An estimator of the location parameter obtained from observations with admixture
Olena Sugakova · Theory of Probability and Mathematical Statistics · 2010
We consider a model of observations from a two-component mixture such that the distribution of the admixture is known, while the distribution of the other component (treated as the primary one) is unknown. We assume that the distribution of the primary component is symmetric about the location parameter. We propose a method for constructing an unbiased estimating equation for the location parameter of the primary component. The asymptotic normality of the corresponding estimators is proved. The exact lower bound for the asymptotic variance is found and estimating functions for which this bound is attained are described. It is shown that the exact lower bound for the estimators under consideration is close to the corresponding bound of effectiveness of parametric estimators for the case where the distributions of both components are Gaussian.