Article
Biochemical Research Methods
Rebecca A. Deek, Hongzhe Li
Summary: This article proposes the use of copula models with mixed zero-beta margins to estimate taxon-taxon covariations using normalized microbial relative abundance data. The method accurately estimates model parameters and enables the construction of biologically meaningful microbial networks.
Article
Management
Konstantinos Petridis, Nikolaos E. Petridis, Ali Emrouznejad, Fouad Ben Abdelaziz
Summary: This paper proposes a two-stage approach for prioritizing volatility models. In the first stage, a novel method is used to rank the models, and in the second stage, the impact of model characteristics on efficiency scores is analyzed.
INTERNATIONAL TRANSACTIONS IN OPERATIONAL RESEARCH
(2023)
Article
Meteorology & Atmospheric Sciences
Marcel Carvalho Abreu, Amaury de Souza, Gustavo Bastos Lyra, Ivana Pobocikova, Roberto Avelino Cecilio
Summary: This study aims to estimate the monthly and annual mean rainfall in Mato Grosso do Sul state, Brazil, by fitting polynomial models based on latitude, longitude, and altitude coordinates. It was found that an increased number of predictor variables enhances the performance of regression methods, and fitting regressions in hydrologically similar groups through cluster analysis can improve regression performance. However, the limited number of rain gauge stations in Brazil makes applying this technique difficult due to the possibility of having more parameters in the regression models than rain gauge stations in the cluster.
INTERNATIONAL JOURNAL OF CLIMATOLOGY
(2021)
Article
Computer Science, Artificial Intelligence
Zhijian Li, Sulin Pang, Hongying Qu, Wanmin Lian
Summary: This article focuses on the key influencing factors and prediction accuracy of diabetes. By designing data experiment method and logistic regression prediction algorithm, it analyzes nine test indexes and establish two logistic regression models for prediction. The study found that age and blood sugar are the key influencing factors of diabetes. The research method has high application value and can provide scientific solutions for medical institutions to predict, analyze and early diagnose diabetes.
NEURAL COMPUTING & APPLICATIONS
(2023)
Article
Geosciences, Multidisciplinary
Huadan Fan, Yuefeng Lu, Shiwei Shao, Li Li, Yanjun Wang, Miao Lu, Jing Li, Kaizhong Yao, Ying Sun
Summary: This study conducted landslide susceptibility mapping in four counties of the central Ganzi Tibetan Autonomous Prefecture in Sichuan Province, China, based on the calculation of six factors and the use of six models. The IV-AHP model was found to be the most appropriate for assessing landslide disasters in the entire region, while the IV model exhibited the highest accuracy in assessing landslide susceptibility in high and very high-susceptibility regions.
GEOMATICS NATURAL HAZARDS & RISK
(2023)
Article
Energy & Fuels
Maria Soledad Callen, Isabel Martinez, Gemma Grasa, Jose Manuel Lopez, Ramon Murillo
Summary: Gasification is a potential technology for converting biomass into energy, with key parameters such as biomass type, composition, and gas composition influencing the process. Multivariate statistical analysis identified correlations between these parameters and methane, ethylene, and tar contents in the outlet gas. Adjusting gasification bed temperature, CaO/C ratio, and other variables can help improve the quality of the outlet gas.
BIOMASS CONVERSION AND BIOREFINERY
(2022)
Article
Environmental Sciences
Girma Moges, Kevin McDonnell, Mulugeta Admasu Delele, Addisu Negash Ali, Solomon Workneh Fanta
Summary: In this study, two advanced machine learning models were developed for pesticide drift prediction, showing better predictive power compared to traditional regression models. The ability to model complex relationships is beneficial in addressing the variability in pesticide drift.
ENVIRONMENTAL SCIENCE AND POLLUTION RESEARCH
(2023)
Article
Biochemical Research Methods
Shimeng Huang, Elisabeth Ailer, Niki Kilbertus, Niklas Pfister
Summary: In this study, a kernel-based nonparametric regression and classification framework called KernelBiome is proposed for compositional data. It captures complex signals and automatically adapts model complexity. Experimental results on 33 publicly available microbiome datasets demonstrate its superior predictive performance and interpretability compared to state-of-the-art machine learning methods. Additionally, two novel quantities are proposed to interpret contributions of individual components and the connection between kernels and distances aids interpretability.
PLOS COMPUTATIONAL BIOLOGY
(2023)
Article
Multidisciplinary Sciences
Mintode Nicodeme Atchade, Paul P. Tchanati
Summary: This paper develops a robust method for estimating nonlinear regression models that addresses the issues of heteroscedasticity and autocorrelation of errors. The proposed approach is found to be accurate in predicting the COVID-19 cases in Africa and can be useful for decision making in various fields.
Article
Materials Science, Multidisciplinary
Azhari A. Elhag, Tahani A. Aloafi, Taghreed M. Jawa, Neveen Sayed-Ahmed, F. S. Bayones, J. Bouslimi
Summary: The biggest challenge facing the world in 2020 was the COVID-19 pandemic, causing sadness and anxiety. Statistical analysis using artificial neural networks and logistic regression models were utilized to study the pandemic trends. The rationale for methodological tools was based on the high classification accuracy rates.
RESULTS IN PHYSICS
(2021)
Article
Statistics & Probability
M. T. Pratola, E. I. George, R. E. McCulloch
Summary: Bayesian Classification and Regression Trees (BCART) and Bayesian Additive Regression Trees (BART) are popular Bayesian regression models that can flexibly model complex responses and quantify uncertainties. However, there has been little work on evaluating the sensitivity of these models to violations of assumptions.
JOURNAL OF COMPUTATIONAL AND GRAPHICAL STATISTICS
(2023)
Article
Oncology
Shigeyuki Matsui, Jennifer Le-Rademacher, Sumithra J. Mandrekar
Summary: This article emphasizes the importance of using statistical models in clinical studies and suggests collaboration with a statistician to better interpret results and ensure appropriate study design and analysis.
JOURNAL OF THORACIC ONCOLOGY
(2021)
Article
Environmental Sciences
Xiao Pan, Gokhan Yildirim, Ataur Rahman, Khaled Haddad, Taha B. M. J. Ouarda
Summary: This paper presents the development of peaks-over-threshold (POT) based regional flood frequency analysis (RFFA) techniques, comparing them to the traditional annual maximum (AM) flood model. The results show that the regularised linear models provide more accurate flood quantile estimates and greater flexibility than the AM-based RFFA techniques.
Article
Energy & Fuels
Marcello Congro, Alexandre S. Zanatta, Karoline Nunes, Roberto Quevedo, Bruno R. B. M. Carvalho, Deane Roehl
Summary: This article proposes a new method to predict the width of fault damage zones in sandstone reservoirs, considering the impact of geomechanical properties. The method combines a deterministic model, numerical experiments, and statistical analysis to establish a regression model that fits with the predicted information from the deterministic model and field data.
GEOMECHANICS FOR ENERGY AND THE ENVIRONMENT
(2023)
Article
Multidisciplinary Sciences
Tadas Zvirblis, Darius Vainorius, Jonas Matijosius, Kristina Kilikeviciene, Alfredas Rimkus, Akos Bereczky, Kristof Lukacs, Arturas Kilikevicius
Summary: This paper presents a three-step statistical analysis algorithm that shows increased prediction accuracy when using vibration and sound pressure data as a covariate variable in the exhaust emission prediction model. Statistical analysis reveals that non-negative time domain statistics are the best predictors, with only one statistic being a statistically significant predictor for all 11 exhaust parameters. The study also analyzes the ecological and energy parameters of the engine using symmetric methods in terms of fuel type and adjustable engine parameters.