Article
Statistics & Probability
Pangpang Liu, Yichuan Zhao
Summary: In this paper, a novel smoothed empirical likelihood method is proposed for the difference of quantiles with paired samples. The method includes two estimating equations and introduces a nuisance parameter to account for the dependence between paired samples. The approach yields a limiting distribution following the standard chi (2) distribution and outperforms other methods in simulation studies.
STATISTICAL PAPERS
(2023)
Article
Mathematics
Sajid Hussain, Mahmood Ul Hassan, Muhammad Sajid Rashid, Rashid Ahmed
Summary: The study of hydrological characteristics plays a crucial role in water resources design, planning, and management. The selection of appropriate probability distributions and estimation methods are fundamental in hydrology analyses. This article proposes a new family called the 'exponentiated power alpha index generalized' (EPAIG)-G to develop various new distributions. Based on this family, a new model called the EPAIG-exponential (EPAIG-E) is developed, and its structural properties are obtained. The EPAIG-E parameters are estimated using the method of maximum likelihood (MML), and Monte Carlo simulation (MCS) is conducted to assess the model's performance using real data.
Article
Biology
Jinyuan Chang, Song Xi Chen, Cheng Yong Tang, Tong Tong Wu
Summary: High-dimensional statistical inference with general estimating equations is challenging and little explored. Two problems in this area, confidence set estimation and model specifications tests, are studied with new approaches proposed for constructing estimating equations and test statistics. Theoretical validity and promising performance of the new methods are demonstrated through numerical studies.
Article
Statistics & Probability
Matthew Reimherr, Xiao-Li Meng, Dan L. Nicolae
Summary: This paper presents a framework for quantifying the contribution of the prior to posterior inference in the presence of prior-likelihood discordance, expanding the classic notion of prior sample size in three directions. The framework is demonstrated using simulated and real data, showing the potential of quantifying the impact of a prior in specific inference problems. The discussion also touches upon conceptual and theoretical issues, such as the use of improper priors and priors with asymptotically non-vanishing influence.
JOURNAL OF THE ROYAL STATISTICAL SOCIETY SERIES B-STATISTICAL METHODOLOGY
(2021)
Article
Biology
Yukitoshi Matsushita, Taisuke Otsu
Summary: This article discusses inference problems for statistical models under alternative or nonstandard asymptotic frameworks using the perspective of jackknife empirical likelihood. The study establishes Wilks' theorem for jackknife empirical likelihood statistic and proposes a modification to recover asymptotic pivotalness under both conventional and nonstandard asymptotics, providing a unified framework for investigating nonstandard asymptotic problems.
Article
Statistics & Probability
Giuseppe De Luca, Jan R. Magnus, Franco Peracchi
Summary: The study derives explicit expressions for moments and quantiles of the posterior distribution of the location parameter eta in the normal location model with Laplace prior, and then uses these results to approximate the posterior distribution of sums of independent copies of eta.
COMMUNICATIONS IN STATISTICS-THEORY AND METHODS
(2021)
Article
Mathematics, Applied
Guili Liao, Liang Peng, Rongmao Zhang
Summary: The paper proposes a new method for testing the equality of covariance matrices, which is not affected by dimension divergence and shows stable performance and higher power in simulation studies. The method is further illustrated using a breast cancer dataset.
SCIENCE CHINA-MATHEMATICS
(2021)
Article
Management
Sareh Nabi, Houssam Nassif, Joseph Hong, Hamed Mamani, Guido Imbens
Summary: Adding domain knowledge as a prior in learning systems has been shown to improve results. This study proposes a hierarchical empirical Bayes approach that addresses the challenges of lacking informative priors and controlling parameter learning rates. By learning empirical meta-priors and decoupling learning rates of different feature groups, the method improves performance and convergence time.
MANAGEMENT SCIENCE
(2022)
Article
Computer Science, Information Systems
Rory Mitchell, Eibe Frank, Geoffrey Holmes
Summary: This study evaluates lightweight moment estimators for single-pass quantile approximation, demonstrating how stable summation formulas can offset numerical precision issues and providing a GPU-accelerated quantile approximation algorithm. Experiments show that moment-based quantile approximation methods are reliable and high-performing in terms of efficient summarization.
ACM TRANSACTIONS ON DATABASE SYSTEMS
(2021)
Review
Mathematics, Interdisciplinary Applications
Nicole A. Lazar
Summary: Empirical likelihood is a popular nonparametric analog of parametric likelihood, inheriting many large-sample properties. This article reviews the development of empirical likelihood, including computation, connections with other likelihood-type quantities, and future research directions.
ANNUAL REVIEW OF STATISTICS AND ITS APPLICATION, VOL 8, 2021
(2021)
Article
Computer Science, Theory & Methods
Yiming Liu, Shaochen Wang, Wang Zhou
Summary: This paper introduces a novel Jackknife Empirical Likelihood approach for both smooth and unsmooth statistical functionals. It considers the estimation of unknown statistical functionals using the general delete-d jackknife and a subsampling method to reduce computational burden. The paper also investigates the statistical inference issues and asymptotic properties of the proposed method, and provides application examples and finite sample simulation studies to support its superiority.
STATISTICS AND COMPUTING
(2023)
Article
Mathematics
Amor Keziou, Aida Toma
Summary: This paper introduces a robust empirical likelihood estimator for semiparametric moment condition models, obtained by minimizing the modified Kullback-Leibler divergence using truncated orthogonality functions. The robustness and consistency of the new estimator are proven, and its performance is illustrated through examples based on Monte Carlo simulations.
Article
Mathematics, Interdisciplinary Applications
Yan Liu, Mingyang Ren, Sanguo Zhang
Summary: This paper introduces an empirical likelihood method for testing regression coefficients in high dimensional partially linear models, which shows good performance in controlling the error rate and power. The method is validated through simulations and analysis of Skin Cutaneous Melanoma data.
JOURNAL OF SYSTEMS SCIENCE & COMPLEXITY
(2021)
Article
Computer Science, Information Systems
Yang Li, Shijie Guo, Zhongxue Gan
Summary: This study introduces the application of probabilistic dynamics and probabilistic inference neural network (PINN) in reinforcement learning. The PINN approach shows remarkable effectiveness in data regression and policy learning frameworks, outperforming the currently prevailing methods. The experimental results also demonstrate high data efficiency and superior policy performance of the PINN-based policy learning framework in control tasks.
INFORMATION SCIENCES
(2022)
Article
Hospitality, Leisure, Sport & Tourism
Haoyu Shu, Jianping Zha, Ting Tan, Cheng Li
Summary: This study examines the relationship between high-speed railway (HSR) and tourism efficiency using a difference-in-differences (DID) model and a global data envelopment analysis (DEA) decomposition analytical framework. The findings reveal that the opening of HSR can contribute to efficiency growth, but it can also negatively affect efficiency growth by widening the technological gap between regions.
CURRENT ISSUES IN TOURISM
(2023)
Article
Cell Biology
Xiao-zheng Du, Chun-ling Bao, Gui-rong Dong, Xu-ming Yang
NEURAL REGENERATION RESEARCH
(2016)
Article
Biology
Ji-Yeon Yang, Xuming He
Article
Metallurgy & Metallurgical Engineering
Xiyun Yang, Michael S. Moats, Jan D. Miller, Xuming Wang, Xichang Shi, Hui Xu