It is well known that the Barankin bound is locally the tightest lower bound on the variance of any unbiased estimator. If the minimum-variance unbiased-parameter estimation problem is viewed as a minimum-norm problem in an appropriate Hilbert space, the Barankin bound has a very simple geometric interpretation. This interpretation also gives some insight into the influence of bias on the estimation error.
Published in:
Information Theory, IEEE Transactions on
(Volume:19
,
Issue:
4
)
Date of Publication: Jul 1973