Breakdown points and variation exponents of robust $M$-estimators in linear models

Contentas University and Georg-August-University The breakdown point behavior of M-estimators in linear models with fixed designs, arising from planned experiments or qualitative factors, is characterized. Particularly, this behavior at fixed designs is quite different from that at designs which can be corrupted by outliers, the situation prevailing in the literature. For fixed designs, the breakdown points of robust M-estimators (those with bounded derivative of the score function), depend on the design and the variation exponent (index) of the score function. This general result implies that the highest breakdown point within all regression equivariant estimators can be attained also by certain M-estimators : those with slowly varying score function, like the Cauchy or slash maximum likelihood estimator. The M-estimators with variation exponent greater than 0, like the L 1 or Huber estimator, exhibit a considerably worse breakdown point behavior.

[1]  J. B. Copas,et al.  On the unimodality of the likelihood for the Cauchy distribution , 1975 .

[2]  V. Yohai,et al.  Bias- and efficiency-robustness of general M-estimators for regression with random carriers , 1979 .

[3]  Jana Jurečková.  Tail-Behavior of Location Estimators , 1981 .

[4]  Gorm Gabrielsen,et al.  On the unimodality of the likelihood for the Cauchy distribution: Some comments , 1982 .

[5]  Michael Stuart,et al.  Understanding Robust and Exploratory Data Analysis , 1984 .

[6]  P. J. Huber.  Finite Sample Breakdown of $M$- and $P$-Estimators , 1984 .

[7]  Werner A. Stahel,et al.  Robust Statistics: The Approach Based on Influence Functions , 1987 .

[8]  S. Resnick.  Extreme Values, Regular Variation, and Point Processes , 1987 .

[9]  Peter J. Rousseeuw,et al.  Robust regression and outlier detection , 1987 .

[10]  D. Ruppert.  Robust Statistics: The Approach Based on Influence Functions , 1987 .

[11]  M. Meerschaert.  Regular Variation in R k , 1988 .

[12]  Jeremy MG Taylor,et al.  Robust Statistical Modeling Using the t Distribution , 1989 .

[13]  Xuming He.  TAIL BEHAVIOR OF REGRESSION ESTIMATORS AND THEIR BREAKDOWN POINTS , 1990 .

[14]  D. Ruppert.  Computing S Estimators for Regression and Multivariate Location/Dispersion , 1992 .

[15]  S. P. Ellis,et al.  Leverage and Breakdown in L 1 Regression , 1992 .

[16]  Ivan Mizera,et al.  On consistent M-estimators: Tuning constants, unimodality and breakdown , 1994, Kybernetika.

[17]  C. Müller.  Optimal breakdown point maximizing designs , 1995 .

[18]  Christine H. Müller,et al.  Breakdown points for designed experiments , 1995 .

[19]  C. Müller.  High Breakdown Point Designs , 1996 .

[20]  C. W. Coakley,et al.  Robust estimation in structured linear regression , 1996 .

[21]  B. Price,et al.  Robust Planning and Analysis of Experiments , 1997 .