Workshop FDA: Registration
View Submission - CRONOSFDA2018
A0162
Title: Robust functional regression based on principal components Authors:  Ioannis Kalogridis - KU Leuven (Belgium) [presenting]
Abstract: Functional data analysis is a fast evolving branch of modern statistics, yet despite the popularity of the functional linear model in recent years, almost all estimation methods rely on generalized least squares procedures and as such are sensitive to atypical observations. To remedy this, we propose a two-step estimation procedure that combines robust functional principal components and robust linear regression. We further propose a transformation that reduces the curvature of the estimates and can be advantageous in many settings. For these methods we prove Fisher-consistency for elliptical distributions and consistency under mild regularity conditions. Simulation experiments show that the proposed estimators have reasonable efficiency, protect against outlying observations, produce smooth estimates and compare favourably with the few existing robust approaches.