A5010565380- Statistical Methods and Inference
- Advanced Causal Inference Techniques
- Child Nutrition and Water Access
- Statistical Methods and Bayesian Inference
- Child Nutrition and Feeding Issues
- Statistical Methods in Clinical Trials
- Poverty, Education, and Child Welfare
- Global Maternal and Child Health
- Advanced Statistical Methods and Models
- Data Analysis with R
- Birth, Development, and Health
- Bioinformatics and Genomic Networks
- Medical Imaging Techniques and Applications
- Privacy-Preserving Technologies in Data
- Genetics, Bioinformatics, and Biomedical Research
- Graph Theory and Algorithms
- Gene expression and cancer classification
- Advanced Graph Neural Networks
University of California, Berkeley
2018-2024
Sustainable Development Goal 2.2-to end malnutrition by 2030-includes the elimination of child wasting, defined as a weight-for-length z-score that is more than two standard deviations below median World Health Organization standards for growth
Globally, 149 million children under 5 years of age are estimated to be stunted (length more than 2 standard deviations below international growth standards)
Abstract Growth faltering in children (low length for age or low weight length) during the first 1,000 days of life (from conception to 2 years age) influences short-term and long-term health survival 1,2 . Interventions such as nutritional supplementation pregnancy postnatal period could help prevent growth faltering, but programmatic action has been insufficient eliminate high burden stunting wasting low- middle-income countries. Identification windows population subgroups on which focus...
Abstract The Highly-Adaptive least absolute shrinkage and selection operator (LASSO) Targeted Minimum Loss Estimator (HAL-TMLE) is an efficient plug-in estimator of a pathwise differentiable parameter in statistical model that at minimal (and possibly only) assumes the sectional variation norm true nuisance functions (i.e., relevant part data distribution) are finite. It relies on initial (HAL-MLE) by minimizing empirical risk over space under constraint candidate bounded constant, where...
Summary Sustainable Development Goal 2.2, to end malnutrition by 2030, includes elimination of child wasting, defined as weight-for-length more than 2 standard deviations below international standards. Prevailing methods measure wasting rely on cross-sectional surveys that cannot onset, recovery, and persistence — key features inform preventive interventions disease burden estimates. We analyzed 21 longitudinal cohorts show is a highly dynamic process onset with incidence peaking between...
Summary Growth faltering (low length-for-age or weight-for-length) in the first 1000 days — from conception to two years of age influences short and long-term health survival. Interventions such as nutritional supplementation during pregnancy postnatal period could help prevent growth faltering, but programmatic action has been insufficient eliminate high burden stunting wasting low- middle-income countries. Future preventive efforts will benefit understanding age-windows population...
We consider estimation of a functional parameter realistically modeled data distribution based on observing independent and identically distributed observations. The highly adaptive lasso estimator the is defined as minimizer empirical risk over class cadlag functions with finite sectional variation norm, where parametrized in terms such functions. In this article we establish that HAL yields an asymptotically efficient any smooth feature under global undersmoothing condition. It formally shown
Researchers in observational survival analysis are interested not only estimating curve nonparametrically but also having statistical inference for the parameter. We consider right-censored failure time data where we observe n independent and identically distributed observations of a vector random variable consisting baseline covariates, binary treatment at baseline, subject to right censoring, censoring indicator. assume covariates allowed affect so that an estimator ignores covariate...
Summary Globally 149 million children under five are estimated to be stunted (length more than 2 standard deviations below international growth standards). Stunting, a form of linear faltering, increases risk illness, impaired cognitive development, and mortality. Global stunting estimates rely on cross-sectional surveys, which cannot provide direct information about the timing onset or persistence faltering— key consideration for defining critical windows deliver preventive interventions....
Doubly robust estimators are a popular means of estimating causal effects. Such combine an estimate the conditional mean outcome given treatment and confounders (the so-called regression) with probability propensity score) to generate effect interest. In addition enjoying double-robustness property, these have additional benefits. First, flexible regression tools, such as those developed in field machine learning, can be utilized relevant regressions, while effects retain desirable...
Targeted Learning is a subfield of statistics that unifies advances in causal inference, machine learning and statistical theory to help answer scientifically impactful questions with confidence. driven by complex problems data science has been implemented diversity real-world scenarios: observational studies missing treatments outcomes, personalized interventions, longitudinal settings time-varying treatment regimes, survival analysis, adaptive randomized trials, mediation networks...
Current Targeted Maximum Likelihood Estimation (TMLE) methods used to analyze time-to-event data estimate the survival probability for each time point separately, which result in estimates that are not necessarily monotone. In this paper, we present an extension of TMLE observational data, one-step Estimator treatment-rule specific curve. We construct a one-dimensional universal least favorable submodel targets entire curve, and thereby requires minimal extra fitting with achieve its goal...
Current statistical inference problems in areas like astronomy, genomics, and marketing routinely involve the simultaneous testing of thousands -- even millions null hypotheses. For high-dimensional multivariate distributions, these hypotheses may concern a wide range parameters, with complex unknown dependence structures among variables. In analyzing such hypothesis procedures, gains efficiency power can be achieved by performing variable reduction on set prior to testing. We present this...
The adaptest R package contains an implementation of a methodology based on using data-adaptive statistics for estimating effect sizes, complete with appropriate inference, in high-dimensional settings while avoiding the inferential burdens multiple testing corrections.To address issue situations where dimensionality is high but sample size comparatively small (e.g., analysis RNA-seq data), we expose method statistical inference target parameters (Hubbard, Kherad-Pajouh, & van der Laan,...