Abstract:
|
Functional data analysis (FDA) deal with data in the form of curves, surfaces or more complex objects. Our work is motivated by a biomedical study using Raman spectroscopy to monitor hemodialysis treatment near real-time. The Raman spectrum of a biological sample provides a fingerprint of its chemical composition and constituent concentrations. In this presentation, we introduce a locally sparse contrast test to compare groups of spectra data. For a significant contrast spectrum, it can identify the region of significant difference and thus pinpointing the chemicals in the waste dialysate that are potential signatures of dialysis adequacy.
|