Toggle Main Menu Toggle Search

Open Access padlockePrints

A general robust t-process regression model

Lookup NU author(s): Dr Jian Shi


Full text for this publication is not currently held within this repository. Alternative links are provided below where available.


© 2020 Elsevier B.V.The Gaussian process regression (GPR) model is well-known to be susceptible to outliers. Robust process regression models based on t-process or other heavy-tailed processes have been developed to address the problem. However, due to the current definitions of heavy-tailed processes, the unknown process regression function and the random errors are always defined jointly. This definition, mainly owing to mix-up of the regression function modeling and the distribution of the random errors, is not justified in many practical problems and thus limits the application of those robust approaches. It also results in a limitation of the statistical properties and robust analysis. A general robust process regression model is proposed by separating the nonparametric regression model from the distribution assumption of the random error. An efficient estimation procedure is developed. It shows that the estimated random-effects are useful in detecting outlying curves. Statistical properties, such as unbiasedness and information consistency, are provided. Numerical studies show that the proposed method is robust against outliers and outlying curves, and has a better performance in prediction compared with the existing models.

Publication metadata

Author(s): Wang Z, Noh M, Lee Y, Shi JQ

Publication type: Article

Publication status: Published

Journal: Computational Statistics and Data Analysis

Year: 2021

Volume: 154

Print publication date: 01/02/2021

Online publication date: 24/09/2020

Acceptance date: 08/09/2020

ISSN (print): 0167-9473

ISSN (electronic): 1872-7352

Publisher: Elsevier B.V.


DOI: 10.1016/j.csda.2020.107093


Altmetrics provided by Altmetric