Title: Wavelet-based Probit Functional Models
Abstract: Recent advances in functional regression include the development of a wide range of function-on-scalar regression models where outcome is a function of time regressed onto a set of scalar covariates. Most of these methodologies require the outcome to be normally distributed. However, very little work has been done on the case where the outcome is categorical. In this talk, we will discuss a flexible Bayesian procedure for function-on-scalar regression for categorical functional outcomes of varying levels. Simulations and data examples will be presented to illustrate the methodology.