Expectile regression with errors-in-variables

• Published in: IEEE access Vol. 11; p. 1
• Main Authors: He, Xiaoxia, Zhou, Xiaodan, Li, Chunli
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 01-01-2023; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Data models; Error analysis; Error reduction; Errors-in-variables; Expectile regression; IRWLS Algorithm; Load modeling; Mathematical models; Measurement errors; Measurement uncertainty; Normality; Orthogonal distance regression; Parameter estimation; Regression; Training data
• ISSN: 2169-3536; 2169-3536
• DOI: 10.1109/ACCESS.2023.3287571

This paper studies the expectile regression with error-in-variables to reduce the data error and describe the overall data distribution. Specifically, the asymptotic normality of the proposed estimator is thoroughly investigated, and an IRWLS algorithm based on orthogonal distance expectile regression (ODER) is proposed to estimate the parameters. Extensive simulation studies and real data applications evaluate our method's capabilities in reducing the measurement error bias, demonstrating our model's parameter estimation effectiveness, and its capability in reducing the simulation error compared with linear and quantile regression schemes.