| Summary: | The minimax estimation is an upgraded non-clasical approach method in
the estimation area of statistical inference which is closely related with Bayes
estimation. The most important element in this estimation are the prior
distribution and the loss function. In the thesis, parameter estimation of the
Rayleigh distribution using minimax estimation method under quadratic and
modified linier exponential loss function is studied. The prior distribution as used
in this estimation is invers moment prior distribution. In the data simulation, the
estimated values of the parameter and mean squared error (MSE) of the
estimators are computed. And the result indicate that for small sample size
(n < 25) dan c in interval (-2,3), semi minimax estimator under modified linier
exponential loss function appear to be better (efficient) than the semi minimax
estimator under quadratic loss function. But for sample size (n < 25) and other c
show the opposite result. And for large sample size (n > 25), both of estimators
have approximately the same mean squared error (MSE).
|
|---|