Non-regular statistical estimation (Q1899197)

In the theory of statistical inference, parametric estimation plays a major role. Under some ``regularity'' conditions, it is known that the maximum likelihood estimators are consistent, asymptotically normal and asymptotically efficient in the sense that they achieve the Cramér-Rao lower bound. However there are situations where these ``regularity conditions'' are not satisfied and yet the estimations are ``super- efficient'' in some sense. The subject matter of this monograph is the study of such ``nonregular'' and other related problems. The contents of the monograph are as follows: Chapter 1: General discussion on unbiased estimation; Chapter 2: Lower bound for the variance of unbiased estimators; Chapter 3: Amounts of information and the minimum variance unbiased estimators; Chapter 4: Loss of information associated with the order statistics and related estimators in the case of double exponential distributions; Chapter 5: Estimation of a common parameter for pooled samples from the uniform distributions and the double exponential distribution; Chapter 6: Higher-order asymptotics in estimation for two-sided Weibull type distributions; Chapter 7: ``3/2- th'' and second order asymptotics for the generalised Bayes estimators for a family of truncated distributions; Supplement: The bound for the asymptotic distribution of estimators when the maximum order of consistency depends on the parameter. The monograph has a large number of examples sprinkled throughout and is a welcome addition to the list of books on asymptotic theory of estimation.