Trimmed likelihood estimation of location and scale of the normal distribution. (English) Zbl 0798.62043

The authors consider trimmed likelihood estimators of location and scale of a normal distribution such that the estimators are robust. The approach employs empirical processes and differentiability of estimating functionals and is different from P. J. Huber [Am. Math. Stat. 35, 73-101 (1964; Zbl 0136.398); ibid. 43, 1041-1067 (1972; Zbl 0254.62023)] and F. R. Hampel’s [J. Am. Stat. Assoc. 69, 383-393 (1974; Zbl 0305.62031)] approaches. The estimator for location does not depend on the scale estimate and is robust against asymmetric contamination. Simulation results support their theory.


62F35 Robustness and adaptive procedures (parametric inference)
62F10 Point estimation
62H12 Estimation in multivariate analysis
Full Text: DOI