Fully Functional, Historical Functional
Most of my work has been on varying function-on-function regression models including fully functional models with random effects and historical functional models or HFLMs. I’ve largely worked with wavelets for my basis expansion and in the Bayesian context. My publications page contains links to additional work in this area.
Function-on-Function Regression
As part of my dissertation work, I developed a wavelet-based Bayesian function-on-function method for repeated samples. This method considered some innovative basis-space techniques for modeling the functional predictor.
Historical Functional Linear Models
I’ve also worked on a wavelet-packet based historical functional linear model. One of the challenges of historical models is enforcing the constraint which we implement via a novel use of wavelet-packets.
Additional Function-on-Function work
I’ve also done some other work in function-on-function methods including an extension to non-square surfaces for analyzing epigenetic data as well as a review article on historical functional linear models.