Rank estimation for the functional linear model

作者：Denhere Melody^*; Bindele Huybrechts F

来源：Journal of Applied Statistics, 2016, 43(10): 1928-1944.

DOI：10.1080/02664763.2015.1125863

• Functional linear regression
• robust methods
• functional data
• outliers
• rank estimation
• Primary: 62J05
• 62G05
• Secondary: 62J99
• 62G35

This article discusses the estimation of the parameter function for a functional linear regression model under heavy-tailed errors' distributions and in the presence of outliers. Standard approaches of reducing the high dimensionality, which is inherent in functional data, are considered. After reducing the functional model to a standard multiple linear regression model, a weighted rank-based procedure is carried out to estimate the regression parameters. A Monte Carlo simulation and a real-world example are used to show the performance of the proposed estimator and a comparison made with the least-squares and least absolute deviation estimators.

• 出版日期2016-8