Name in Chinese: 吴振科 . Pronounced: “Jen-Kuh Wu”.
[Publications] [CV] [Contact] [Bio]
[Google Scholar] [GitHub] [Twitter].
The best way to contact me is email. Direction to my office is here.
I am an Assistant Professor in the Department of Biostatistics at University of Michigan, with joint appointment as Research Assistant Professor in Michigan Institute for Data Science (MIDAS). I am also Faculty Associate in Quantitative Methodology Program, Survey Research Center of Institute for Social Research (ISR), University of Michigan.
Research Theme:
My research is motivated by biomedical and public health problems and is centered on the design and application of statistical methods that inform health decisions made by individuals, or precision medicine. Towards this goal, I focus on two lines of methodological research: a) structured Bayesian latent variable models for clustering and disease subtyping, and b) study design, causal and reinforcement learning methods for evaluating sequential interventions that tailor to individuals’ changing circumstances such as in interventional mobile health studies. I am committed to developing robust, scalable, and interpretable statistical methods to harness real-world, high-dimensional, dynamic data for individualized health. The methods and software developed so far have supported studies in diverse scientific fields including infectious disease epidemiology, autoimmune diseases, mental health, behavioral health, and cancer.
Keywords:
Statistical: Hierarchical Bayesian models; Latent variable models; Nonparametric Bayes; Bayesian scalable computation; Causal inference; Reinforcement learning.
Substantive: Precision medicine; Mobile health; Infectious diseases; Mental health; Electronic health records/claims data; Healthcare policy; Clinical trials; Just-in-time adaptive interventions for behaviorial and psychiatric research.
Advising: We are recruiting motivated and hard-working people interested in Bayesian methods and computation, graphical models, causal inference, sequential decision making, reinforcement learning and large-scale health data analytics. If you want to get involved, please say hi.
Working Group:
I collaborate closely with
Precision Health Use Case: PROviding Mental Health Precision Treatment (PROMPT)
Congratulations to Tsung-Hung Yao (co-advisor: Veera Baladandayuthapani) for receiving a competitive travel award to 2022 International Society of Bayesian Analysis (ISBA), Montreal. This is based on his work “Probabilistic Learning of Treatment Trees in Cancer”.
Congratulations to Hera Shi (co-advisor: Walter Dempsey) for receiving a competitive Junior Researcher Travel Grant to attend 2022 American Causal Inference Conference (ACIC) at Berkeley, CA. This is based on her work “Assessing Time-Varying Causal Effect Moderation in the Presence of Cluster-Level Treatment Effect Heterogeneity”.
baker has a first public release (v1.0.0) at CRAN! Discussion related to version 1.0.0 can be submitted to here; [vignette] [source code].
The definitive reference for baker R package can be found here.
This vignette describes and illustrates the functionality of the baker R package. The package provides a suite of nested partially-latent class models (NPLCM) for multivariate binary responses that are observed under a case-control design. The baker package allows researchers to flexibly estimate population- and individual-level class distributions that may also depend on additional explanatory covariates.
Functions in baker implement recent methodological developments in our group (here, here, and here). Estimation is accomplished by calling a cross-platform automatic Bayesian inference software JAGS through a wrapper R function that parses model specifications and data inputs. The baker package provides many useful features, including data ingestion, exploratory data analyses, model diagnostics, extensive plotting and visualization options, catalyzing vital communications between practitioners and domain scientists. Package features and workflows are illustrated using simulated and real data sets.
The focus of this document is on guiding a new user to utilize some useful functions in baker for simulation studies and data analyses, aided by other powerful R packages. We refer readers of this document to the accompanying main software paper for more details about the software design considerations and review of model formulations. Since baker’s first appearance on Github, the authors have not been able to track other recent substantive publications that have used this package; we hope the main software paper and this vignette serve as the definitive reference for future scientific studies that find the baker package useful.
An exciting new collaboration is supported by 2021 Propelling Original Data Science (PODS) grant from Michigan Institute for Data Science (MIDAS). The project title is “Structured Latent Variable Methods for High-Dimensional Electronic Health Records and Administrative Claims Data”. The faculty investigators are Zhenke Wu (PI; Biostatistics, Public Health), Jordan Schaefer (Co-I; Hematology, Internal Medicine), and Andrew Ryan (Co-I; Health Management and Policy, Public Health). The primary data source will be based on UnitedHealthcare OptumInsight claims data via Data and Methods Hub (DMH) at UM Institute for Healthcare Policy and Innovation (IHPI).
Huge congrats to Irena and Hera for passing their PhD qualifying exams!!
Huge congrats to Mengbing for winning “Best Doctoral Qualifying Exam Award”.
Another huge congrats to Irena for winning a poster award today in the 2019 MIDAS Annual Symposium in the category of “Most Likely to Make an Impact in the Field” for her poster “Regression Analysis for Probabilistic Cause-of-disease Assignment using Case-control Diagnostic Tests: A Hierarchical Bayesian Approach”!
Way to go!!