Article
Economics
Javier Alvarez, Manuel Arellano
Summary: Likelihood-based estimators for autoregressive panel data models are developed to be consistent in the presence of time series heteroskedasticity. The empirical application suggests evidence against unit roots in individual earnings processes.
JOURNAL OF ECONOMETRICS
(2022)
Article
Management
Kien C. Tran, Mike G. Tsionas, Artem B. Prokhorov
Summary: This paper considers the estimation of a spatial autoregressive stochastic frontier model with dependencies on observed environmental factors. A two-step semiparametric procedure is developed, and estimators for various quantities are derived. Monte Carlo simulations and an empirical application demonstrate the effectiveness of the proposed methods.
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH
(2023)
Article
Statistics & Probability
Chaonan Jiang, Davide La Vecchia, Elvezio Ronchetti, Olivier Scaillet
Summary: We propose new higher-order asymptotic techniques for the Gaussian maximum likelihood estimator in a spatial panel data model, and develop saddlepoint density and tail area approximations that demonstrate good performance in density approximation and testing experiments.
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION
(2023)
Article
Computer Science, Information Systems
Yuming Lin, Zejun Xu, Yinghao Zhang, You Li, Jingwei Zhang
Summary: Cardinality estimation is a fundamental task in database query processing and optimization. Recently, deep autoregressive models have been used to obtain joint probability distributions for accurate estimation. However, the sparsity of data and error propagation challenges hinder the accuracy of estimation. To address these issues and improve accuracy, a random smoothing autoregressive cardinality estimation model (SAM-CE) is proposed, which combines random smoothing and deep autoregressive models. The proposed model achieves state-of-the-art effectiveness in cardinality estimation.
WORLD WIDE WEB-INTERNET AND WEB INFORMATION SYSTEMS
(2023)
Article
Operations Research & Management Science
Guangyuan Gao, Yanlin Shi
Summary: This paper uses the generalized autoregressive score (GAS) models to study the long memory and regime switching in the second comment. It demonstrates the robustness of both the long memory GAS (LMGAS) and Markov switching GAS (MS-GAS) models against outliers through simulation studies. It proposes an MS-LMGAS model to resolve the spurious long memory and provides empirical evidence supporting its superiority over existing models.
ANNALS OF OPERATIONS RESEARCH
(2023)
Article
Multidisciplinary Sciences
Jie Zhang, Dehui Wang, Kai Yang, Xiaogang Dong
Summary: This paper proposes a first-order random coefficient multinomial autoregressive model for complex and asymmetric finite range multi-state integer-valued time series data. The model's basic probabilistic and statistical properties are discussed, and estimators for the model parameters are derived with established asymptotic properties. Simulation studies compare different estimation methods to verify the proposed procedure, with a real example illustrating the advantages of the model.
Article
Economics
Gary Chamberlain
Summary: This study relaxes the assumption of strict exogeneity in panel data analysis to allow for lagged dependent variables and feedback from lagged dependent variables to current predictor variables. An information bound is derived for a semiparametric regression model and applied to a model with a multiplicative random effect. The study also addresses identification issues when the random effect is a vector with two or more components.
JOURNAL OF ECONOMETRICS
(2022)
Article
Engineering, Environmental
Cristobal Gallego-Castillo, Alvaro Cuerva-Tejero, Mohanad Elagamy, Oscar Lopez-Garcia, Sergio Avila-Sanchez
Summary: This article presents a novel method for determining optimal autoregressive models to reproduce a predefined target autocovariance function, utilizing flexibility and genetic algorithms to optimize the generated time series.
STOCHASTIC ENVIRONMENTAL RESEARCH AND RISK ASSESSMENT
(2022)
Article
Statistics & Probability
Ke-Ang Fu, Ting Li, Chang Ni, Wenkai He, Renshui Wu
Summary: A new conditional self-weighted M-estimator is proposed for the generalized random coefficient AR(1) model, with established asymptotic normality even when Eu2t may be infinite. Simulation experiments and a real data example are used to evaluate the performance of the theory and method in finite samples.
STATISTICAL PAPERS
(2021)
Article
Mathematics, Interdisciplinary Applications
Jana Holtmann, Michael Eid, Philip S. Santangelo, Tobias D. Kockler, Ulrich W. Ebner-Priemer
Summary: Longitudinal models for panel data analysis assume population homogeneity in studying the temporal dynamics of variables, which may be too restrictive. We propose extension of autoregressive and cross-lagged latent state-trait models to mixture distribution models. These models allow researchers to account for unobserved person heterogeneity and qualitative differences in longitudinal dynamics with limited observations per person, while considering temporal dependencies and measurement error. The models can also incorporate categorical covariates to investigate distribution of latent classes in observed groups. We illustrate the potential of these models with an application on self-esteem and affect data in patients with borderline personality disorder, an anxiety disorder, and healthy control participants. We also conduct an extensive simulation study to investigate requirements for model applicability and provide recommendations.
MULTIVARIATE BEHAVIORAL RESEARCH
(2023)
Article
Mathematics, Applied
Sergio Contreras-Espinoza, Christian Caamano-Carrillo, Javier E. Contreras-Reyes
Summary: Models with time-varying parameters, such as Generalized Autoregressive Score (GAS) models, are popular for analyzing time series data. GAS models capture the dynamic behavior of time series processes better than models with fixed parameters. This paper extends the distribution setting of GAS models to include the sinh-arcsinh (SAS) family, particularly the SAS-Gaussian and SAS-t distributions. The SAS family provides flexible distributions for modeling asymmetry. The proposed method's performance is demonstrated through simulations and an application to a fish condition dataset, with the SAS-Gaussian distribution fitting the dataset best.
JOURNAL OF COMPUTATIONAL AND APPLIED MATHEMATICS
(2023)
Article
Mathematical & Computational Biology
Junhui Park, Seung-Ho Kang
Summary: The article explores the application of hierarchical linear models and hierarchical generalized linear models in multiregional clinical trials, discussing the details of the models under different response variable distributions. Simulation studies indicate that the empirical powers of HGLM exceed those of random effects models when incorporating region-level covariates.
STATISTICS IN BIOPHARMACEUTICAL RESEARCH
(2022)
Article
Computer Science, Interdisciplinary Applications
H. R. Khorshidi, A. R. Nematollahi, T. Manouchehri
Summary: This paper extends the generalized autoregressive models to vector-valued autoregressive models, providing a flexible framework for modeling dependent data. The properties of the new model, such as stationary conditions, explicit forms of the auto-covariance function, and spectral density matrices, are investigated. Unknown parameters are estimated and compared with traditional methods. Numerical results from simulation studies are reported. Finally, the performance of the proposed models and estimation methods are discussed by fitting the traditional autoregressive model and generalized autoregressive model to a well-known bivariate time series.
JOURNAL OF STATISTICAL COMPUTATION AND SIMULATION
(2023)
Article
Statistics & Probability
Zixuan Liu, Junmo Song
Summary: This paper considers the problem of testing for normality of the two unobservable random processes in first order random coefficient autoregressive models. An information matrix based test is proposed and its limiting null distribution is derived. Simulations are conducted to evaluate the performance and characteristics of the introduced test, and a real data analysis is provided.
JOURNAL OF THE KOREAN STATISTICAL SOCIETY
(2023)
Article
Mathematics, Interdisciplinary Applications
Fien Gistelinck, Tom Loeys, Nele Flamant
Summary: This paper discusses the performance of different implementations of the multilevel autoregressive model in handling the endogeneity assumption and the initial conditions problem. The study found bias in some commonly used approaches for the autoregressive parameter.
STRUCTURAL EQUATION MODELING-A MULTIDISCIPLINARY JOURNAL
(2021)
Article
Statistics & Probability
Marta Regis, Alberto Brini, Nazanin Nooraee, Reinder Haakma, Edwin R. van den Heuvel
Summary: This paper provides an in-depth analysis of the t linear mixed model (tLMM), evaluating a direct maximum likelihood estimation method through extensive simulations and investigating identifiability properties.
COMMUNICATIONS IN STATISTICS-SIMULATION AND COMPUTATION
(2022)
Article
Statistics & Probability
Edwin R. van den Heuvel, Stephan A. W. van Driel, Zhuozhao Zhan
Summary: In this paper, a bivariate zero-inflated Poisson control chart is proposed for rare events in an industrial process. The authors address the dependency between the bivariate counts using a specific copula. However, they find that the copula provided does not meet the criteria in cases where the correlation parameter is negative, and the upper control limit for the sum of events is not calculated correctly. This study presents proper copulas and upper control limits, and discusses the choice of parameter estimators. The results are illustrated using simulations.
COMMUNICATIONS IN STATISTICS-THEORY AND METHODS
(2022)
Article
Statistics & Probability
Edwin van den Heuvel, Zhuozhao Zhan
Summary: Pearson's correlation coefficient is considered a measure of linear association between bivariate random variables X and Y. However, for nonlinear monotonic associations, alternative measures like Spearman's rank and Kendall's tau correlation coefficients are more appropriate. Existing views that rule out Pearson's correlation coefficient for measuring nonlinear monotonic associations are incorrect.
AMERICAN STATISTICIAN
(2022)
Article
Health Care Sciences & Services
Francesco Ungolo, Edwin R. van den Heuvel
Summary: This study introduces a joint model for addressing the problem of informative censoring in survival studies, utilizing latent variables and a fully Bayesian approach. Results suggest that ignoring informative censoring may lead to serious biases.
STATISTICAL METHODS IN MEDICAL RESEARCH
(2022)
Article
Mathematical & Computational Biology
Mona Emampour, Pieta C. IJzerman-Boon, Md Abu Manju, Edwin R. van den Heuvel
Summary: The European and United States Pharmacopoeia require a noninferiority study to be conducted when considering an alternative microbiological method. By assuming the accuracy of the alternative method is homogeneous across microorganisms, a joint statistical analysis can help reduce the required sample size. This study provides a test statistic for noninferiority, an optimal spiking experiment, and a sample size calculation approach based on mild modeling assumptions of microorganism-specific detection proportions.
STATISTICS IN BIOPHARMACEUTICAL RESEARCH
(2023)
Meeting Abstract
Cardiac & Cardiovascular Systems
Edwin R. van den Heuvel, Robert P. Agans, William D. Kalsbeek, Suzanne E. Judd, Vasan S. Ramachandran
Article
Surgery
Dieuwke C. Broekstra, Rosanne Lanting, Paul M. N. Werker, Edwin R. van den Heuvel
Summary: This study found that progression of Dupuytren disease is mainly associated with disease extent, particularly on the small finger side of the hand. However, none of the traditional risk and diathesis factors were found to be associated with progression.
PLASTIC AND RECONSTRUCTIVE SURGERY
(2022)
Article
Mathematical & Computational Biology
Osama Almalik, Zhuozhao Zhan, Edwin R. van den Heuvel
Summary: The DerSimonian-Laird weighted average method is widely used for overall effect size estimation in meta-analysis, but it underestimates the standard error when effect sizes are heterogeneous. Several alternative approaches, such as profile likelihood and corrections factors, have been proposed to improve the estimation. This study presents a bivariate likelihood approach that jointly estimates the overall effect size, between-study variance, and potentially heteroskedastic within-study variances. Simulation results show that the proposed method has better or similar performance compared to other approaches.
BIOMETRICAL JOURNAL
(2022)
Article
Infectious Diseases
Anette Veringa, Roger J. Bruggemann, Lambert F. R. Span, Bart J. Biemond, Mark G. J. de Boer, Edwin R. van den Heuvel, Saskia K. Klein, Doris Kraemer, Monique C. Minnema, Niek H. J. Prakken, Bart J. A. Rijnders, Jesse J. Swen, Paul E. Verweij, Marielle J. Wondergem, Paula F. Ypma, Nicole Blijlevens, Jos G. W. Kosterink, Tjip S. van der Werf, Jan-Willem C. Alffenaar
Summary: This study investigated whether TDM-guided voriconazole treatment is superior to standard treatment for invasive aspergillosis. The results showed no significant difference in treatment outcome and adverse reactions between the TDM and non-TDM groups, but a higher proportion of voriconazole concentrations within the acceptable range were found in the TDM group.
INTERNATIONAL JOURNAL OF ANTIMICROBIAL AGENTS
(2023)
Article
Statistics & Probability
Michiel H. J. Paus, Edwin R. van den Heuvel, Marc J. M. Meddens
Summary: This study proposes a quantile-based prediction (QBP) method for binary disease classification, which selects multiple continuous biomarkers and utilizes the tail differences in biomarker distributions between cases and controls. QBP outperforms alternative methods when biomarkers predominantly show variance differences.
JOURNAL OF NONPARAMETRIC STATISTICS
(2023)
Article
Statistics & Probability
Roxana A. Ion, Chris A. J. Klaassen, Edwin R. van den Heuvel
Summary: This article proves sharp upper bounds for the probability of a standardized random variable taking on a value outside a possibly asymmetric interval around 0. Six classes of distributions for the random variable are considered, namely the general class of 'distributions', the class of 'symmetric distributions', 'concave distributions', 'unimodal distributions', 'unimodal distributions with coinciding mode and mean', and 'symmetric unimodal distributions'. The results generalize previous work by Gauss (1823), Bienayme (1853), Chebyshev (1867), and Cantelli (1928), and provide alternative proofs for some known inequalities such as Gauss inequality.
Article
Statistics & Probability
Maliheh Heidari, Md Abu Manju, Pieta C. IJzerman-Boon, Edwin R. van den Heuvel
Summary: This paper generalizes the optimal designs for Mitscherlich's function from normally distributed responses with homoscedastic variances to discrete and continuous responses with distribution functions in the exponential family. The study explores the application of this function in other distribution scenarios.
MATHEMATICAL METHODS OF STATISTICS
(2022)
Article
Pediatrics
Anouk W. J. Scholten, Zhuozhao Zhan, Hendrik J. Niemarkt, Marieke Vervoorn, Ruud W. van Leuteren, Frans H. de Jongh, Anton H. van Kaam, Edwin R. van den Heuvel, G. Jeroen Hutten
Summary: Cardiorespiratory monitoring is crucial in the NICU. Current monitoring techniques have disadvantages such as the usage of wired adhesive electrodes that may damage the skin. The Bambi(R) belt, a wireless and non-adhesive alternative, enables cardiorespiratory monitoring by measuring electrical activity of the diaphragm.
BMJ PAEDIATRICS OPEN
(2022)
Article
Medical Informatics
Ramachandran S. Vasan, Edwin van den Heuvel
Summary: The use of sex-specific and race-specific pooled cohort equations (PCEs) may result in substantially divergent cardiovascular disease risk estimates for Black and White individuals with identical risk profiles, introducing race-related variations in clinical recommendations for cardiovascular disease prevention.
LANCET DIGITAL HEALTH
(2022)