Article
Engineering, Multidisciplinary
Sergio A. Carvajal, Andres F. Medina, Andres J. Bohorquez, Ciro A. Sanchez
Summary: The paper presents an approach that combines Bayesian inference and experimental data to determine the calibration intervals of instruments. This method is suitable for small sample sizes and allows for better utilization of previous knowledge about instrument performance compared to other methodologies.
Article
Computer Science, Artificial Intelligence
Nicolas Dewolf, Bernard De Baets, Willem Waegeman
Summary: In this independent comparative study, four classes of methods, including Bayesian methods, ensemble methods, direct interval estimation methods, and conformal prediction methods, are reviewed for their validity and calibration in the regression setting. Results on benchmark data sets show large fluctuations in performance across different domains. Conformal prediction can be used as a general calibration procedure for methods that deliver poor results without calibration.
ARTIFICIAL INTELLIGENCE REVIEW
(2023)
Article
Thermodynamics
Elie Solai, Heloise Beaugendre, Ulrich Bieder, Pietro Marco Congedo
Summary: This paper proposes an innovative method to deal with uncertainties in the internal resistance of Li-ion batteries using experimental data and numerical simulation. A CFD model is used to replicate the behavior of heated Li-ion battery cells under constant discharging current conditions. A methodology based on Uncertainty Quantification is proposed to represent and quantify the impact of uncertainties on the temperature evolution of Li-ion cells. Experimental measurements are used to perform a Bayesian inference of the internal resistance model parameters, significantly reducing prediction uncertainty. An enhanced internal model is constructed by considering the state of charge and temperature dependency on internal resistance, and its performance is compared to a standard model under low state of charge situations.
APPLIED THERMAL ENGINEERING
(2023)
Article
Engineering, Electrical & Electronic
Gersende Fort, Barbara Pascal, Patrice Abry, Nelly Pustelnik
Summary: Monitoring the COVID19 pandemic is important and has received considerable research attention. The reproduction number is a critical measure of the intensity of the pandemic on a given territory. However, the current method for estimating the reproduction number lacks the ability to provide credibility interval based estimates, which is a limitation for practical use in pandemic monitoring. This study aims to overcome this limitation by incorporating Monte Carlo sampling into a Bayesian framework.
IEEE TRANSACTIONS ON SIGNAL PROCESSING
(2023)
Article
Biochemistry & Molecular Biology
Daniele Salvi, Catarina Pinho, Joana Mendes, D. James Harris
Summary: Podarcis wall lizards are endemic to the Mediterranean Basin, with their diversification occurring rapidly during the Middle Miocene, associated with European climate changes and geological history. Multiple dispersal and vicariant events at different time frames are needed to explain current allopatric distributions and Mediterranean biota assembly, cautioning against biogeographic calibrations based on the assumption of vicariance.
MOLECULAR PHYLOGENETICS AND EVOLUTION
(2021)
Article
Engineering, Electrical & Electronic
Xiaoyue Qiao, Guoqing Ding, Xin Chen, Ping Cai, Li Shao
Summary: Three-dimensional self-calibration is a cost-effective and efficient method for calibrating high-precision 3-D measurement instruments. This study extended the iterative optimization method in 2-D self-calibration to 3-D and compared it with the established equation method based on least squares. The effectiveness and uncertainties of these two methods were analyzed through simulation and experiments.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2022)
Article
Energy & Fuels
Adrian Chong, Godfried Augenbroe, Da Yan
Summary: This study investigates the impact of occupancy data on energy prediction accuracy through Bayesian calibration, finding that using WiFi data to derive occupancy and plug-load schedules can improve the accuracy of building energy models. However, increasing the spatial resolution of occupancy data may result in trade-offs between parameter uncertainty and model bias.
Article
Energy & Fuels
Adrian Chong, Godfried Augenbroe, Da Yan
Summary: This study investigates the impact of different spatial resolutions of occupancy data on building energy use prediction accuracy, finding that deriving occupancy and plug-load schedules from WiFi data can improve the accuracy of building energy models. However, increasing spatial resolution may lead to trade-offs between parameter uncertainty and model bias/inadequacy.
Article
Mathematical & Computational Biology
Hisashi Noma, Tomohiro Shinozaki, Katsuhiro Iba, Satoshi Teramukai, Toshi A. Furukawa
Summary: In clinical prediction models, optimism corrections should be applied to correct for biased results and undercoverage issues in confidence intervals of prediction accuracy measures. The proposed generic bootstrap methods showed favorable coverage performances in numerical evaluations by simulations, highlighting the importance of adjusting confidence intervals for accurate predictions.
STATISTICS IN MEDICINE
(2021)
Article
Environmental Sciences
Shenghai Huang, Caiyun Lu, Hongwen Li, Jin He, Qingjie Wang, Panpan Yuan, Jing Xu, Shan Jiang, Dong He
Summary: In order to improve the accuracy of the acoustic-soil discrete element simulation model, this study focused on the parameter calibration of the model. Sensitivity ranking of soil parameters affecting the dominant frequency and velocity of the acoustic wave was obtained using the Plackett-Burman test scheme. The regression relationship between the dominant frequency and the velocity of the sound wave and soil parameters was established using the Box-Behnken test scheme. This study provided insights into the construction accuracy of the acoustic wave-soil discrete element model and offered a reference for other fields' discrete element models.
Article
Automation & Control Systems
Hui Liu, Si-ying Ling, Li-ding Wang, Zhen-jiang Yu, Xiao-dong Wang
Summary: This study proposes a parametric form of volumetric error modeling and verification methods to overcome theoretical calculation errors, with a new optimized algorithm. Through numerical simulations, the algorithm is shown to improve modeling accuracy, and the two verification methods can check the veracity of parametric modeling precision and isolate theoretical calculation errors.
INTERNATIONAL JOURNAL OF ADVANCED MANUFACTURING TECHNOLOGY
(2021)
Article
Engineering, Environmental
Xiao Zhou, Shuyi Guo, Kunlun Xin, Weirong Xu, Tao Tao, Hexiang Yan
Summary: The upgrading of water supply services requires accurate and adaptive numerical models, and data assimilation methods can enhance the long-term accuracy and stability of the models by reducing uncertainties.
Article
Engineering, Civil
Rajesh Ranjan, Ashok Mishra
Summary: This study compared the impact of three calibration methods on model performance and parameter uncertainty using the SWAT model, finding that the sequential-simultaneous calibration method outperformed the other two methods in simulating streamflow and sediment load.
JOURNAL OF HYDROLOGY
(2022)
Article
Agriculture, Multidisciplinary
Tan Jun-wei, Duan Qing-yun, Gong Wei, Di Zhen-hua
Summary: This study compares different parameter estimation methods and evaluates their impact on model predictions. The results show that different methods yield different parameter estimates, which have a significant effect on model predictions. The frequentist methods are sensitive to initial values, and the SCE-UA method is recommended. All parameter estimates improve the model fit, but the Bayesian-based methods are relatively less effective.
JOURNAL OF INTEGRATIVE AGRICULTURE
(2022)
Article
Engineering, Industrial
Marion Goedel, Nikolai Bode, Gerta Koester, Hans-Joachim Bungartz
Summary: Crowd simulation plays a crucial role in assessing risks and ensuring crowd safety at events and in built environment. This study compares point estimates and Approximate Bayesian Computation as calibration techniques for a microscopic model and its emulator in crowd dynamics. By calibrating on data measuring the flow through a bottleneck scenario, the advantages and shortcomings of the two techniques are demonstrated through three case studies. It is concluded that while point estimation often suffices in practice, Bayesian inference is necessary to capture important structural information about uncertain parameters and the physics of safety.
Article
Zoology
Carlos G. Schrago, Hector N. Seuanez
AMERICAN JOURNAL OF PRIMATOLOGY
(2019)
Article
Ecology
Ian V. Caldas, Carlos G. Schrago
ECOLOGY AND EVOLUTION
(2019)
Article
Virology
Lucia P. Barzilai, Carlos G. Schrago
ARCHIVES OF VIROLOGY
(2019)
Article
Biochemistry & Molecular Biology
Qiqing Tao, Koichiro Tamura, Beatriz Mello, Sudhir Kumar
MOLECULAR BIOLOGY AND EVOLUTION
(2020)
Article
Evolutionary Biology
Alessandra P. Lamarca, Carlos G. Schrago
BIOLOGICAL JOURNAL OF THE LINNEAN SOCIETY
(2020)
Review
Ecology
Carlos G. Schrago, Beatriz Mello
Article
Biochemistry & Molecular Biology
Beatriz Mello, Qiqing Tao, Jose Barba-Montoya, Sudhir Kumar
Summary: Simultaneous molecular dating of population and species divergences is crucial in various biological investigations, and methods such as Bayesian relaxed clock and non-Bayesian RelTime are commonly used. The study indicates that Bayesian approach generally provides accurate molecular date estimates for datasets containing both population and species variations, while RelTime shows similar performance in terms of computational efficiency.
MOLECULAR ECOLOGY RESOURCES
(2021)
Editorial Material
Multidisciplinary Sciences
Carlos G. Schrago
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA
(2022)
Article
Biochemistry & Molecular Biology
Lucia P. Barzilai, Carlos G. Schrago
Summary: Tree shape metrics offer a fast alternative to statistical methods and evolutionary models in analyzing vast amounts of data. This study indicates that these metrics can successfully distinguish between different evolutionary regimes, including negative, positive, frequency-dependent selection, and neutral evolution. The genetic diversity of the viral population at the start of the simulation also has an impact on the differentiation of evolutionary scenarios. These findings highlight the importance of considering tree topologies in understanding viral evolutionary dynamics.
MOLECULAR PHYLOGENETICS AND EVOLUTION
(2023)
Article
Biochemistry & Molecular Biology
Carlos G. Schrago, Lucia P. Barzilai
Summary: Estimating evolutionary parameters in short time intervals can be challenging due to the complexity of virus population dynamics, and using tree topologies shape as proxies is important for assessing the effectiveness of virus phylogeny. Estimations at shorter timescales have larger uncertainties, and caution should be exercised when evaluating estimates from these datasets.
GENETICS AND MOLECULAR BIOLOGY
(2021)
Article
Evolutionary Biology
Guilherme Rezende Dias, Eduardo Guimaraes Dupim, Thyago Vanderlinde, Beatriz Mello, Antonio Bernardo Carvalho
BMC EVOLUTIONARY BIOLOGY
(2020)
Article
Biochemistry & Molecular Biology
Lucas Freitas, Rafael D. Mesquita, Carlos G. Schrago
GENETICS AND MOLECULAR BIOLOGY
(2020)
Article
Evolutionary Biology
Beatriz Mello, Carlos G. Schrago
EVOLUTIONARY BIOINFORMATICS
(2019)
Meeting Abstract
Zoology
E. M. Costa-Paiva, C. G. Schrago, K. M. Halanych
INTEGRATIVE AND COMPARATIVE BIOLOGY
(2018)