Research

People are very interested in understanding how subject-level characteristics, including clinic variables and genetic factors, would influence imaging phenotypes. We consider a functional data approach to investigate how the clinic variables influence the brain imaging data. The proposed method can be applied to subjects with irregular shapes.

Multivariate Spline Estimation and Inference for Image-on-scalar Regression

Spatial or spatiotemporal data are generated at varying scales and levels of complexity. We study flexible and computationally efficient non-/semi- parametric models for spatiotemporal analysis methods.

Estimation and Inference for Generalized Geoadditive Models

Spatiotemporal Autoregressive Partially Linear Varying Coefficient Models

Distributed Heterogeneity Learning for Generalized Partially Linear Models with Spatially Varying Coefficients

We aim to provide a user-friendly tool to visualize, track and predict real-time infected/death cases of COVID-19 in the U.S., based on our collected data and proposed methods, and thus further illustrate the spatiotemporal dynamics of the disease spread and guide evidence-based decision making. 

Spatiotemporal Epidemic Modeling (STEM)

Realtime 7-day forecast and longterm forecast dashboard

We propose a sparse multi-group functional linear regression model to simultaneously estimate multiple coefficient functions and identify groups, such that coefficient functions are identical within groups and distinct across groups. By borrowing information from relevant subgroups of subjects, our method enhances estimation efficiency while preserving heterogeneity in model parameters and coefficient functions. We use an adaptive fused lasso penalty to shrink coefficient estimates to a common value within each group. 

Fusion Learning of Functional Linear Regression with Application to Genotype-by-Environment Interaction Studies