A5091306803- Statistical Methods and Inference
- Advanced Causal Inference Techniques
- Advanced Statistical Methods and Models
- Bayesian Methods and Mixture Models
- Statistical Methods and Bayesian Inference
- Mental Health Research Topics
- Statistical Methods in Clinical Trials
- Functional Brain Connectivity Studies
- Soil Geostatistics and Mapping
- Treatment of Major Depression
- Advanced Clustering Algorithms Research
- Health Systems, Economic Evaluations, Quality of Life
- Gaussian Processes and Bayesian Inference
- Cardiac Arrest and Resuscitation
- Statistical Distribution Estimation and Applications
- Autism Spectrum Disorder Research
- Statistical and numerical algorithms
- Pain Management and Placebo Effect
- COVID-19 Clinical Research Studies
- Advanced Control Systems Optimization
- Traumatic Brain Injury and Neurovascular Disturbances
- COVID-19 Impact on Reproduction
- Gene expression and cancer classification
- Advanced Statistical Process Monitoring
- Sensory Analysis and Statistical Methods
New York University
2016-2025
Columbia University Irving Medical Center
2023
Wright State University
2009-2021
Columbia University
2009-2020
NYU Langone Health
2019
Nathan Kline Institute for Psychiatric Research
2010
New York Psychoanalytic Society and Institute
2009
Inter-Mountain Laboratories (United States)
1995
National Institute of Standards and Technology
1993-1994
United States Department of Commerce
1994
<h3>Importance</h3> Identifying which patients with COVID-19 are likely to benefit from convalescent plasma (CCP) treatment may have a large public health impact. <h3>Objective</h3> To develop an index for predicting the expected relative CCP compared without hospitalized using patients' baseline characteristics. <h3>Design, Setting, and Participants</h3> This prognostic study used data COMPILE study, ie, meta-analysis of pooled individual patient 8 randomized clinical trials (RCTs)...
The term "self-consistency" was introduced in 1989 by Hastie and Stuetzle to describe the property that each point on a smooth curve or surface is mean of all points project orthogonally onto it. We generalize this concept self-consistent random vectors: vector Y for X if $\mathscr{E}[X|Y] = Y$ almost surely. This allows us construct unified theoretical basis principal components, curves surfaces, points, variables, modes variation other statistical methods. provide some general results give...
Objectives We explored medications for opioid use disorder treatment (MOUD) utilization in six New York City public hospitals that implemented the “Consultation Addiction Care and Treatment Hospitals (CATCH)” program. Methods CATCH rolled out between October 2018 February 2020. Data from electronic health record were analyzed first year post-implementation. Eligible cases included adults with an opioid-related diagnosis admitted to inpatient departments served by CATCH, a stay of ≥1 night....
ABSTRACT This paper presents a Bayesian regression model relating scalar outcomes to brain functional connectivity represented as symmetric positive definite (SPD) matrices. Unlike many proposals that simply vectorize the matrix-valued predictors, thereby ignoring their geometric structure, method presented here respects Riemannian geometry of SPD matrices by using tangent space modeling. Dimension reduction is performed in space, resulting low-dimensional representations responses. The...
Functional data can be clustered by plugging estimated regression coefficients from individual curves into the k-means algorithm. Clustering results differ depending on how are fit to data. Estimating using different sets of basis functions corresponds linear transformations clustering is not invariant The optimal transformation for will stretch distribution so that primary direction variability aligns with actual differences in clusters. It shown raw often give similar obtained an...
The $k$ principal points of a $p$-variate random vector $\mathbf{X}$ are those $\xi_1, \ldots, \xi_k \in \mathbb{R}^p$ which approximate the distribution by minimizing expected squared distance from nearest $\xi_j$. Any set $\mathbf{y}_1, \mathbf{y}_k$ partitions $\mathbb{R}^p$ into "domains attraction" $D_1, D_k$ according to minimal distance; following Hastie and Stuetzle we call self-consistent if $E\lbrack\mathbf{X}\mid\mathbf{X} D_j\rbrack = \mathbf{y}_j$ for $j 1, k$. Principal special...
Abstract There is a well-known simple formula for computing prediction sum of squares (PRESS) residuals in regression problem without having to refit the curve each observation. This note shows that same basic result holds fitting function when coefficients are subject linear constraints. Key Words: Leave-one-outLinear constraintsRegression
Abstract There is a well-known simple formula for computing prediction sum of squares (PRESS) residuals in regression problem without having to refit the curve each observation. This note shows that same basic result holds fitting function when coefficients are subject linear constraints.
Summary The amount and complexity of patient-level data being collected in randomized-controlled trials offer both opportunities challenges for developing personalized rules assigning treatment a given disease or ailment. For example, examining treatments major depressive disorder are not only collecting typical baseline such as age, gender, scores on various tests, but also that measure the structure function brain images from magnetic resonance imaging (MRI), functional MRI (fMRI),...
Summary Many techniques of functional data analysis require choosing a measure distance between functions, with the most common choice being distance. In this article we show that using weighted distance, judiciously chosen weight function, can improve performance various statistical methods for data, including k ‐medoids clustering, nonparametric classification, and permutation testing. Assuming quadratically penalized (e.g., spline) basis representation consider three nontrivial functions:...
A recent meta‐regression of antidepressant efficacy on baseline depression severity has caused considerable controversy in the popular media. central source is a lack clarity about relation parameters to corresponding models for subject‐level data. This paper focuses linear regression with continuous outcome and predictor, case that often considered less problematic. We frame general mixture setting encompasses both finite infinite models. In many applications meta‐analysis, goal evaluate...
Abstract The “mother knows best” hypothesis states that adults should choose hosts for oviposition on which their offspring will best perform, maximizing own fitness. It has been hypothesized this preference—performance relationship wood-boring insects is especially important because larvae are not able to switch hosts, although no study examined choices these insects. We preferences of the emerald ash borer, Agrilus planipennis Fairmaire (Coleoptera: Buprestidae), in two common gardens, one...
Antidepressant medications are commonly used to treat depression, but only about 30% of patients reach remission with any single first-step antidepressant. If the treatment fails, response and rates at subsequent steps even more limited. The literature on biomarkers for is largely based secondary analyses studies designed answer primary questions efficacy, rather than a planned systematic evaluation decision. lack evidence-based knowledge guide decisions depression has lead recognition that...
Abstract I examine the self-consistency of a principal component axis; that is, when distribution is centered about axis. A axis random vector X self-consistent if each point on corresponds to mean given projects orthogonally onto point. large class symmetric multivariate distributions are examined in terms subspaces. Elliptical characterized by preservation axes after arbitrary linear transformations. “lack-of-fit” test proposed tests for The applied two real datasets.
We propose a penalized spline approach to performing large numbers of parallel non-parametric analyses either two types: restricted likelihood ratio tests parametric regression model versus general smooth alternative, and nonparametric regression. Compared with naïvely each analysis in turn, our techniques reduce computation time dramatically. Viewing the collection scatterplot smooths produced by methods as functional data, we develop clustering summarize visualize these results. Our is...