Article
Computer Science, Information Systems
Masoud Fazlalipour Miyandoab, Parviz Nasiri, Ali M. Mosammam
Summary: Recognizing and presenting the appropriate statistical model for time series data is crucial. The Auto Regressive Fractionally Integrated Moving Average (ARFIMA) model is widely used in analyzing economic, meteorological, geographical, and financial data. Parameters of this model, as well as other time series models, are estimated by assuming a constant average. This article introduces Bayesian estimation for the fractional difference parameter (d) in the ARFIMA model, considering an appropriate prior distribution. Simulation and Akaike information criterion (AIC) demonstrate the superior performance of Bayesian estimation compared to other methods. The goodness of fit of the ARFIMA model is evaluated using Bayesian estimation of parameters with a real data set.
INFORMATION SCIENCES
(2023)
Article
Astronomy & Astrophysics
Jing Niu, Tong-Jie Zhang
Summary: This study compares the significance of the traditional combined method and Linder's joint method in constraining the density parameter QM. The results show that Linder's joint method is more significant than the traditional combined method.
PHYSICS OF THE DARK UNIVERSE
(2023)
Article
Mathematical & Computational Biology
Xiaoming Lu, Thierry Chekouo, Hua Shen, Alexander R. de Leon
Summary: In this article, a two-level copula joint model is proposed to analyze clinical data with multiple disparate continuous longitudinal outcomes and multiple event-times in the presence of competing risks. The model constructs submodels for the observed event-time and longitudinal outcomes using a copula and Gaussian copula respectively, and combines them in a joint model that incorporates conditional dependence. Linear quantile mixed models are proposed to accommodate skewed data and examine covariate effects. Bayesian framework and Markov Chain Monte Carlo sampling are used for model estimation and inference. Simulation study and analysis of clinical data on renal transplantation demonstrate the superior performance of the proposed method compared to conventional approaches.
STATISTICS IN MEDICINE
(2023)
Article
Astronomy & Astrophysics
Aishwarya Bhave, Soham Kulkarni, Shantanu Desai, P. K. Srijith
Summary: This article investigates the classification problem of Gamma-Ray Bursts (GRBs) using a two-dimensional approach combining GRB hardness and duration. The analysis reveals the existence of two or three distinct classes, depending on the dataset and the information theory criteria.
ASTROPHYSICS AND SPACE SCIENCE
(2022)
Article
Mathematics
Nitzan Cohen, Yakir Berchenko
Summary: This article proposes a new approach that enables the use of classic information criteria for model selection with missing data by normalizing the information criteria theory, which is found to be exponentially better in computational complexity than traditional imputation methods, leading to increased statistical efficiency.
Article
Acoustics
Okon Johnson Esua, Da-Wen Sun, Clement Kehinde Ajani, Jun-Hu Cheng, Kevin M. Keener
Summary: The combination of UPFB technology combining ultrasound with plasma functionalized buffer was more effective in inactivating Escherichia coli and Listeria monocytogenes. Non-linear models, particularly the biphasic model, demonstrated superior performance in describing the inactivation kinetics. This study provided valuable insights for the evaluation of decontamination methods.
ULTRASONICS SONOCHEMISTRY
(2022)
Article
History & Philosophy Of Science
Alireza Fatollahi
Summary: There has been a lively debate in the philosophy of science regarding predictivism, with the argument that successful predictions provide stronger evidence for a theory than mere accommodation of the same data. The author presents a strong version of predictivism, drawing on statistical results from the model selection problem. By highlighting the inverse relationship between the support for a hypothesis and the number of adjustable parameters in the model, the author argues that when data can be predicted, the model associated with the hypothesis has fewer adjustable parameters compared to when the data needs to be accommodated.
EUROPEAN JOURNAL FOR PHILOSOPHY OF SCIENCE
(2023)
Article
Multidisciplinary Sciences
Meryem Bekar Adiguzel, Mehmet Ali Cengiz
Summary: Multivariate Adaptive Regression Splines (MARS) is a non-parametric regression analysis method that is useful for model selection in high-dimensional data. It has the advantage of identifying and modeling complex, non-linear relationships between variables without requiring assumptions, as well as automatically selecting variables to simplify the model building process and prevent overfitting.
Article
Agriculture, Dairy & Animal Science
Ana Martins-Bessa, Miguel Quaresma, Belen Leiva, Ana Calado, Ander Arando, Carmen Marin, Francisco Javier Navas
Summary: This study aimed to model the evolution of body weight, testicular dimensions, and gonadosomatic index in Miranda donkeys using in vivo ultrasonography. The cubic function modeling showed the best-fitting properties for various testicular measurements, capturing higher interindividual variability compared to other functions. The Bayesian information criterion values suggest that in vivo ultrasonography may be a more efficient and accurate tool compared to three-dimensional testicular measurements in predicting the evolution of testicular volume or gonadosomatic index.
ITALIAN JOURNAL OF ANIMAL SCIENCE
(2021)
Article
Physics, Multidisciplinary
Luca Spolladore, Michela Gelfusa, Riccardo Rossi, Andrea Murari
Summary: The study suggests that in complex systems and highly correlated variables, the proposed versions of model selection criteria outperform the traditional ones.
Article
Environmental Sciences
Dmitrii Shadrin, Artyom Nikitin, Polina Tregubova, Vera Terekhova, Raghavendra Jana, Sergey Matveev, Maria Pukalchik
Summary: This study emphasizes the importance of sustainable environmental management and proposes a method for constructing a water quality index and spatial prediction map using machine learning techniques. The approach was validated on actual groundwater quality data, showing better performance compared to traditional Kriging models and highlighting its potential for predicting the spatial distribution of natural resource properties.
Article
Energy & Fuels
Ge Xinmin, Xue Zong'an, Zhou Jun, Hu Falong, Li Jiangtao, Zhang Hengrong, Wang Shuolong, Niu Shenyuan, Zhao Ji'er
Summary: In this study, an unsupervised clustering method based on Gaussian mixture model (GMM) was developed to quantitatively analyze the type and pore structure of reservoirs using nuclear magnetic resonance (NMR) transverse relaxation (T2) spectra. The results showed that the clustering results based on GMM had good correlations with the T2 spectrum features, pore structure, and petroleum productivity. This research provides a new approach for quantitative identification of pore structure, reservoir grading, and oil and gas productivity evaluation.
PETROLEUM EXPLORATION AND DEVELOPMENT
(2022)
Article
Ecology
Jeremy D. Wilson, Nicolas Mongiardino Koch, Martin J. Ramirez
Summary: Research suggests that selecting branch lengths most correlated with a character can enhance accuracy in ancestral state estimation. Phylogenetic signal statistics have limited utility in choosing the correct branch lengths, while model-fit statistics are more accurate.
METHODS IN ECOLOGY AND EVOLUTION
(2022)
Article
Economics
Peter S. Sephton
Summary: Recent research has shown that wavelet transformations can improve the performance of unit root and stationarity tests by separating a variable's high and low frequency components. This note provides response surface estimates for an Augmented Dickey-Fuller type wavelet test that includes a Fourier term, allowing for smooth breaks in the series. Applications highlight the practical benefits.
COMPUTATIONAL ECONOMICS
(2023)
Article
Mathematical & Computational Biology
Siyun Liu, Tao Yu
Summary: In this article, a method for density estimation of data with a mixture structure is proposed, which nonparametrically estimates component density functions through weighted kernel density estimation. Extensive simulation studies and real data examples demonstrate the superiority of the proposed method over existing methods in most cases.
STATISTICS IN MEDICINE
(2021)