scientific article; zbMATH DE number 3890515

From MaRDI portal
Publication:3223689





[For the entire collection see Zbl 0517.00012.] From the authors' introduction: In its simplest form the Cramér-Rao inequality asserts: under regularity conditions, the variance of an unbiased estimator of a parametric function is at least the square of the derivative of that function divided by the Fisher information in the sample for all values of the unknown parameter. Unfortunately, the regularity conditions may be severe. In early work, the regularity conditions placed restrictions on the estimator itself, which is undesirable if the goal is to obtain an inequality which must be satisfied by all estimators. Here it is shown that the Cramér-Rao inequality holds for almost every value of the parameter (Lebesgue) under the sole condition that information be definable; and even the definition of information... is more general than previous ones.... an a.e. version of the Cramér-Rao inequality is established for sequential estimators, under the condition that the information in a single observation be positive.... By using approximate limits and derivatives, in place of real ones, the present proofs avoid requiring that the densities be absolutely continuous in the parameter or that the variance of the estimator be locally integrable.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3223689)