A0277
Title: Low-rank regularization of Frechet regression models for distribution function response
Authors: Kyunghee Han - University of Illinois at Chicago (United States) [presenting]
Hsin-Hsiung Huang - University of Central Florida (United States)
Abstract: Frechet regression has emerged as a promising approach for modeling non-Euclidean response variables in relation to Euclidean covariates. A novel estimation method is introduced that integrates rank regularization into the underlying model for metric-space-valued responses. Specifically focusing on distribution function responses, it is demonstrated how this framework employs low-rank regularization techniques for functional linear regression to enhance the efficiency and accuracy of the global Frechet regression estimator. By leveraging the low-rank structure, the method enables more robust modeling and estimation, particularly in high-dimensional settings. A detailed theoretical analysis of the large-sample properties of the proposed estimator is presented. Numerical experiments further validate these theoretical results.
