He, Xuming; Simpson, Douglas G.; Wang, Guangyu Breakdown points of \(t\)-type regression estimators. (English) Zbl 1120.62320 Biometrika 87, No. 3, 675-687 (2000). Summary: To bound the influence of a leverage point, generalised \(M\)-estimators have been suggested. However, the usual generalised \(M\)-estimator of regression has a breakdown point that is less than the inverse of its dimension. This paper shows that dimension-independent positive breakdown points can be attained by a class of well-defined generalised \(M\)-estimators with redescending scores. The solution can be determined through optimisation of \(t\)-type likelihood applied to properly weighted residuals. The highest breakdown point of \(\frac12\); is attained by Cauchy score. These bounded-influence and high-breakdown estimators can be viewed as a fully iterated version of the one-step generalised \(M\)-estimates of Simpson, Ruppert and Carroll (1992) with the two advantages of easier interpretability and avoidance of undesirable roots to estimating equations. Given the design-dependent weights, they can be computed via EM algorithms. Empirical investigations show that they are highly competitive with other robust estimators of regression. Cited in 15 Documents MSC: 62J05 Linear regression; mixed models 62F35 Robustness and adaptive procedures (parametric inference) 62F10 Point estimation Keywords:breakdown point; generalised \(M\)-estimator; linear regression; likelihood; robustness; \(t\)-distribution PDFBibTeX XMLCite \textit{X. He} et al., Biometrika 87, No. 3, 675--687 (2000; Zbl 1120.62320) Full Text: DOI