A0445
Title: Probabilistic analysis of multi-way random functions
Authors: Donatello Telesca - UCLA (United States) [presenting]
Abstract: Experimental and observational settings are considered, where data can be conceptualized as independent realizations of a univariate or multivariate functional process defined over multiple arguments. We discuss how the probabilistic formulation of latent random features can be useful in characterizing several classes of covariance operators, which, in turn, describe the second order properties of the underlying stochastic processes. In this setting, we show how regularized estimation can be achieved naturally within the Bayesian inferential framework. Several applications to longitudinal functional data and multi-way imaging data are used for illustration.
