研究园简易版 · 经典书籍 · 期刊文章 · 有原文 · 完整版(可多选筛选、搜索) · 首页
2026 年的期刊文章 · 第 11 页
共 1756 篇,按发表时间排,新的在前,每页 40 篇。这一页是 2026-08-04 到 2026-07-27。
按期刊、按主题看 →
SiM Clinical Trials Testing & Uncertainty · 发表于 2026-08-04,收录于 2026-08-07
Siyu Zhu, Yuxuan Yang, Minggang Yin, Yixin Luo, Xueqing Liang, Shijie Yu, Chongyang Duan
摘要
To avoid unnecessary resource expenditure resulting from the sample size overestimation in seamless phase II/III designs, multi‐stage early stopping may be incorporated in the phase III component. However, existing estimation methods for seamless phase II/III designs are primarily developed for two‐stage settings and do not directly accommodate designs with multi‐stage early stopping in phase III. In this paper, we develop a Score‐statistics‐based framework for point and interval estimation in this specific class of designs. The framework provides a design‐specific formulation of conditional bias adjustment, conditional median‐unbiased estimation, and Rao–Blackwellization for seamless phase II/III designs with multi‐stage early stopping. Conditional on the trial continuing to phase III, the proposed framework targets valid and efficient estimation of the treatment effect for the selected arm and is applicable to a range of common endpoint distributions. Within this framework, we formulate and evaluate several representative conditional estimation procedures: the multiple iterations‐based conditional bias‐adjusted estimator (CBAE‐MI), the single iteration‐based conditional bias‐adjusted estimator (CBAE‐SI), the conditional median unbiased estimators with the unselected treatment group effects be estimated by their maximum likelihood estimators (CMUE‐MLE) or set as zero (CMUE‐ZERO), and the Rao–Blackwellized (RB) estimator. In addition, to properly quantify uncertainty around the point estimates, we derive confidence intervals based on CMUE‐MLE, CMUE‐ZERO, and RB. Through simulation under various endpoint types and parameter configurations, we recommend employing RB for point estimation and confidence interval, as it demonstrates superior robustness and conservative coverage probability across treatment selection and trial design settings.
JASA Bayesian Methods Nonparametric & Semiparametric · 发表于 2026-08-03,收录于 2026-06-19
Not available
摘要
We develop a nonparametric Bayesian modeling approach to ordinal regression based on priors placed directly on the discrete distribution of the ordinal responses. The prior probability models are built from a structured mixture of multinomial distributions. We leverage the continuation-ratio logits representation to formulate the mixture kernel, with mixture weights defined through the logit stick-breaking process that incorporates the covariates through a linear function. The implied regression functions for the response probabilities can be expressed as weighted sums of parametric regression functions, with covariate-dependent weights. Thus, the modeling approach achieves flexible ordinal regression relationships, avoiding linearity or additivity assumptions in the covariate effects. Model flexibility is formally explored through the Kullback-Leibler support of the prior probability model. A key model feature is that the parameters for both the mixture kernel and the mixture weights can be associated with the continuation-ratio logits regression structure. Hence, an efficient and relatively easy to implement posterior simulation method can be designed, using Pólya-Gamma data augmentation. Moreover, the model is built from a conditional independence structure for category-specific parameters, which results in additional computational efficiency gains through partial parallel sampling. In addition to the general mixture structure, we study simplified model versions that incorporate covariate dependence only in the mixture kernel parameters or only in the mixture weights. For all proposed models, we discuss approaches to prior specification and develop Markov chain Monte Carlo methods for posterior simulation. The methodology is illustrated with several synthetic and real data examples. *(abstract via OpenAlex)*
JASA Other · 发表于 2026-08-03,收录于 2026-07-24
Not available
摘要
Dynamic decision-making in rapidly evolving research domains, including marketing, finance, and pharmaceutical development, presents a significant challenge. Researchers frequently confront the need for real-time action within a doubly sequential framework characterized by the continuous influx of high-volume data streams and the intermittent arrival of novel tasks. This calls for the development and implementation of new online inference protocols capable of handling both the continuous processing of incoming information and the efficient allocation of resources to address emerging priorities. We introduce a novel class of Safe and Always-Valid Alpha-investing (SAVA) rules that leverages powerful tools including always valid p-values, e-processes, and online false discovery rate methods. The SAVA algorithm effectively integrates information across all tasks, mitigates the alpha-death problem, and controls the false selection rate (FSR) at all decision points. We validate the efficacy of the SAVA framework through rigorous theoretical analysis and extensive numerical experiments. Our results demonstrate that SAVA not only offers effective control of the FSR but also significantly improves statistical power compared to traditional online testing approaches. *(abstract via OpenAlex)*
SiM Genetics & Omics Bayesian Methods · 发表于 2026-08-03,收录于 2026-08-07
Kevin McGregor, Todd Parsons, Elinor Simons, James Scott, Anita Kozyrskyj, Christopher Quince
摘要
The human microbiome comprises the microorganisms that inhabit the various locales of the human body and plays a vital role in human health. The composition of a microbial population is often quantified through measures of species diversity, which summarize the number of species along with their relative abundances into a single value. In a finite microbiome sample, there will be species missing from the target population, which will affect the diversity estimates. We employ a model based on the hierarchical Pitman–Yor (HPY) process to model the species abundance distributions over multiple populations. The model parameters are estimated using a Gibbs sampler. We also derive estimates of species diversity, conditional and unconditional on the observed data, as a function of the HPY parameters. Finally, we derive a general formula for the Hill numbers in the HPY context. We show that the Gibbs sampler for the HPY model performs well in simulations. We also show that the conditional estimates of diversity from the HPY model improve over naïve estimates when species are missing. Similarly, the conditional HPY estimates tend to perform better than the naïve estimates especially when the number of individuals sampled from a population is small. Finally, we illustrate results of applying the HPY model in an infant gut microbiome dataset." (Wiley's article page was blocked; retrieved from the PMC mirror, PMC13433823.) *Publication/online date: 3 August 2026*
JASA Bayesian Methods · 发表于 2026-08-03,收录于 2026-08-07
Estevão Prado, Christopher Nemeth, Chris Sherlock
摘要
The Metropolis–Hastings (MH) algorithm is one of the most widely used Markov Chain Monte Carlo schemes for generating samples from Bayesian posterior distributions. The algorithm is asymptotically exact, flexible and easy to implement. However, in the context of Bayesian inference for large datasets, evaluating the likelihood on the full data for thousands of iterations until convergence can be prohibitively expensive. This paper introduces a new subsample MH algorithm that satisfies detailed balance with respect to the target posterior and utilises control variates to enable exact, efficient Bayesian inference on datasets with large numbers of observations. Through theoretical results, simulation experiments and real-world applications on certain generalised linear models, we demonstrate that our method requires substantially smaller subsamples and is computationally more efficient than the standard MH algorithm and other exact subsample MH algorithms. *(abstract via OpenAlex)*
JASA Nonparametric & Semiparametric · 发表于 2026-08-03,收录于 2026-07-24
Not available
摘要
We consider robust conditional mean learning in the mathematical framework of reproducing kernel Hilbert spaces, assuming the errors can be heavy-tailed and at the same time some responses are subject to arbitrary contamination. Simultaneous robustness to both heavy tails and contamination are of interest. When there is no contamination, we establish rates when the (1+θ)-moment ( θ>0) of the error exists, which matches the minimax rate for the least squares regression as soon as θ≥1, although the existing upper bounds for the least squares regression usually assume sub-Gaussian or sub-exponential error distribution. The existence of contamination adds an additional term to the bound. We also consider learning using random features, which is a computationally efficient kernel approximation method to relieve the computational burden of kernel-based learning. We show the same rates can be achieved with the number of random features much smaller than the sample size, extending previous results on random features for least squares kernel regression. *(abstract via OpenAlex)*
JASA Networks & Graphs High-Dimensional · 发表于 2026-08-03,收录于 2026-06-19
Not available
摘要
Glioblastoma multiforme (GBM) is a highly aggressive brain cancer with largely ineffective treatment. It is imperative to explore more effective therapies, such as gene-based treatments. For co-expression QTL studies of GBM, we develop fairness-aware Gaussian graphical regression models (Fair RegGGMs), which can determine how genetic variants modulate subject-level gene networks, and recover both population-level and subject-level gene graphs, while ensuring that the developed learning and inference neither propagate biases nor amplify disparities. We introduce pairwise graph disparity risk to quantify fairness and propose fairness-aware multi-task learning (Fair-MTL) via a cross-task group sparsity penalty, within-task element-wise sparsity penalty, and pairwise fairness regularization. We also develop a projected Fair-SAGE debiasing method for statistical inference. In Fair-MTL, we strive for a balance among group-/element-wise sparsity of graphical network structures, fairness across different subgroups, and statistical effectiveness of RegGGMs. For the ultrahigh-dimensional overparameterized models, an efficient nonsmooth fair multi-objective optimization algorithm (Fair-MOO) is developed. Both Fair-MOO and the projected Fair-SAGE dramatically reduce computational costs. Under a general dependence structure, the nonasymptotic l2 convergence rate of Fair-MTL, asymptotic normality of Fair-SAGE debiased estimator, and Pareto optimality of Fair-MOO algorithm are established. The simulation study and brain co-expression QTL analysis confirm the fairness and effectiveness of our Fair RegGGMs and provide valuable insights for the gene graph of GBM. *(abstract via OpenAlex)*
JASA Diagnostics & Prediction Bayesian Methods Clinical Trials · 发表于 2026-08-03,收录于 2026-08-07
Li Wang, Yujie Zhao, Ying Yuan, Peter F. Thall, Ruitao Lin
摘要
Assessing the long-term benefits of new treatments can be expensive and time-consuming, particularly in disease areas with unmet medical needs. While platform trials enable the evaluation of multiple interventions simultaneously, they currently cannot assess studies involving multiple agents and doses or utilize longitudinal biomarkers in decision-making. We propose a Bayesian biomarker-assisted platform design that offers a unified framework for evaluating multiple investigational agents and their doses in a multi-stage, randomized controlled trial. The design streamlines the drug evaluation process and decreases development costs by including proof-of-concept, futility and superiority monitoring, and dose optimization in a single trial, while avoiding over-allocating patients to a shared placebo or active control arm. To facilitate making real-time interim group sequential decisions, temporarily unobserved long-term responses are estimated from longitudinal biomarker measurements. Design parameters and the maximum sample size are fine-tuned to achieve good frequentist properties. The proposed design is illustrated by a trial of three targeted agents for systemic lupus erythematosus, evaluated by their 24-week response rates. Extensive simulations show that the proposed design compares favorably to several conventional platform designs. *(abstract via OpenAlex)*
AoS Nonparametric & Semiparametric · 发表于 2026-08-01
Matias D. Cattaneo, Yingjie Feng, Boris Shigida
摘要
This paper presents uniform estimation and inference theory for a large class of nonparametric partitioning-based M-estimators. The main theoretical results include: (i) uniform consistency for convex and nonconvex objective functions; (ii) rate-optimal uniform Bahadur representations; (iii) rate-optimal uniform (and mean square) convergence rates; (iv) valid strong approximations and feasible uniform inference methods; and (v) extensions to functional transformations of underlying estimators. Uniformity is established over both the evaluation point of the nonparametric functional parameter and a Euclidean parameter indexing the class of loss functions. The results also account explicitly for the smoothness degree of the loss function (if any), and allow for a possibly nonidentity (inverse) link function. We illustrate the theoretical and methodological results in four examples: quantile regression, distribution regression, Lp regression, and logistic regression. Many other possibly nonsmooth, nonlinear, generalized, robust M-estimation settings are covered by our results. We provide detailed comparisons with the existing literature and demonstrate substantive improvements: we achieve the best (in some cases optimal) known results under improved (in some cases minimal) requirements in terms of regularity conditions and side rate restrictions. The Supplementary Material reports complementary technical results that may be of independent interest, including a novel uniform strong approximation result based on Yurinskii’s coupling.
AoS Optimization & Learning Theory · 发表于 2026-08-01
Neil K. Chada, Benedict Leimkuhler, Daniel Paulin, Peter A. Whalley
摘要
We present an unbiased method for Bayesian posterior means based on kinetic Langevin dynamics that combines advanced splitting methods with enhanced gradient approximations. Our approach avoids Metropolis correction by coupling Markov chains at different discretization levels in a multilevel Monte Carlo approach. Theoretical analysis demonstrates that our proposed estimator is unbiased, attains finite variance, and satisfies a central limit theorem. It can achieve accuracy ϵ>0 for estimating expectations of Lipschitz functions in d dimensions with O(d1/4ϵ−2) expected gradient evaluations, without assuming warm start. We exhibit similar bounds using both approximate and stochastic gradients, and our method’s computational cost is shown to scale independently of the size of the dataset. The proposed method is tested using a multinomial regression problem on the MNIST dataset and a Poisson regression model for soccer scores. Experiments indicate that the number of gradient evaluations per effective sample is independent of dimension, even when using inexact gradients. For product distributions, we give dimension-independent variance bounds. Our results demonstrate that in large-scale applications, the unbiased algorithm we present can be 2–3 orders of magnitude more efficient than the “gold-standard” randomized Hamiltonian Monte Carlo.
AoS Machine Learning Econometrics & Economics · 发表于 2026-08-01,收录于 2026-08-20
Kaizhao Liu, Qi Long, Zhekun Shi, Weijie Su, Jiancong Xiao
摘要
(verbatim from arXiv:2503.10990v2)*: "Aligning large language models (LLMs) with diverse human preferences is critical for ensuring fairness and informed outcomes when deploying these models for decision-making. In this paper, we seek to uncover fundamental statistical limits concerning aligning LLMs with human preferences, with a focus on the probabilistic representation of human preferences and the preservation of diverse preferences in aligned LLMs. We first show that human preferences can be represented by a reward model if and only if the preference among LLM-generated responses is free of any Condorcet cycle. Moreover, we prove that Condorcet cycles exist with probability converging to one exponentially fast under a general probabilistic preference model called the Luce model, thereby demonstrating the impossibility of fully aligning human preferences using reward-based approaches such as reinforcement learning from human feedback. Next, we explore the conditions under which LLMs would employ mixed strategies—meaning they do not collapse to a single response—when aligned in the limit using a non-reward-based approach, such as Nash learning from human feedback. We identify a necessary and sufficient condition for mixed strategies: the absence of a response that is preferred over all others by a majority. As a blessing, we prove that this condition holds with high probability under the Luce model, thereby highlighting the statistical possibility of preserving minority preferences without explicit regularization in aligning LLMs.
AoS Causal Inference Clinical Trials Survival Analysis · 发表于 2026-08-01,收录于 2026-08-20
Bo Liu, Chengxin Yang, Fan Li
摘要
*(arXiv preprint abstract, arXiv:2501.11181, verbatim; the journal version may differ; preprint title: "Sample size and power calculations for causal inference of observational studies")* This paper investigates the theoretical foundation and develops analytical formulas for sample size and power calculations for causal inference with observational data. By analyzing the variance of an inverse probability weighting estimator of the average treatment effect, we decompose the power calculation into three components: propensity score distribution, potential outcome distribution, and their correlation. We show that to determine the minimal sample size of an observational study, in addition to the standard inputs in the power calculation of randomized trials, it is sufficient to have two parameters, which quantify the strength of the confounder-treatment and the confounder-outcome association, respectively. For the former, we propose using the Bhattacharyya coefficient, which measures the covariate overlap and, together with the treatment proportion, leads to a uniquely identifiable and easily computable propensity score distribution. For the latter, we propose a sensitivity parameter bounded by the R-squared statistic of the regression of the outcome on covariates. Our procedure relies on a parametric propensity score model and a semiparametric restricted mean outcome model, but does not require distributional assumptions on the multivariate covariates. We develop an associated R package PSpower and an online calculator.
AoS High-Dimensional · 发表于 2026-08-01
Curtis Grant, Aukosh Jagannath, Justin Ko
摘要
We develop a pseudo-likelihood theory for rank one matrix estimation problems in the high-dimensional limit. We prove a variational principle for the limiting pseudo-maximum likelihood, which also characterizes the performance of the corresponding pseudo-maximum likelihood estimator. We show that this variational principle is universal and depends only on four parameters determined by the corresponding null model. Through this universality, we introduce a notion of equivalence for estimation problems of this type and, in particular, show that a broad class of estimation tasks, including community detection, sparse submatrix detection and nonlinear spiked matrix models, are equivalent to spiked matrix models. As an application, we obtain a complete description of the performance of the least-squares (or “best rank one”) estimator for any rank one matrix estimation problem.
AoS Bayesian Methods · 发表于 2026-08-01
Shenggang Hu, Louis Aslett, Hongsheng Dai, Murray Pollock, Gareth O. Roberts
摘要
In recent years, differential privacy has been adopted by tech-companies and governmental agencies as the standard for measuring privacy in algorithms. In this article, we study differential privacy in Bayesian posterior sampling settings. We begin by considering differential privacy in the most common privatisation setting in which Laplace or Gaussian noise is injected into the output. In an effort to achieve better differential privacy, we consider adopting Huber’s contamination model for use within privacy settings, and replace at random data points with samples from a heavy-tailed distribution (instead of injecting noise into the output). We derive bounds for the differential privacy level (ϵ,δ) of our approach, without requiring bounded observation and parameter spaces, a restriction commonly imposed in the literature. We further consider for our approach the effect of sample size on the privacy level and the rate at which (ϵ,δ) converges to zero. Asymptotically, our contamination approach is fully private with no information loss. We also provide examples of inference models for which our approach applies, with theoretical convergence rate analysis and simulation studies.
AoS Causal Inference · 发表于 2026-08-01,收录于 2026-08-20
Yuehao Bai, Jizhou Liu, Azeem Shaikh, Max Tabord-Meehan
摘要
(verbatim from arXiv:2307.15181 — the IMS Future Papers page posts no abstract; accepted but not yet issue-assigned, so no online date)*: "This paper studies the use of highly stratified designs for the efficient estimation of a large class of treatment effect parameters that arise in the analysis of experiments. By a 'highly stratified' design, we mean experiments in which units are divided into blocks of a fixed size and a proportion within each block is assigned to a binary treatment uniformly at random. The class of parameters considered are those that can be expressed as the solution to a set of moment conditions constructed using a known function of the observed data. They include, among other things, average treatment effects, quantile treatment effects, and local average treatment effects as well as the counterparts to these quantities in experiments in which the unit is itself a cluster. In this setting, we establish three results. First, we show that under a highly stratified design, the naïve method of moments estimator achieves the same asymptotic variance as what could typically be attained under alternative treatment assignment mechanisms only through ex post covariate adjustment. Second, we argue that the naïve method of moments estimator under a highly stratified design is asymptotically efficient by deriving a lower bound on the asymptotic variance of regular estimators of the parameter of interest in the form of a convolution theorem. In this sense, highly stratified experiments are attractive because they lead to efficient estimators of treatment effect parameters 'by design.' Finally, we strengthen this conclusion by establishing conditions under which a 'fast-balancing' property of highly stratified designs is in fact necessary for the naïve method of moments estimator to attain the efficiency bound.
AoS Testing & Uncertainty · 发表于 2026-08-01
Nick W. Koning
摘要
We study e-values for quantifying evidence against exchangeability and general invariance of a random variable under a compact group. We start by characterizing such e-values, and explaining how they nest traditional group invariance tests as a special case. We show they can be easily designed for an arbitrary test statistic, and computed through Monte Carlo sampling. We prove a result that characterizes optimal e-values for group invariance against optimality targets that satisfy a mild orbit-wise decomposition property. We apply this to design expected-utility-optimal e-values for group invariance, which include both Neyman–Pearson-optimal tests and log-optimal e-values. Moreover, we generalize the notion of rank- and sign-based testing to compact groups, by using a representative inversion kernel. In addition, we characterize e-processes for group invariance for arbitrary filtrations, and provide tools to construct them. We also describe test martingales under a natural filtration, which are simpler to construct. Peeking beyond compact groups, we encounter e-values and e-processes based on ergodic theorems. These nest e-processes based on de Finetti’s theorem for testing exchangeability.
AoS High-Dimensional Latent Variables & Mixtures · 发表于 2026-08-01
Gérard Ben Arous, Reza Gheissari, Jiaoyang Huang, Aukosh Jagannath
摘要
We study the local geometry of empirical risks in high dimensions via the spectral theory of their Hessian and information matrices. We focus on settings where the data, (Yℓ)ℓ=1n∈Rd, are i.i.d. draws of a k-Gaussian mixture model, and the loss depends on the projection of the data into a fixed number of vectors, namely x⊤Y, where x∈Rd×C are the parameters, and C need not equal k. This setting captures a broad class of problems such as classification by one and two-layer networks and regression on multi-index models. We provide exact formulas for the limits of the empirical spectral distribution and outlier eigenvalues and eigenvectors of such matrices in the proportional asymptotics limit, where the number of samples and dimension n,d→∞ and n/d=ϕ∈(0,∞). These limits depend on the parameters x only through the summary statistic of the (C+k)×(C+k) Gram matrix of the parameters and class means, G=(x,μ)⊤(x,μ). It is known that under general conditions, when x is trained by online stochastic gradient descent, the evolution of these same summary statistics along training converges to the solution of an autonomous system of ODEs, called the effective dynamics. This enables us to connect the training dynamics to the spectral theory of these matrices generated with test data. We demonstrate our general results by analyzing the effective spectrum along the effective dynamics in the case of multiclass logistic regression. In this setting, the empirical Hessian and information matrices have substantially different spectra, each with their own static and dynamical spectral transitions.
AoS Bayesian Methods · 发表于 2026-08-01
Geerten Koers, Botond Szabó, Aad van der Vaart
摘要
We consider the recovery of an unknown function f from a noisy observation of the solution uf to a partial differential equation that can be written in the form Luf=c(f,uf), for a differential operator L that is rich enough to recover f from Luf. Examples include the time-independent Schrödinger equation Δuf=2uff, the heat equation with absorption term (∂t−Δx/2)uf=fuf and the Darcy problem ∇·(f∇uf)=h. We transform this problem into the linear inverse problem of recovering Luf under the Dirichlet boundary condition, and show that Bayesian methods with priors placed either on uf or Luf for this problem yield optimal recovery rates not only for uf, but also for f. We also derive frequentist coverage guarantees for the corresponding Bayesian credible sets. Adaptive priors are shown to yield adaptive contraction rates for f, thus eliminating the need to know the smoothness of this function. The results are illustrated by numerical experiments on synthetic data sets.
AoS Testing & Uncertainty · 发表于 2026-08-01
Christopher Blier-Wong, Ruodu Wang
摘要
The rejection threshold used for e-values and e-processes is by default set to 1/α for a guaranteed type-I error control at α, based on Markov’s and Ville’s inequalities. This threshold can be wasteful in practical applications. We discuss how this threshold can be improved under additional distributional assumptions on the e-values; some of these assumptions are naturally plausible and empirically observable, without knowing explicitly the form or model of the e-values. For small values of α, the threshold can roughly be improved (divided) by a factor of 2 for decreasing or unimodal densities, and by a factor of e for decreasing or unimodal-symmetric densities of log-transformed e-values. Moreover, we propose to use the supremum of comonotonic e-values, which is shown to preserve the type-I error guarantee. We also propose some preliminary methods to boost e-values in the e-BH procedure under some distributional assumptions while controlling the false discovery rate. Through a series of simulation studies, we demonstrate the effectiveness of our proposed methods in various testing scenarios, showing enhanced power.
AoS Networks & Graphs · 发表于 2026-08-01
Anirban Chatterjee, Soham Dan, Bhaswar B. Bhattacharya
摘要
Exchangeable random graphs, which include some of the most widely studied network models, have emerged as the mainstay of statistical network analysis in recent years. Graphons, which are the central objects in graph limit theory, provide a natural way to sample exchangeable random graphs. It is well known that network moments (motif/subgraph counts) identify a graphon (up to an isomorphism); hence, understanding the sampling distribution of subgraph counts in random graphs sampled from a graphon is pivotal for nonparametric network inference. Although there are a few results regarding the asymptotic normality of subgraph counts in graphon models, for many commonly appearing graphons this distribution is degenerate. This degeneracy phenomenon was overlooked until very recently and its consequences in network inference have remained unexplored. Toward this, we obtain the following results: We derive the joint asymptotic distribution of any finite collection of network moments in random graphs sampled from a graphon, which includes both the nondegenerate case (where the distribution is Gaussian) as well as the degenerate case (where the distribution has both Gaussian or non-Gaussian components). This provides the higher-order fluctuation theory for subgraph counts in the graphon model. Furthermore, we develop a novel multiplier bootstrap for graphons that consistently approximates the limiting distribution of the network moments (both in the Gaussian and non-Gaussian regimes). Using this and a procedure for testing degeneracy, we construct joint confidence sets for any finite collection of motif densities. This provides a general framework for statistical inference based on network moments in the graphon model. Examples and simulations are provided to validate the general theory. To illustrate the broad scope of our results, we also consider the problem of detecting global structure (i.e., testing whether the graphon is a constant function) based on small subgraphs. We propose a consistent test for this problem, invoking celebrated results on quasirandom graphs, and derive its limiting distribution both under the null and the alternative.
AoS Testing & Uncertainty · 发表于 2026-08-01
Chengzhu Huang, Anru R. Zhang
摘要
We present a new framework for statistical inference on Riemannian manifolds that achieves high-order accuracy, addressing the challenges posed by non-Euclidean parameter spaces frequently encountered in modern data science. Our approach leverages a novel and computationally efficient procedure to reach higher-order asymptotic precision. In particular, we develop a bootstrap algorithm on Riemannian manifolds that is both computationally efficient and accurate for hypothesis testing and confidence region construction. Although locational hypothesis testing can be reformulated as a standard Euclidean problem, constructing high-order accurate confidence regions necessitates careful treatment of manifold geometry. To this end, we establish high-order asymptotics under an appropriate coordinate representation induced by a second-order retraction, thereby enabling precise expansions that incorporate curvature effects. We demonstrate the versatility of this framework across various manifold settings, including spheres, the Stiefel manifold, fixed-rank matrix manifolds, and rank-one tensor manifolds; for Euclidean submanifolds, we also introduce a class of projection-like coordinate charts with strong consistency properties. Finally, numerical studies confirm the practical merits of the proposed procedure.
AoS Spatial & Time Series · 发表于 2026-08-01
Gan Yuan, Chun Yip Yau
摘要
We consider a k-dimensional multiple-regime vector threshold autoregressive model, in which the regime-switching mechanism is governed by a bivariate threshold variable. Specifically, the regimes are induced by a partition of the threshold space by an unknown number of threshold lines. Within each regime, the process follows a specific vector autoregressive (VAR) model. We formulate the model selection and parameter estimation into a minimization problem based on the Minimum Description Length (MDL) principle and estimate the number of threshold lines, parametric forms of threshold lines and VAR model parameters in each regime simultaneously. Theoretically, we show that the MDL estimators of threshold lines are n-consistent and characterize their limiting distribution. This requires novel proving techniques of introducing a new functional space G for the local MDL difference functions and establishing weak convergence results therein. Finally, we conduct some empirical studies with simulated datasets and perform real data analyses on U.S. interest rates and U.S. GNP Data.
AoS Testing & Uncertainty Causal Inference · 发表于 2026-08-01,收录于 2026-08-20
Zhaoyang Shi, Chinmoy Bhattacharjee, Krishnakumar Balasubramanian, Wolfgang Polonik
摘要
*(publisher's abstract as deposited in Crossref, verbatim)* We establish Gaussian approximation bounds for covariate and rank-matching-based Average Treatment Effect (ATE) estimators. By analyzing these estimators through the lens of stabilization theory, we employ the Malliavin–Stein method to derive our results. Our bounds precisely quantify the impact of key problem parameters, including the number of matches and treatment balance, on the accuracy of the Gaussian approximation. Additionally, we develop multiplier bootstrap procedures to estimate the limiting distribution in a fully data-driven manner, and we leverage the derived Gaussian approximation results to further obtain bootstrap approximation bounds. Our work not only introduces a novel theoretical framework for commonly used ATE estimators, but also provides data-driven methods for constructing nonasymptotically valid confidence intervals.
AoS Bayesian Methods High-Dimensional · 发表于 2026-08-01
Holden Lee, Kexin Zhang
摘要
We propose using a modified conductance-based method to study the mixing time of an important class of two-block Gibbs samplers, the data augmentation (DA) algorithm. Using this method, we prove the first non-asymptotic polynomial upper bounds on mixing times of three important DA algorithms: DA algorithms for Bayesian Probit regression (ProbitDA) and Bayesian Logit regression (LogitDA), and Bayesian Lasso Regression (LassoDA). Concretely, for ProbitDA and LogitDA, we demonstrate a tight bound that explicitly depends on the design matrix and prior covariance matrix. Under the assumption that data are independently generated from either a sub-Gaussian or log-concave distribution and properly scaled, the bound implies that with η-warm start, parameter dimension d, and sample size n, with high probability over data, the two algorithms require O(nlog( logη ϵ)) steps to obtain samples with at most ϵ error in TV, KL or χ2 distance. Meanwhile, we show that under minimal data assumptions, LassoDA requires O(d2(dlogd+nlogn)2log(η ϵ)) steps to achieve ϵ-accuracy in TV distance. The results are generally applicable to settings with large n and large d, including settings with highly imbalanced response data in Probit and Logit regression. We compare them with the best known guarantees of Langevin Monte Carlo and Metropolis Adjusted Langevin Algorithm. We evaluate our theoretical results using numerical examples, and discuss the mixing times of the three algorithms under feasible initialization.
AoS Missing Data & Measurement Error · 发表于 2026-08-01
Tianyi Ma, Kabir A. Verchand, Thomas B. Berrett, Tengyao Wang, Richard J. Samworth
摘要
We study the effects of missingness on the estimation of population parameters. Moving beyond restrictive missing completely at random (MCAR) assumptions, we first formulate a missing data analogue of Huber’s arbitrary ϵ-contamination model. For mean estimation with respect to squared Euclidean error loss, we show that the minimax quantiles decompose as a sum of the corresponding minimax quantiles under a heterogeneous, MCAR assumption and a robust error term, depending on ϵ, that reflects the additional error incurred by departure from MCAR. We next introduce natural classes of realisable ϵ-contamination models, where an MCAR version of a base distribution P is contaminated by an arbitrary missing not at random (MNAR) version of P. These classes are rich enough to capture various notions of biased sampling and sensitivity conditions, yet we show that they enjoy improved minimax performance relative to our earlier arbitrary contamination classes for both parametric and nonparametric classes of base distributions. For instance, with a univariate Gaussian base distribution, consistent mean estimation over realisable ϵ-contamination classes is possible even when ϵ and the proportion of missingness converge (slowly) to 1. We extend our results to the setting of departures from missing at random (MAR) in normal linear regression with a realisable missing response, and also demonstrate that our methods can be made adaptive to the case of unknown ϵ.
AoS High-Dimensional Testing & Uncertainty · 发表于 2026-08-01,收录于 2026-08-20
Yixiang Luo, William Fithian, Lihua Lei
摘要
(verbatim from arXiv:2408.07231)*: "We introduce a generic estimator for the false discovery rate of any model selection procedure, in common statistical modeling settings including the Gaussian linear model, Gaussian graphical model, and model-X setting. We prove that our method has a conservative (non-negative) bias in finite samples under standard statistical assumptions, and provide a bootstrap method for assessing its standard error. For methods like the Lasso, forward-stepwise regression, and the graphical Lasso, our estimator serves as a valuable companion to cross-validation, illuminating the tradeoff between prediction error and variable selection accuracy as a function of the model complexity parameter.
AoS Other · 发表于 2026-08-01
Yaqing Chen, Paromita Dubey
摘要
This paper develops a novel unified framework for testing mutual independence among random objects residing in possibly different metric spaces. The framework generalizes existing methodologies and introduces new measures of mutual independence, and proposes associated tests that achieve minimax rate optimality and exhibit strong empirical power. The foundation of the proposed tests is the new concept of joint distance profiles, which uniquely characterize the joint law of random objects under a mild condition on either the joint law or the metric spaces. Our test statistics quantify the difference of the joint distance profiles of each data point with respect to the joint law and the product of marginal laws of the vector of random objects. To enhance power, we consider integrating this difference with respect to different measures and incorporate flexible data-adaptive weight profiles in the test statistics. We derive the limiting distribution of the test statistics under the null hypothesis of mutual independence and show that the proposed tests with certain weight profiles are asymptotically distribution-free if the marginal distance profiles are continuous. Furthermore, we establish the consistency of the tests under sequences of alternative hypotheses converging to the null. For practical implementations, we employ a permutation scheme to approximate the p-values and provide theoretical guarantees that the permutation-based tests maintain type I error control under the null and achieve consistency under alternatives. We demonstrate the power of the proposed tests across various types of data objects through simulations and real data applications, where the new tests exhibit better performance compared with popular existing approaches.
AoS Other · 发表于 2026-08-01
Hang Zhou, Hans-Georg Müller
摘要
We develop a unified framework for testing independence and quantifying association between random objects that are located in general metric spaces. Special cases include functional and high-dimensional data as well as networks, covariance matrices and data on Riemannian manifolds, among other metric space-valued data. A key concept is the profile association, a measure based on distance profiles that intrinsically characterize the distributions of random objects in metric spaces. We establish a connection between Hoeffding’s D statistic and the profile association and derive a permutation test with theoretical guarantees for consistency and power under alternatives to the null hypothesis of independence (equivalently, no association). We extend this framework to the conditional setting, where the independence between random objects given a Euclidean predictor is of interest. In simulations across various metric spaces, the proposed profile independence test is found to outperform existing approaches. The practical utility of this framework is demonstrated with applications to brain connectivity networks derived from magnetic resonance imaging and age-at-death distributions for males and females obtained from human mortality data.
AoS High-Dimensional · 发表于 2026-08-01,收录于 2026-08-20
Gen Li, Yuting Wei
摘要
Characterizing the distribution of high-dimensional statistical estimators is a challenging task, due to the breakdown of classical asymptotic theory in high dimension. This paper makes progress towards this by developing non-asymptotic distributional characterizations for approximate message passing (AMP)—a family of iterative algorithms that prove effective as both fast estimators and powerful theoretical machinery—for both sparse and robust regression. Prior AMP theory, which focused on high-dimensional asymptotics for the most part, failed to describe the behavior of AMP when the number of iterations exceeds o(logn/loglogn) (with n the sample size). Our AMP theory is further leveraged to derive a refined rate for the Lasso estimator, built upon a nonasymptotic convergence of AMP to Lasso. To the best of our knowledge, this is the first result of its kind, breaking the rate barrier resulting from the widely adopted Gaussian min-max theorem. Our results establish approximate accuracy of Gaussian approximation of the AMP iterates, which improves upon prior results and implies enhanced distributional characterizations for both optimally tuned Lasso and robust M-estimator.
Biometrics Other · 发表于 2026-07-31
Lu Chen, Xuerong Chen, Xinzhou Guo, Yi Li
摘要
In clinical trials, a treatment rarely benefits every patient, underscoring the need to identify subgroups that are more likely to respond. Traditional subgroup analysis approaches, including finite mixture and threshold models, often rely on stringent distributional assumptions and prespecified subgroup structures that may be unrealistic in practice. Moreover, the resulting subgroups can be difficult to interpret and may not generalize well to new patients. To address these challenges, we propose a new least-squares regression framework that accommodates flexible subgroup structure in heterogeneous data. Our model is distribution-free and allows subgroup membership to depend on covariates, while permitting both the number and the organization of coefficient groups to vary across covariates. Building on regularization, we develop a computationally efficient procedure to detect subgroup structure in linear regression coefficients and then use a support vector machine to recover the corresponding partitions, enabling subgroup membership prediction for future individuals. Relative to pairwise fused regularization, our approach substantially reduces computational complexity. We also establish theoretical guarantees for estimation of group-specific parameters and recovery of the underlying partitions. Simulation studies and a real-data application illustrate the practical effectiveness of the proposed method.
Biometrics Spatial & Time Series Survival Analysis · 发表于 2026-07-30
Jamshid Namdari, Robert T Krafty, Amita Manatunga
摘要
Many studies collect data that can be considered as a realization of a point process. Partial least squares (PLS) is a popular analytic approach that combines features from linear modeling as well as dimension reduction to provide parsimonious prediction and classification. However, existing PLS methodologies do not include the analysis of point process predictors. In this article, we introduce point process PLS ($P^3LS$) for analyzing latent time-varying intensity functions from collections of inhomogeneous point processes. We develop a novel estimation procedure for $P^3LS$ that utilizes the properties of log-Gaussian Cox processes and examine its empirical properties via simulation studies. In addition, we establish the predictive consistency for the $P^3LS$. Finally, we apply the proposed method to a renal radionuclide imaging study to predict kidney obstruction from patient renograms. The method yields a clinically meaningful and interpretable model, demonstrating its potential to support data-driven clinical decision-making in renal studies.
SMMR Survival Analysis · 发表于 2026-07-30
Saptangshu Nandi, Sandip Barui, Debanjan Mitra, Narayanaswamy Balakrishnan
摘要
In lifetime studies, a proportion of subjects may never experience the event of interest, giving rise to the notion of cure rate. While traditional cure models address this in univariate set-up, many applications involve paired lifetimes, such as twin survival studies, paired organs, or dependent components in a system. In such cases, both the possibility of cure in one or both marginals and the dependence between lifetimes need to be modeled simultaneously, motivating bivariate cure models. We propose here a flexible framework for analyzing bivariate cure data by combining parametric modeling of each marginal survival distribution with a shared frailty term for capturing dependence. The shared frailty is assumed to follow a generalized gamma distribution, providing substantial flexibility in representing diverse dependence structures induced by unobserved frailty. This formulation accommodates cured individuals in one or both marginals and captures the joint survival dynamics of susceptible pairs. Likelihood-based inference for the proposed model is then developed, with estimation carried out using the expectation–maximization algorithm. The performance of the model and the associated inferential method are then evaluated through extensive simulation studies, demonstrating the efficiency and accuracy in estimating model parameters. A real data application, concerning diabetic retinopathy, further illustrates how the model provides insights into cure dynamics and associations that will not be evident in univariate analyses. The proposed bivariate cure model offers a comprehensive statistical tool for jointly studying cure fractions and survival dependence in paired lifetime settings, broadening the applicability of cure rate methodology in biomedical research.
SiM Diagnostics & Prediction Longitudinal Data Survival Analysis · 发表于 2026-07-30,收录于 2026-08-07
Siddharth Roy, Paul S. Albert, Anindya Roy, Danping Liu
摘要
Longitudinal cancer biomarker studies aim to identify markers useful for long‐term associations with cancer risk and early detection of cancer diagnosis. Early detection biomarkers show acute associations, meaning longitudinal trajectory changes sharply just before diagnosis, while long‐term risk prediction biomarkers are often represented as changes in slopes or levels. We compare three common approaches to identify longitudinal biomarkers associated with survival outcomes: joint models, conditional models, and Cox models with time‐varying covariates. Each of the three methods uses a different modeling framework for the joint density of the biomarkers and survival time. Thus, they have distinct advantages and disadvantages for detecting acute and long‐term associations. We investigate the power and Type I error rates for the three approaches under different data‐generating settings with regular yearly visits and moderate measurement error to evaluate robustness to assumptions and the power gain from using test statistics that match the data's association structure. The Cox model controlled Type I errors rates and maintained high power to detect associations across the scenarios considered. The conditional and joint models had inflated Type I errors under longitudinal model misspecification and only outperformed the Cox model for power when the correct model is used. However, only the conditional model can effectively disentangle the acute and long‐term effects. The Cox model's convex likelihood also provides the fastest convergence. We apply all three approaches to a study of the association between CA‐125 and ovarian cancer in the National Cancer Institute Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial. The results follow similar patterns to those of our simulations.
SiM Clinical Trials Bayesian Methods · 发表于 2026-07-30,收录于 2026-08-07
Yining Li, Jiaying Guo, Samer Gawrieh, Wei Shen, Yong Zang, Wanzhu Tu
摘要
Clinical syndromes or diseases with complex etiology often require multiple treatments, each addressing one specific aspect of the disease. Such treatments are typically administered sequentially in an integrated fashion. For example, in alcohol‐associated hepatitis (AH), an acute liver disease caused by excessive alcohol drinking, successful treatment requires therapies that reduce liver inflammation during the acute phase, followed by interventions that address the underlying alcohol use disorder (AUD) to achieve optimal outcomes. Selecting an optimal treatment combination, however, presents great challenges in study design due to the involvement of multiple drug combinations and treatment phases. In this paper, we describe a new trial design based on a Bayesian model that connects the outcomes from the acute and subsequent treatment phases. We refer to this design as the Bayesian Optimal Adaptive Design for Sequentially Integrated Therapies (BIT) design. BIT incorporates multiple interim analyses with adaptive stopping rules for both futility and superiority, enhancing efficiency while controlling the family‐wise type I error rate and maximizing statistical power. We illustrate the use of BIT design in a clinical trial comparing the efficacy of integrated therapy for severe AH, with sample size determination and design parameter optimization. Simulation studies show that the design possesses excellent operating characteristics, including experiment‐wise type I error rate control. Although the design's development is motivated by studies of AH treatment, the design framework is broadly applicable to other complex diseases requiring sequentially delivered therapies. R codes are provided for BIT implementation.
SiM Bayesian Methods Longitudinal Data · 发表于 2026-07-28,收录于 2026-07-31
Bijit Roy, Emmanuel Lesaffre
摘要
In this article, we adapt a Bayesian model selection criterion for use with derived response variables, motivated by modeling BMI as a function of its components, weight and height. We argue that the best model for the original responses (weight and height) may not be the optimal model for a derived, non‐linear quantity like BMI. Existing model selection criteria such as , , and are limited because they measure the out‐of‐sample predictive accuracy of the original responses. To address this limitation, we propose to adapt the . The Derived () is specifically designed to measure the out‐of‐sample predictive accuracy of the derived quantity of interest. This also allows for direct comparison of models based on their performance in predicting the target variable (BMI), regardless of whether they model the original components (height and weight) or the derived variable directly. We illustrate the utility of by comparing a bivariate growth model for height and weight with a spline smoothing model for BMI, a comparison that cannot be done with traditional criteria like or .
JASA Other · 发表于 2026-07-27,收录于 2026-07-31
Moshe Pollak
摘要
In the framework of the Cusum procedure, the evolution of a false alarm has a well-understood stochastic behavior. So, if observations preceding an alarm were to exhibit a behavior that is significantly different, there would be reason to reject the hypothesis that the alarm is false.We develop a test of this difference. The method is applied to detecting a change in a Covid-19 context involving a possible increase of a mean and in a context involving a possible increase in the probability of a rare event. Supplementary materials for this article are available online. *(abstract via OpenAlex)*
JASA High-Dimensional Spatial & Time Series Latent Variables & Mixtures · 发表于 2026-07-27,收录于 2026-06-05
Xinghao Qiao, Zihan Wang, Qiwei Yao, Bo Zhang
摘要
The factor modeling for high-dimensional time series is powerful in discovering latent common components for dimension reduction and information extraction. Most available estimation methods can be divided into two categories: the covariance-based under asymptotically-identifiable assumption and the autocovariance-based with white idiosyncratic noise. This paper follows the autocovariance-based framework and develops a novel weight-calibrated method to improve the estimation performance. It adopts a linear projection to tackle high-dimensionality, and employs a reduced-rank autoregression formulation. The asymptotic theory of the proposed method is established, relaxing the assumption on white noise. Additionally, we make the first attempt in the literature by providing a systematic theoretical comparison among the covariance-based, the standard autocovariance-based, and our proposed weight-calibrated autocovariance-based methods in the presence of factors with different strengths. Extensive simulations are conducted to showcase the superior finite-sample performance of our proposed method, as well as to validate the newly established theory. The superiority of our proposal is further illustrated through the analysis of one financial and one macroeconomic data sets. *(abstract via OpenAlex)*
JASA Bayesian Methods Spatial & Time Series · 发表于 2026-07-27,收录于 2026-07-31
Rihui Ou, Lachlan Astfalck, Deborshee Sen, David Dunson
摘要
Bayesian computation often scales poorly with increasing data size, motivating developments such as divide-and-conquer approaches for scalable inference. These methods partition the data into subsets, perform parallel inference on each subset, and aggregate the results into a single posterior. Appealing theoretical properties and practical performance have been demonstrated for independent data; however, methods for dependent data remain challenging. Existing methods rely on ad hoc approximations with limited theoretical guarantees that lead to potentially poor accuracy in practice. Here, we focus on time-series data and propose a simple, scalable divide-and-conquer method for dependent time series, with theoretically rigorous accuracy guarantees. Simulation studies and real-data examples are used to empirically verify the effectiveness of our approach. *(abstract via OpenAlex)*