Article
Public, Environmental & Occupational Health
Daniel E. Sack, Bryan E. Shepherd, Carolyn M. Audet, Caroline De Schacht, Lauren R. Samuels
Summary: Inverse probability weighting (IPW) is a well-established method used in observational studies to control for confounding. This study extended the use of IPW to analyze continuous exposures, specifically in cases of quasicontinuous exposures. Different approaches were assessed and compared using simulations and cluster-randomized clinical trial data. The results showed that certain methods, such as covariate balancing generalized propensity scores (CBGPS) and nonparametric covariate balancing generalized propensity scores (npCBGPS), achieved excellent covariate balance and lowest bias, while quantile binning (QB) and cumulative probability model (CPM) had the lowest mean squared error. The IPW approaches were also successfully applied to assess the influence of session attendance on postpartum contraceptive uptake in a partners-based clustered intervention in Mozambique.
AMERICAN JOURNAL OF EPIDEMIOLOGY
(2023)
Article
Engineering, Electrical & Electronic
Prabhu Babu, Petre Stoica
Summary: In this study, the problem of smoothed nonparametric spectral estimation via cepstrum thresholding is revisited. The cepstrum thresholding problem is formulated as a multiple hypothesis testing problem, and the false discovery rate (FDR) and familywise error rate (FER) procedures are used to threshold the cepstral coefficients. The FDR and FER approaches are compared with a previously proposed individual hypothesis testing approach, and it is shown that cepstrum thresholding based on FDR and FER can yield spectral estimates with lower mean square error (MSE).
IEEE SIGNAL PROCESSING LETTERS
(2022)
Article
Statistics & Probability
Tommaso Lando
Summary: This paper proposes a nonparametric test for detecting violations of the increasing hazard rate property, which is based on the distance between the empirical distribution function and a shape-constrained estimator. The test is consistent, and the power function is evaluated through simulations in critical cases.
STATISTICS & PROBABILITY LETTERS
(2023)
Article
Statistics & Probability
Anqi Zhao, Youjin Lee, Dylan S. Small, Bikram Karmakar
Summary: This article proposes a balanced block design method to offset the possible violation of the exclusion restriction by balancing the instruments, in order to construct approximate evidence factors. It also introduces a novel stratification method for using multiple nested candidate instruments.
ANNALS OF STATISTICS
(2022)
Article
Computer Science, Hardware & Architecture
Hongyi Pan, Diaa Badawi, Ahmet Enis Cetin
Summary: This article proposes using binary block Walsh-Hadamard transform (WHT) instead of Fourier transform for convolutions in deep neural networks. By replacing some convolution layers with WHT-based binary layers, the number of trainable parameters can be significantly reduced with negligible loss in accuracy. The experimental results also show that the 2D-FWHT layer runs much faster and consumes less RAM compared to the regular 3 x 3 convolution layer.
ACM TRANSACTIONS ON EMBEDDED COMPUTING SYSTEMS
(2022)
Article
Mathematics, Applied
Luai Al-Labadi, Mohammed Hamlili, Anna Ly
Summary: This paper presents a novel methodology for computing varentropy and varextropy, drawing inspiration from Bayesian nonparametric methods. The approach is implemented using a computational algorithm in R and its effectiveness is demonstrated across various examples. Furthermore, these new estimators are applied to test uniformity in data.
Article
Astronomy & Astrophysics
Niccolo Muttoni, Danny Laghi, Nicola Tamanini, Sylvain Marsat, David Izquierdo-Villalba
Summary: This study investigates the use of binary black holes detected by 3G interferometers as dark sirens to extract cosmological parameters. Results show that a network of ET and two CEs can provide promising constraints on H-0 and Omega(m) within one year.
Article
Automation & Control Systems
Molei Liu, Yin Xia, Kelly Cho, Tianxi Cai
Summary: Identifying informative predictors in a high-dimensional regression model is crucial for association analysis and predictive modeling. Signal detection often fails in high-dimensional settings due to limited sample size, but meta-analyzing multiple studies can help improve power. Integrative analysis of high-dimensional data from different studies poses challenges, especially with data sharing constraints, but a new method called DSILT is proposed for signal detection without sharing individual-level data. The method incorporates proper estimation and debiasing procedures to construct test statistics for specific covariates, and a multiple testing procedure is developed to control false discovery rate and identify significant effects. Simulation studies show the proposed testing procedure performs well in controlling false discoveries and achieving power.
JOURNAL OF MACHINE LEARNING RESEARCH
(2021)
Article
Business, Finance
Simon C. Smith
Summary: Bayesian methods can avoid significance threshold correction in multiple testing, and controlling the Type-S error rate yields more reliable inferences. Our study identifies two breaks in the sample period of 1980-2018, with different characteristics selected after each break. In a portfolio application, the method outperforms benchmark methods in generating higher Sharpe ratios after transaction costs.
INTERNATIONAL REVIEW OF FINANCIAL ANALYSIS
(2022)
Article
Engineering, Mechanical
Hongya Qu, Tiantian Li, Ruilong Wang, Jianzhong Li, Zhongguo Guan, Genda Chen
Summary: This paper introduces adaptive wavelet analysis (AWT) as a preprocessing method to obtain a clearer, smoother, and more accurate time-frequency representation. Optimized analytical mode decomposition (AMD) is utilized for signal component extraction, successfully decomposing signals and performing system identification in a shake table test of a 1/20-scale cable-stayed bridge model.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2021)
Article
Astronomy & Astrophysics
Anarya Ray, Michael Camilo, Jolien Creighton, Shaon Ghosh, Soichiro Morisaki
Summary: Bayesian hierarchical inference is developed to improve understanding of neutron star structure and nuclear force by inferring phenomenological parametrized neutron star equations of state from gravitational wave observations of binary neutron star mergers. A novel algorithm is introduced that utilizes a priori knowledge of neutron star physics and reuses single-event parameter estimation samples to reduce computational cost. The method is tested on real and simulated data, producing consistent results. The fast analysis scheme also allows for studying the variability of the equations of state constraints with different event properties.
Article
Optics
Lina Zhou, Yin Xiao, Zilan Pan, Yonggui Cao, Wen Chen
Summary: A new scheme based on SIMO and AOHs is proposed in this paper, which can retrieve a large number of different secret images from one single host image during optical retrieval, while reducing optical implementation complexity.
Article
Statistics & Probability
Siqi Xiang, Wan Zhang, Kai Zhang, J. S. Marron
Summary: Binary expansion testing (BET) is a powerful tool for detecting interesting nonlinear dependence among variables in large-scale data analysis. However, the commonly used Bonferroni adjusted p-values may be too conservative in determining the significant testing pairs. This paper introduces a novel contribution of applying extreme value theory analysis to BET, proposing a potentially powerful new significance threshold for the maximal BET z-statistics.
SANKHYA-SERIES A-MATHEMATICAL STATISTICS AND PROBABILITY
(2023)
Article
Environmental Sciences
Cristina Prieto, Dmitri Kavetski, Nataliya Le Vine, Cesar Alvarez, Raul Medina
Summary: In hydrological modeling, a statistical hypothesis-testing perspective on model identification challenge is presented. A mechanism identification framework is proposed, combining Bayesian estimation, test statistic, and flexible modeling framework. The method is reliable in identifying dominant mechanisms, but statistical power decreases as data/model errors increase. Insights on process identifiability are reported, and the method is expected to contribute to improving model identification in hydrology.
WATER RESOURCES RESEARCH
(2021)
Article
Engineering, Electrical & Electronic
Martin Goelz, Abdelhak M. Zoubir, Visa Koivunen
Summary: In this study, a general framework based on multiple hypothesis testing is developed to identify regions with spatially interesting, different or adversarial behavior. The framework utilizes a discrete spatial grid and a large-scale sensor network to acquire measurements, and involves estimating local false discovery rates and interpolating them to assign locations to regions associated with different hypotheses.
IEEE TRANSACTIONS ON SIGNAL AND INFORMATION PROCESSING OVER NETWORKS
(2022)
Editorial Material
Dentistry, Oral Surgery & Medicine
Matthew W. Brown, Lorne Koroluk, Ching-Chang Ko, Kai Zhang, Mengqi Chen, Tung Nguyen
AMERICAN JOURNAL OF ORTHODONTICS AND DENTOFACIAL ORTHOPEDICS
(2015)
Article
Computer Science, Information Systems
Kai Zhang
IEEE TRANSACTIONS ON INFORMATION THEORY
(2017)
Article
Neurosciences
Qunqun Yu, Benjamin B. Risk, Kai Zhang, J. S. Marron
Article
Statistics & Probability
Siliang Gong, Kai Zhang, Yufeng Liu
JOURNAL OF MULTIVARIATE ANALYSIS
(2018)
Article
Statistics & Probability
Daniel McCarthy, Kai Zhang, Lawrence D. Brown, Richard Berk, Andreas Buja, Edward George, Linda Zhao
Article
Statistics & Probability
Kelly Bodwin, Kai Zhang, Andrew Nobel
ANNALS OF APPLIED STATISTICS
(2018)
Article
Statistics & Probability
Andreas Buja, Lawrence Brown, Richard Berk, Edward George, Emil Pitkin, Mikhail Traskin, Kai Zhang, Linda Zhao
STATISTICAL SCIENCE
(2019)
Editorial Material
Statistics & Probability
Kentaro Hoffman, Jan Hannig, Kai Zhang
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION
(2021)
Article
Biology
D. Lee, H. El-Zaatari, M. R. Kosorok, X. Li, K. Zhang
Article
Economics
Jialu Li, Wan Zhang, Peiyao Wang, Qizhai Li, Kai Zhang, Yufeng Liu
Summary: In this article, a novel nonparametric resolution-wise regression procedure is proposed to estimate the distribution of the response by decomposing and modeling the information of the response and predictors. Simulations and a real estate valuation dataset demonstrate the effectiveness of the proposed method.
JOURNAL OF BUSINESS & ECONOMIC STATISTICS
(2023)
Article
Computer Science, Interdisciplinary Applications
Hyowon An, Kai Zhang, Hannu Oja, J. S. Marron
Summary: Identification of important variables in big data is a crucial challenge. To tackle this, methods for discovering variables with non-standard univariate marginal distributions are proposed. Traditional moments-based summary statistics can be sensitive to outliers, thus L-moments are considered for robustness. However, the limitation of L-moments is addressed by proposing Gaussian Centered L-moments.
COMPUTATIONAL STATISTICS & DATA ANALYSIS
(2023)
Article
Statistics & Probability
Siliang Gong, Kai Zhang, Yufeng Liu
Summary: This study introduces a variable selection method incorporating pairwise effects in covariates and combines it with sure independence screening, showing competitive performance in terms of prediction accuracy and variable selection accuracy.
Article
Health Care Sciences & Services
Kai Zhang, Ding-Geng Chen
STATISTICAL CAUSAL INFERENCES AND THEIR APPLICATIONS IN PUBLIC HEALTH RESEARCH
(2016)
Article
Social Sciences, Mathematical Methods
Richard Berk, Lawrence Brown, Andreas Buja, Edward George, Emil Pitkin, Kai Zhang, Linda Zhao
SOCIOLOGICAL METHODS & RESEARCH
(2014)