Article
Medicine, General & Internal
Ganggang Wang, Xiaowei Shen, Yicun Wang, Huanhua Lu, Hua He, Xiaoliang Wang
Summary: This study provides guidance for the prevention and diagnosis of nonalcoholic fatty liver disease (NAFLD) by analyzing its risk factors and diagnostic value of each index. The results showed that age, BMI, SBP, ALT, AST, FBG, TBIL, TG, and LDL were risk factors for NAFLD in adults, while HDL was a protective factor. Age, BMI, ALT, TG, and HDL had predictive value for NAFLD occurrence, and the combination of these factors had diagnostic value. In conclusion, managing BMI, blood pressure, blood glucose, and lipid levels while monitoring ALT and AST levels is important for NAFLD prevention, and age, BMI, ALT, TG, and HDL are helpful in the diagnosis of NAFLD.
FRONTIERS IN MEDICINE
(2023)
Article
Engineering, Civil
Mengzhu Chen, Konstantinos Papadikis, Changhyun Jun, Neil Macdonald
Summary: Nonstationary flood frequency analysis (NFFA) is an active area of research in recent years due to its importance in water resources management and hydrologic engineering design. This study evaluates different modelling techniques for NFFA and finds that precipitation-informed models have higher rejection rates compared to time-varying models. Fractional polynomial models are identified as a potential alternative due to their flexibility and simplicity. The impact of seasonal flood variation and catchment characteristics on the goodness-of-fit of NFFA models is unclear.
JOURNAL OF HYDROLOGY
(2023)
Article
Automation & Control Systems
Daniel R. Kowal
Summary: Subset selection is a valuable tool for interpretability, scientific discovery, and data compression. We propose a Bayesian approach to address the challenges in classical subset selection, and introduce a strategy that focuses on finding near-optimal subsets rather than a single best subset. We apply Bayesian decision analysis to derive the optimal linear coefficients for any subset of variables, and our approach outperforms competing methods in prediction, interval estimation, and variable selection. By analyzing a large education dataset, we gain unique insights into the factors that predict educational outcomes and identify over 200 distinct subsets of variables that offer near-optimal predictive accuracy.
JOURNAL OF MACHINE LEARNING RESEARCH
(2022)
Article
Clinical Neurology
Dana Pisica, Ruben Dammers, Eric Boersma, Victor Volovici
Summary: This article provides a theoretical and practical tutorial on regression analysis, including data preparation, univariable and multivariable analysis, and model selection. It also demonstrates the application of regression analysis to real-world data and emphasizes the importance of multidisciplinary collaborations with professionals.
WORLD NEUROSURGERY
(2022)
Article
Physics, Multidisciplinary
Xiaolei Zhang, Guohua Yan, Renjun Ma, Jiaxiu Li
Summary: This paper proposes a new approach to handle longitudinal binomial data with positive association, which can handle both random and zero number of trials and can accommodate overdispersion and zero inflation in the data.
Article
Oncology
Lillian Boe, Perri S. Vingan, Minji Kim, Kevin K. Zhang, Danielle Rochlin, Evan Matros, Carrie Stern, Jonas A. Nelson
Summary: This study provides best practice guidelines and pitfalls of regression modeling in surgical oncology research using real working examples. Through analyzing patients who underwent breast reconstruction surgery, it was found that age, marital status, and surgical technique were significantly associated with breast health scores, while body mass index, age, surgical technique, and complications were related to the likelihood of experiencing complications.
JOURNAL OF SURGICAL ONCOLOGY
(2023)
Article
Statistics & Probability
Jared S. Murray
Summary: The study introduces Bayesian additive regression trees (BART) for log-linear models such as multinomial logistic regression and count regression, addressing issues such as zero-inflation and overdispersion. New data augmentation strategies and prior distributions are developed to extend the application of BART beyond Gaussian data models, showcasing its utility with examples and a previously published dataset.
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION
(2021)
Article
Mathematics
Lu Liu, Junheng Gao, Georgia Beasley, Sin-Ho Jung
Summary: This paper discusses the standard approach of using machine learning methods to select features and build prediction models, and the issues with popular methods like LASSO and elastic net. The paper proposes a combination of standard regression methods and stepwise variable selection to overcome these issues and highlights the advantages of this method in terms of statistical significance and prediction accuracy compared to LASSO and elastic net.
Article
Environmental Sciences
Jing Xiu, Xiaoqiang Zang, Zhenggang Piao, Liang Li, Kwansoo Kim
Summary: China has achieved its 2020 Intended National Determined Contribution set out in the Paris Agreement by adopting a strategy of low-carbon economic growth. Research findings show that China has maintained a steady growth efficiency in low-carbon total factor productivity and low-carbon technological progress. The Eastern region of China has a leading advantage in low-carbon economic growth and has the potential to be the first to reach its CO2 turning point. However, China needs to strengthen policy implementation in terms of low-carbon environmental regulations.
ENVIRONMENTAL SCIENCE AND POLLUTION RESEARCH
(2023)
Article
Engineering, Environmental
Nafsika Antoniadou, Hjalte Jomo Danielsen Sorup, Jonas Wied Pedersen, Ida Buelow Gregersen, Torben Schmith, Karsten Arnbjerg-Nielsen
Summary: Extreme precipitation events can have severe negative consequences for society, the economy, and the environment, so it is important to understand when and why they occur. This study compares the performance of logistic regression and three commonly used supervised machine learning algorithms in determining whether extreme events occur locally. The results show that logistic regression performs similarly to more complex machine learning algorithms, highlighting the value of comparing different modeling approaches.
STOCHASTIC ENVIRONMENTAL RESEARCH AND RISK ASSESSMENT
(2023)
Article
Green & Sustainable Science & Technology
Jingjing Li, Xinge Rao, Xianyi Li, Sihai Guan
Summary: The bitcoin market is rapidly growing and has been recognized as a new type of gold that may replace traditional gold as a hedge against inflation and a new investment asset for financial management. This study models and predicts the performance of gold and bitcoin in investment portfolios, providing insights into optimal investment strategies and initial investment values.
Article
Computer Science, Theory & Methods
Marco Scutari, Francesca Panero, Manuel Proissl
Summary: This paper presents a general framework for estimating regression models with a user-defined level of fairness. Fairness is enforced through model selection, where a ridge penalty is chosen to control the impact of sensitive attributes. The proposed framework is mathematically simple and can be extended to various types of models and fairness definitions. Empirical evaluations show that the proposed framework outperforms other models in terms of goodness of fit and predictive accuracy at the same level of fairness.
STATISTICS AND COMPUTING
(2022)
Article
Instruments & Instrumentation
Pauline Ong, Suming Chen, Chao-Yin Tsai, Yung-Kun Chuang
Summary: The study introduced a method for optimally selecting band-pass filters to reduce spectral data dimensionality, applied in determining theanine content in oolong tea. Results indicated the method's potential for accurate prediction of analytes with cost savings in spectral acquisition.
INFRARED PHYSICS & TECHNOLOGY
(2021)
Article
Medical Laboratory Technology
Toshihiko Kobayashi, Kiyoshi Ichihara, Shuhei Goda, Isao Hidaka, Takahiro Yamasaki, Haku Ishida
Summary: A laboratory test-based regression model was developed for early detection of HCV-associated HCC with practical diagnostic accuracy. LRM4 demonstrated high accuracy in distinguishing HCC cases, even in early stages, prompting timely imaging studies. The retroactive validation scheme proved useful in evaluating diagnostic models for other neoplastic diseases.
CLINICA CHIMICA ACTA
(2021)
Article
Environmental Sciences
Sajid Hussain, Sun Hongxing, Muhammad Ali, Meer Muhammad Sajjad, Zeeshan Afzal, Sajid Ali
Summary: The study applied PS-InSAR technique to estimate slope deformation velocity for optimizing landslide susceptibility mapping in the area. Frequency ratio and logistic regression models were used for comparative assessment and predicting correlations with landslide occurrence. LR method showed superior results and in combination with PS-InSAR, an optimized susceptibility model was developed to mitigate landslide disasters and support management of development programs in the area.
GEOCARTO INTERNATIONAL
(2022)