Research

My methodological research involves primarily functional data analysis and cluster analysis, and in recent years has seen a focus on applying innovative methods in the analysis of environmental data. A common theme of much of my early research was using data alteration such as smoothing and shrinkage to improve the quality of some data analysis (cluster analysis in particular).  In later years, interesting applications have driven my research endeavors, such as a project with John Grego to estimate the size of a loggerhead turtle population, given very limited and nonstandard observed capture-recapture-type information. I have collaborated with Ian Dryden and Ph.D. students Nicole Lewis and Wen Cheng in identifying peptides and aligning spectra resulting from mass spectrometry data. Branching off from Cheng's work, Ph.D student Zizhen Wu and I developed a method for the simultaneous registration and clustering of functional data.

Continuing my environmental data emphasis in recent years, working with Ph.D. student Haigang Liu and regular collaborator S.Z. (Vidya) Samadi (Water Resources Engineering, Clemson), I have developed models for spatio-temporal precipitation and flooding data.  Along with John Grego and Ph.D. student Ryan Pittman, I have used functional data regression to reconstruct river stage data from a major flood event.  Another recent functional data application is the analysis of cannulation techniques to guide medical skill training, in collaboration with R. (Joseph) Singapogu and Zhanhe Liu (Bioengineering, Clemson).  Other recent projects have focused on time series models, e.g., deep learning ensemble modeling to forecast stock prices, with Ph.D. student Shan Zhong; and multinomial time series models for music genre popularity, with honors thesis student Aimée Petitbon.