Classifications of Proteomic Mass Spectra
and Other Curve Data

Xiaohui Wang, Ph.D
Department of Mathematics
University of Texas-Pan American

Abstract


            Disease studies based on proteomic mass spectra seem promising. Classification of proteomic mass spectra is a challenging task because of the features of the spectra. Motivated from this problem, we propose classification models for binary and multicategory data where the predictor is a random function. Our methodology is Bayesian, using wavelet basis functions which have nice approximation properties over a large class of functional spaces and can accommodate a variety of functional forms observed in real life applications.