Online Program Home
My Program

Abstract Details

Activity Number: 422 - Statistical Learning for Functional Data
Type: Contributed
Date/Time: Tuesday, July 31, 2018 : 2:00 PM to 3:50 PM
Sponsor: Section on Statistical Learning and Data Science
Abstract #328669 Presentation
Title: Multivariate Calibration with Robust Signal Regression
Author(s): Bin Li* and Brian D. Marx and David C Weindorf and Somsubhra Chakraborty
Companies: Louisiana State University and Louisiana State University and Texas Tech University and Indian Institute of Technology Kharagpur
Keywords: Huber loss; Multivariate calibration; P-splines; Robust Regression; Signal Regression

Motivated by a multivariate calibration problem for a soil characterization study, we proposed tractable and robust variants of penalized signal regression (PSR) using a class of nonconvex Huber-like criteria as the loss function. Standard methods may fail to produce a reliable estimator when there are heavy-tailed errors. We present a computationally efficient algorithm to solve this nonconvex problem. Simulation and empirical examples are extremely promising and show the proposed algorithm substantially improves the PSR performance under heavy-tailed errors.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2018 program