibit variation among different species, particularly in males. However, challenges in obtaining reliable datasets and dealing with measurement errors have limited comparative studies on interspecific variation in survival rates and other life-history traits.
ECOLOGY AND EVOLUTION
(2021)
Article
Ecology
Devin S. Stucki, Thomas J. Rodhouse, Ron J. Reuter
Summary: The common camas bulbs have been a staple food for Indigenous Peoples of western North America for thousands of years, but due to wetland drainage and land conversion, populations have declined. Through a controlled experiment, it was found that a combination of harvesting and burning can promote the growth of adult plants, suggesting a sustainable harvesting interval of approximately 5 years.
ECOLOGY AND EVOLUTION
(2021)
Article
Ecology
Wilson J. Wright, Kathryn M. Irvine, Thomas J. Rodhouse, Andrea R. Litt
Summary: The study utilized a multi-species occupancy model for species distribution predictions, showing significant improvements for species with restricted ranges such as spotted bat, canyon bat, and Brazilian free-tailed bat. In contrast, widespread species like Lasionycteris noctivagans were appropriately modeled using environmental predictors. By incorporating spatial Gaussian processes, the model allows for simultaneous predictions for the entire faunal assemblage even if species have nonoverlapping or restricted ranges within a spatial domain of interest.
ECOLOGY AND EVOLUTION
(2021)
Article
Ecology
Christian Stratton, Kathryn M. Irvine, Katharine M. Banner, Wilson J. Wright, Cori Lausen, Jason Rae
Summary: The increasing complexity and pace of ecological change require natural resource managers to consider entire species assemblages. Acoustic recording units (ARUs) can provide information on relative activity or encounter rates for multiple species, with minimal cost and effort. However, the automated classification process of ARUs can result in species misidentifications, which should be accounted for in conservation decision-making. This study demonstrates that coupled validation methods can reduce bias and uncertainty in estimating relative activity and species classification probabilities, and better adapt to the needs of statistical models.
METHODS IN ECOLOGY AND EVOLUTION
(2022)
Article
Ecology
Mia R. Goldman, Matthew Shinderman, Mackenzie R. Jeffress, Thomas J. Rodhouse, Kevin T. Shoemaker
Summary: Standard occupancy models allow unbiased estimation of occupancy by considering observation errors, and a multi-sign occupancy approach improves estimates of occupancy dynamics for inconspicuous species. Different detection models lead to variations in estimates of occupancy and environmental drivers, indicating that unmodeled heterogeneity in the observation process can introduce biases and uncertainties in occupancy processes and relationships.
ECOLOGY AND EVOLUTION
(2023)
Article
Biodiversity Conservation
Thomas J. Rodhouse, Sara Rose, Trent Hawkins, Rogelio M. Rodriguez
Summary: Bat conservation has been hindered by a lack of basic information on species distributions and abundances. Public participation in citizen science has been limited, but opportunities exist to survey audible bat species globally. Utilizing a structured survey design, a study in western North America focused on rare audible desert bats and successfully updated a Bayesian distribution model for the spotted bat, highlighting the importance of arid cliffs and canyons. Future surveys that integrate citizen science can enhance scientific understanding of rare species and engage the public in conservation efforts.
CONSERVATION SCIENCE AND PRACTICE
(2021)