A5001016275- Statistical Methods and Inference
- Advanced Causal Inference Techniques
- Statistical Methods and Bayesian Inference
- Statistical Methods in Clinical Trials
- Health Systems, Economic Evaluations, Quality of Life
- Survey Sampling and Estimation Techniques
- Bayesian Methods and Mixture Models
- Economic and Environmental Valuation
- Survey Methodology and Nonresponse
- Spatial and Panel Data Analysis
- Bayesian Modeling and Causal Inference
- HIV/AIDS Research and Interventions
- Advanced Statistical Methods and Models
- Air Quality and Health Impacts
- Soil Geostatistics and Mapping
- Hemophilia Treatment and Research
- HIV, Drug Use, Sexual Risk
- COVID-19 Impact on Reproduction
- Distributed Sensor Networks and Detection Algorithms
- E-commerce and Technology Innovations
- COVID-19 Pandemic Impacts
- COVID-19 epidemiological studies
- Qualitative Comparative Analysis Research
- Adolescent Sexual and Reproductive Health
- Pregnancy and Medication Impact
North Carolina State University
2017-2025
Jiangsu University
2025
Fujian Medical University
2024
Qingdao University of Science and Technology
2023-2024
University of North Carolina at Chapel Hill
2020-2024
Union Hospital
2024
Film Independent
2024
Linyi University
2024
Guangxi University
2024
Chengdu University of Traditional Chinese Medicine
2020-2023
In this article, we develop new methods for estimating average treatment effects in observational studies, settings with more than two levels, assuming unconfoundedness given pretreatment variables. We emphasize propensity score subclassification and matching which have been among the most popular binary literature. Whereas literature has suggested that these particular propensity-based do not naturally extend to multi-level case, show, using concept of weak notion generalized score,...
ChatGPT, a large-scale language model based on the advanced GPT-3.5 architecture, has shown remarkable potential in various Natural Language Processing (NLP) tasks. However, there is currently dearth of comprehensive study exploring its area Grammatical Error Correction (GEC). To showcase capabilities GEC, we design zero-shot chain-of-thought (CoT) and few-shot CoT settings using in-context learning for ChatGPT. Our evaluation involves assessing ChatGPT's performance five official test sets...
Marine protected areas (MPAs) are widely used for ocean conservation, yet the relative impacts of various types MPAs poorly understood. We estimated on fish biomass from no-take and multiple-use (fished) MPAs, employing a rigorous matched counterfactual design with global dataset >14,000 surveys in around 216 MPAs. Both generated positive conservation outcomes to no protection (58.2% 12.6% increases, respectively), smaller differences between two MPA when controlling additional confounding...
Causal inference with observational studies often relies on the assumptions of unconfoundedness and overlap covariate distributions in different treatment groups. The assumption is violated when some units have propensity scores close to |$0$| or |$1$|, so both practical theoretical researchers suggest dropping extreme estimated scores. However, existing trimming methods do not incorporate uncertainty this design stage restrict only trimmed sample, due nonsmoothness trimming. We propose a...
"Flexible Imputation of Missing Data, 2nd ed.." Journal the American Statistical Association, 114(527), p. 1421
Summary We consider integrating a non-probability sample with probability which provides high dimensional representative covariate information of the target population. propose two-step approach for variable selection and finite population inference. In first step, we use penalized estimating equations folded concave penalties to select important variables show consistency general samples. second focus on doubly robust estimator mean re-estimate nuisance model parameters by minimizing...
Abstract We propose a test-based elastic integrative analysis of the randomised trial and real-world data to estimate treatment effect heterogeneity with vector known modifiers. When are not subject bias, our approach combines for efficient estimation. Utilising design, we construct test decide whether or use data. characterise asymptotic distribution estimator under local alternatives. provide data-adaptive procedure select threshold that promises smallest mean square error an confidence...
Due to the heterogeneity of randomized controlled trial (RCT) and external target populations, estimated treatment effect from RCT is not directly applicable population. For example, patient characteristics ACTG 175 HIV are significantly different that three populations interest: US early-stage patients, Thailand southern Ethiopia patients. This paper considers several methods transport beyond Most focus on continuous binary outcomes; contrary, we derive discuss for survival outcomes: an...
The era of big data has witnessed an increasing availability multiple sources for statistical analyses. We consider estimation causal effects combining main with unmeasured confounders and smaller validation supplementary information on these confounders. Under the unconfoundedness assumption completely observed confounders, allow constructing consistent estimators effects, but can only give error-prone in general. However, by leveraging a principled way, we improve efficiencies yet preserve...
Abstract Complementary features of randomized controlled trials (RCTs) and observational studies (OSs) can be used jointly to estimate the average treatment effect a target population. We propose calibration weighting estimator that enforces covariate balance between RCT OS, therefore improving trial-based estimator's generalizability. Exploiting semiparametric efficiency theory, we doubly robust augmented achieves bound derived under identification assumptions. A nonparametric sieve method...
Oral anticoagulation (OAC) in atrial fibrillation (AF) reduces the risk of stroke/systemic embolism (SE). The impact OAC discontinuation is less well documented.Investigate outcomes patients prospectively enrolled Global Anticoagulant Registry Field-Atrial Fibrillation study who discontinued OAC.Oral was defined as cessation treatment for ≥7 consecutive days. Adjusted outcome risks were assessed 23 882 with 511 days median follow-up after discontinuation.Patients (n = 3114, 13.0%) had a...
Flexible sensors with high sensitivity and multifunctional integrated stimuli in the environment have been used various applications. Here, we report construction of a type high-performance flexible photonic pressure sensor inspired by principle chameleon color change. The is constructed based on lens-shaped three-dimensional crystal combination an alginate-based hydrogel. Low-pressure can lead to slight change band gaps that, consequently, cause shifting maximal frequency variation...
Adjusting for covariates in randomized controlled trials can enhance the credibility and efficiency of average treatment effect estimation. However, managing numerous their non-linear transformations is challenging, particularly when outcomes have missing data. In this tutorial, we propose a principled covariate adjustment framework, "COADVISE," that enables (i) variable selection most relevant to outcome, (ii) nonlinear adjustments, (iii) robust imputation data both covariates. This...
This study assesses the long-term effects of redlining policies (1935-1974) on present-day fine particulate matter (PM$_{2.5}$) and nitrogen dioxide (NO$_2$) air pollution levels. Redlining enacted in 1930s, so there is very limited documentation pre-treatment covariates. Consequently, traditional methods fails to sufficiently account for unmeasured confounders, potentially biasing causal interpretations. By integrating historical data with 2010 PM$_{2.5}$ NO$_2$ levels, our aims discern...
Abstract Rejective sampling improves design and estimation efficiency of single-phase when auxiliary information in a finite population is available. When such unavailable, we propose to use two-phase rejective (TPRS), which involves measuring variables for the sample units first phase, followed by implementation outcome second phase. We explore asymptotic properties double expansion regression estimators under TPRS. show that TPRS enhances double-expansion estimator, rendering it comparable...
The predictive value of triglyceride-glucose index (TyG) for cardiovascular disease (CVD) in the US elderly diabetic patients is ambiguous. This study aimed to investigate association between TyG and risk CVD an older population with diabetes. examined data from 2007-2016 National Health Nutrition Examination Survey (NHANES). Univariate multivariate regression analysis models were obtained explore baseline CVD. Non-linear investigated using restricted cubic spline (RCS) regression. Subgroup...