Rethinking biased estimation
Document Type
Article
Date of Original Version
1-1-2008
Abstract
In order to improve the accuracy of unbiased estimators used in many signal processing, there are methods that can be used. This approach is based on introducing a bias as a means of reducing the mean-squared error (MSE). In seeking unbiased estimators that perform well, it is typically accomplished by determining the minimum variance unbiased (MVU) estimator, using the theory of sufficient statistics of the Cramér-Rao lower bound. Estimators can be derived to outperform existing approaches for short data records and low signal-to-noise ratios (SNRs). The applications include the design of estimation algorithms for sonar, radar, and communications, as well as many other disciplines that rely heavily on precise measurement of parameters.
Publication Title, e.g., Journal
IEEE Signal Processing Magazine
Volume
25
Issue
3
Citation/Publisher Attribution
Kay, Steven, and Yonina C. Eldar. "Rethinking biased estimation." IEEE Signal Processing Magazine 25, 3 (2008): 133-136. doi: 10.1109/MSP.2008.918027.