Year: 2025
Type: article
Abstract: Unbiased estimation, as one of the core concepts of parameter estimation in statistics, requires that the mathematical expectation of the estimator be strictly equal to the true value of the estimated parameter. This characteristic gives it the important advantage of "no systematic deviation" in theory. However, in the actual process of statistical inference, the existence, uniqueness and validity of unbiased estimation often face many theoretical limitations and practical challenges. This study systematically sorts out the theoretical basis of unbiased estimation and comprehensively explores the applicable conditions, essential limitations and possible improvement directions of unbiased estimation. The study adopts a method combining theoretical derivation and empirical analysis, focusing on the following key issues: Firstly, under what model settings may unbiased estimation generally exist; Secondly, when unbiased estimation exists, how to evaluate its statistical efficiency and what kind of unbiased estimation should be adopted; Finally, in complex scenarios where traditional unbiased estimation performs poorly, what are the feasible improvement strategies? Through this series of studies, this study aims to provide theoretical basis and practical guidance for statistical practitioners to select appropriate estimation methods in different application scenarios.
Source: Theoretical and Natural Science
Author: Junting Liu
Institution: Beijing Institute of Technology
Language: English
Subfield: Statistics and Probability
Field: Mathematics
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W4413962052