Entropy Analysis of Parameter Estimation

The problem of parameter estimation is considered by using the entropy of the error as the criterion function. The described entropy analysis is applicable to a larger class of problems than is usually studied through the use of mean square error analysis, and yields the classical results when applied to Gaussian-linear systems. An important aspect of the derivation of the minimum error entropy solution is that optimum bounds may be evaluated without first determining the optimum estimators. The potential importance of an entropy approach is demostrated with a very simple parameter estimation problem.