Robust regression through robust covariances. (English) Zbl 0639.62023

Summary: This paper discusses the estimation of regression parameters after summarizing the data by a covariance matrix of the concatenated vector of explanatory variables and response variable. A robust estimate of the covariance matrix leads to a robust regression estimator. An M-estimator at the covariance estimation step is studied in the paper, and the resulting regression estimator is compared to a few previously proposed robust regression estimators.


62F35 Robustness and adaptive procedures (parametric inference)
62H12 Estimation in multivariate analysis
62J05 Linear regression; mixed models
Full Text: DOI


[1] Daniel C., Fitting Equations to Data (1980)
[2] Donoho D., The notion of breakdown point (1983) · Zbl 0523.62032
[3] DOI: 10.1214/aoms/1177693054 · Zbl 0229.62041
[4] DOI: 10.2307/2285666 · Zbl 0305.62031
[5] DOI: 10.1109/T-PAS.1975.31858
[6] Huber, P.J. The behavior of maximum likelihood estimates under nonstandard conditions. Proc. Fifth Berkeley Symposium on Mathematical Statistics and Prob. Vol. 1, Berkeley: Univ. of California Press. Cal · Zbl 0212.21504
[7] DOI: 10.1214/aos/1176342503 · Zbl 0289.62033
[8] DOI: 10.1002/0471725250 · Zbl 0536.62025
[9] DOI: 10.1214/aos/1176346396 · Zbl 0557.62034
[10] DOI: 10.2307/2287719 · Zbl 0501.62062
[11] Mallows C.L., On some topics in robustness (1975)
[12] DOI: 10.1214/aos/1176343347 · Zbl 0322.62054
[13] Maronna R.A., Bias- and efficiency-robustness of general M estimators for regression with random carriers (1979) · Zbl 0416.62050
[14] Mosteller F., Data Anal.ysis and Regression: A second course in Statistics (1977)
[15] Stahel W., Robust estimation: infinitesimal optimality and covariance matrix estimators (1981)
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. It attempts to reflect the references listed in the original paper as accurately as possible without claiming the completeness or perfect precision of the matching.