Article
Chemistry, Analytical
Abdulmajid Murad, Frank Alexander Kraemer, Kerstin Bach, Gavin Taylor
Summary: This study applies state-of-the-art techniques of uncertainty quantification in data-driven air quality forecasts, finding that Bayesian neural networks provide a more reliable uncertainty estimate but can be challenging to implement and scale. Other scalable methods, such as deep ensemble, Monte Carlo dropout, and stochastic weight averaging-Gaussian (SWAG), can perform well if applied correctly but with different tradeoffs and slight variations in performance metrics. The results demonstrate the practical impact of uncertainty estimation and highlight the suitability of probabilistic models for informed decision-making.
Article
Environmental Sciences
Henrik Jan Persson, Magnus Ekstrom, Goran Stahl
Summary: This paper discusses the impact of errors in field reference data on the accuracy of forest remote sensing predictions. It presents new theoretical analysis methods and an error characterization model to address these errors. Based on analysis of data from Scandinavian forests, it concludes that field reference errors have a significant impact on remote sensing-based predictions, with the most influential sources of error being residual errors of the biomass model and field plot position errors.
REMOTE SENSING OF ENVIRONMENT
(2022)
Article
Engineering, Geological
Ibsen Chivata Cardenas
Summary: In this paper, a new two-dimensional approach to quantify stratigraphic uncertainty is proposed and demonstrated using a case analysis, showing its usefulness in exploring and representing geological structures and quantifying uncertainty in a computationally inexpensive way.
ENGINEERING GEOLOGY
(2023)
Article
Public, Environmental & Occupational Health
Ivette Raices Cruz, Matthias C. M. Troffaes, Ullrika Sahlin
Summary: Honest communication of uncertainty is important for transparency in scientific assessments. The European Food Safety Authority recommends expressing epistemic uncertainty quantitatively using subjective probability and bounded probability can reflect individual's knowledge strength.
Article
Business
Bobby J. Calder, Sharlene He, Brian Sternthal
Summary: Behavioral research is often limited to independent domains within the effect paradigms. Theories are developed at a fine-grained level, closely tied to specific independent variables or variables that moderate the effects. This lack of integration between different effect paradigms hampers the cumulative knowledge and application of research findings. To address this issue, the Ambiguity-Adoptability-Accessibility (3A) framework is proposed, which provides a comprehensive theoretical framework that encompasses multiple effect paradigms and enhances explanatory understanding. The framework is illustrated with examples from four different effect paradigms, demonstrating its potential for improving the applicability of behavioral research.
JOURNAL OF BUSINESS RESEARCH
(2023)
Article
Agronomy
Alec R. Kowalewski, Charles J. Schmid, Emily T. Braithwaite, Brandon C. McNally, Matt T. Elmore, Clint M. Mattox, Brian W. McDonald, Ruying Wang, John G. Lambrinos, Greg S. Fitzpatrick, Hannah M. Rivedal
Summary: Turfgrass cover can be qualitatively assessed visually, but quantitative measurements are desired for unbiased data. Two quantitative methods, digital image analysis and point intercept, were evaluated in this study. The point intercept method showed accuracy and low variance, while digital image analysis was consistent when quantifying known percent cover. For dollar spot cover assessment, all three methods (point intercept, digital image analysis, and visual ratings) were well-correlated. However, digital image analysis was not correlated with other methods for anthracnose cover assessment. In addition, both point intercept and digital image analysis were correlated when quantifying turfgrass establishment from seed.
Article
Energy & Fuels
Sumanth Yamujala, Priyanka Kushwaha, Anjali Jain, Rohit Bhakar, Jianzhong Wu, Jyotirmay Mathur
Summary: Operational flexibility is crucial in power systems to mitigate load-generation imbalances, requiring a comprehensive consideration of various flexibility aspects such as ramp, power, and energy. This paper proposes a comprehensive metric to quantify flexibility and highlights the importance of considering all three facets for adequate assessment and management of netload intermittency. The proposed tools provide valuable insights for power system planners and operators in handling systemic inflexibility events.
Article
Computer Science, Artificial Intelligence
Aleix Boquet-Pujadas, Jean-Christophe Olivo-Marin
Summary: This article introduces a Bayesian PDE-constrained framework that directly transforms visual information into physical measurements. By tracking brightness and satisfying the physical model, the aperture problem can be translated from motion to underlying physics, while measurement error is derived from posterior covariance. This approach offers advantages such as accurate reconstructions, flexibility in experiment design, and exclusive measurement error.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Article
Computer Science, Information Systems
Jeahan Jung, Minseok Choi
Summary: Polynomial chaos (PC) is an efficient method for uncertainty quantification, but it is limited by assumptions of mutual independence and exact knowledge of distribution. We propose a new data-driven method for dealing with correlated multivariate random variables, transforming them into independent random variables using singular value decomposition. The transformed variables can be used to construct a PC basis for building a surrogate model, and when combined with ANOVA, it can accurately quantify high-dimensional uncertainties with fewer simulations compared to Monte Carlo method.
Article
Engineering, Multidisciplinary
Seyed Mohammad Jafar Jalali, Sajad Ahmadian, Md Kislu Noman, Abbas Khosravi, Syed Mohammed Shamsul Islam, Fei Wang, Joaao P. S. Catalao
Summary: In this work, a deep learning-based prediction interval framework is proposed for modeling the forecasting uncertainties of tidal current datasets. The framework optimizes the architecture automatically and improves the performance of the prediction intervals. The correlation between the PI coverage probability and PI normalized average width is established using the coverage width criterion.
IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS
(2022)
Article
Automation & Control Systems
Peng Wang, Shaobu Wang, Renke Huang
Summary: This paper investigates the bounds estimation for LTV systems with known inputs and unknown but bounded inputs, and analyzes the gap between synchronous generators and their models in different scenarios using simulation.
IET CONTROL THEORY AND APPLICATIONS
(2022)
Article
Environmental Sciences
Jina Yin, Josue Medellin-Azuara, Alvar Escriva-Bou, Zhu Liu
Summary: This study introduces a novel machine learning-based groundwater ensemble modeling framework combined with Bayesian model averaging to predict groundwater storage change in agricultural regions with improved reliability.
SCIENCE OF THE TOTAL ENVIRONMENT
(2021)
Article
Ecology
Jeremy Lobry, Florence Mounier, Marine Ballutaud, Xavier Chevillot, Didier Gascuel, Helene Budzinski, Pierre Labadie, Hilaire Drouineau
Summary: Food-web modelling is a crucial tool for understanding community structure, biodiversity, ecosystem processes, and functioning. This study presents an innovative approach that combines a Bayesian mixing model with classical mass-balance equations to estimate diet composition, isotopic enrichment, contaminant biomagnification, and contaminants and biomass flows in the whole food web.
METHODS IN ECOLOGY AND EVOLUTION
(2022)
Article
Meteorology & Atmospheric Sciences
Yuejian Zhu, Bing Fu, Bo Yang, Hong Guan, Eric Sinsky, Wei Li, Jiayi Peng, Xianwu Xue, Dingchen Hou, Xin-Zhong Liang, Sanghoon Shin
Summary: The Global Ensemble Forecast System version 12 (GEFSv12) has been implemented into National Centers For Environmental Prediction operations since September 2020, which improved forecast skills in many categories by increasing horizontal resolution, ensemble members, and extended forecasts. The improvements were achieved through upgrades in model resolution, data assimilation, and stochastic schemes. Coupled GEFS experiments further improved sub-seasonal forecast skill by coupling atmospheric, land surface, ocean, ice, and wave models, reducing forecast uncertainties and improving correlation.
JOURNAL OF GEOPHYSICAL RESEARCH-ATMOSPHERES
(2023)
Article
Engineering, Electrical & Electronic
Hoang Duc Pham, Katja Tueting, Heyno Garbe
Summary: The quality of field probe calibration directly impacts the uncertainty in radiated immunity tests, requiring precise calibration. IEEE Std 1309 offers three calibration methods, with Method B using a calculated reference field, though mechanical tolerances and misalignment can affect actual results. This study proposes a new method for computing electromagnetic fields inside a TEM-cell.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2021)
Review
Industrial Relations & Labor
Marko Sarstedt, Nicholas P. Danks
Summary: It is crucial for researchers to clearly distinguish between explanatory power and predictive power when evaluating statistical models to avoid confusion. Researchers often focus solely on explanatory power but derive practical recommendations inherently from a predictive scenario.
HUMAN RESOURCE MANAGEMENT JOURNAL
(2022)
Article
Mathematics, Interdisciplinary Applications
Gyeongcheol Cho, Marko Sarstedt, Heungsun Hwang
Summary: Structural equation modelling (SEM) has two domains, factor-based and component-based, each with their own population models. Component-based approaches are generally more robust to construct misrepresentation and should be preferred over factor-based approaches for component models. Among component-based approaches, GSCA is recommended over PLSPM regardless of construct misrepresentation.
BRITISH JOURNAL OF MATHEMATICAL & STATISTICAL PSYCHOLOGY
(2022)
Article
Computer Science, Interdisciplinary Applications
Dan-Andrei Sitar-Taut, Daniel Mican, Lena Froembling, Marko Sarstedt
Summary: This study investigates the impact of social support on the transition of university students during the COVID-19 pandemic and finds that social media plays a crucial role in this process. The study results indicate that perceived social support, as well as bonding and bridging social capital, are particularly important during the transition.
SOCIAL SCIENCE COMPUTER REVIEW
(2023)
Article
Business
Marko Sarstedt, Joseph F. Hair, Mandy Pick, Benjamin D. Liengaard, Lacramioara Radomir, Christian M. Ringle
Summary: This paper presents a new analysis of the use of partial least squares structural equation modeling (PLS-SEM) in marketing research, focusing on articles published between 2011 and 2020 in the top 30 marketing journals. The study finds that although researchers are more aware of the when's and how's of PLS-SEM use, there is still some delay in adopting the best practices for model evaluation. Recommendations for future PLS-SEM use, guidelines for its application, and areas of further research interest are provided based on the review results.
PSYCHOLOGY & MARKETING
(2022)
Article
Business
Marko Sarstedt, Joseph F. Hair, Christian M. Ringle
Summary: The passage discusses the significance of a paper published in the Journal of Marketing Theory & Practice in 2011, which became a cornerstone contribution in the field. The authors offer a review of their own paper from the perspective of reviewers with current knowledge, clarifying ambiguities and updating outdated descriptions.
JOURNAL OF MARKETING THEORY AND PRACTICE
(2023)
Article
Computer Science, Information Systems
Matthias Soellner, Abhay Nath Mishra, Jan-Michael Becker, Jan Marco Leimeister
Summary: This paper examines the continued system use (CSU) by individuals in utilitarian, volitional contexts using a longitudinal perspective. It focuses on the impact of habit and continuance intention on CSU and analyzes how these relationships evolve over time. The results suggest that the impact of continuance intention on CSU and the interaction effect between habit and intention are increasing over time.
EUROPEAN JOURNAL OF INFORMATION SYSTEMS
(2022)
Review
Hospitality, Leisure, Sport & Tourism
Jan-Michael Becker, Jun-Hwa Cheah, Rasoul Gholamzade, Christian M. Ringle, Marko Sarstedt
Summary: This study addresses the uncertainties and questions surrounding the application of PLS-SEM by analyzing a well-known discussion forum. The authors identify the most prominent topics in PLS-SEM and provide guidelines and explanations for each topic. This research provides valuable guidance for PLS-SEM applications in hospitality management research.
INTERNATIONAL JOURNAL OF CONTEMPORARY HOSPITALITY MANAGEMENT
(2023)
Article
Business
Edward Rigdon, Marko Sarstedt, Ovidiu-Ioan Moisescu
Summary: Multimodel inference allows researchers to accurately represent uncertainty about the best model by using Akaike weights and bootstrapping to quantify model selection uncertainty and minimize bias from dubious assumptions.
INTERNATIONAL JOURNAL OF CONSUMER STUDIES
(2023)
Review
Business
Linda D. Hollebeek, Marko Sarstedt, Choukri Menidjel, David E. Sprott, Sigitas Urbonavicius
Summary: This study reviews the major scales used to measure consumer engagement (CE) with a brand or specific brand elements and finds theoretical contamination in some measures, highlighting the need for scholars to verify the theoretical foundations of their adopted CE scales.
PSYCHOLOGY & MARKETING
(2023)
Article
Multidisciplinary Sciences
Christian M. Ringle, Marko Sarstedt, Noemi Sinkovics, Rudolf R. Sinkovics
Summary: This perspective article provides a guide for authors who want to publish stand-alone data articles that can be analyzed using partial least squares structural equation modeling (PLS-SEM). It offers actionable recommendations for the conceptualization phase, suitable data types, quality criteria, and presents adjusted versions of the HTMT metric for discriminant validity testing. The article also highlights the benefits of linking data articles to published research papers that use PLS-SEM.
Article
Multidisciplinary Sciences
Marko Sarstedt, Christian M. Ringle, Denis Iuklanov
Summary: Corporate reputation is crucial for a company's competitiveness, and managers need to understand its relationship with antecedents and consequences. This article presents a dataset that replicates a model of corporate reputation and its impact on customer satisfaction and loyalty. Mediators and moderators in these relationships extend the model to clarify the mechanism through which corporate reputation affects satisfaction and loyalty. The main effects of the model are documented using PLS-SEM.
Article
Business
Marko Sarstedt, Ovidiu-Ioan Moisescu
Summary: By introducing a new quality assessment dimension, researchers can view mediation analysis as a type of model comparison, quantifying the degree of uncertainty in model effects induced by the introduction of a mediator. This procedure provides a new means for deciding whether to introduce a mediator in a PLS path model and improves the replicability of research results.
JOURNAL OF MARKETING ANALYTICS
(2023)
Article
Business
Marko Sarstedt, Susanne J. Adler
Summary: Recent research proposed an efficient way of streamlining p-hacking: a metric of randomly generated p-values that always indicates significant results. However, the metric's applicability is limited, as it covers only a small range of use cases relevant to p-hacking researchers. To address this issue, we introduce extra pointless, a new metric that allows researchers to freely specify desired p-value ranges while ensuring full reproducibility. Our newly introduced extra pointless metric not only addresses the limitations of the original metric but also meets the demand for alternative and innovative statistical approaches from editors and reviewers. An R-based, interactive shiny web app assists researchers with limited coding background in applying our metric.
JOURNAL OF BUSINESS RESEARCH
(2023)
Article
Business
Michael Canty, Felix Josua Lang, Susanne Jana Adler, Marcel Lichters, Marko Sarstedt
Summary: Psychological state alterations induced by substance-related physiological mechanisms, such as caffeine intake, have been found to influence consumer decision-making. This study specifically focuses on the attraction effect, and finds that high caffeine intake is associated with a larger attraction effect in real product choice tasks.