A0376
Title: Bayesian inference for sphere-on-sphere regression with optimal transport map
Authors: Tin Lok James Ng - Trinity College Dublin (Ireland) [presenting]
Andrew Zammit Mangion - University of Wollongong (Australia)
Kwok-Kun Kwong - University of Wollongong (Australia)
Jiakun Liu - University of Wollongong (Australia)
Abstract: The field of spherical regression, where both covariate and response variables take values on the sphere, has seen extensive methodological development over time. Despite the creation of various parametric and non-parametric techniques to tackle spherical regression, it remains a challenging problem due to the complexities involved in parameterizing regression models between spherical domains. Additionally, there is a notable gap in methods for quantifying uncertainties associated with the estimated regression maps. To address these challenges, optimal transport theory is utilized, and a Bayesian approach is employed. This framework eliminates the necessity for directly parameterizing the regression map and enables uncertainty quantification.
