Sign In
Skip Navigation Links.
Expand AboutAbout
Programs & Degrees
Our People
Expand Quick LinksQuick Links
Expand MediaMedia
Recent

Seminar-An Improved Pretest Estimator

Zoom 25 JAN 20224 PM

Title: An Improved Pretest Estimator

 

Abstract:

In this seminar , a new estimator called an improved pretest estimator for estimating the population mean parameter of a normal distribution when a prior non-sample information about the parameter is available will be discussed. The regular pretest estimator is based on testing a hypothesis about the population mean; if the null hypothesis is not rejected, it chooses the estimator value under the null hypothesis, and the estimator value based on the data sample otherwise. The new pretest estimator gives a small weight to the estimator value under the null hypothesis even if it is rejected, and a small weight to the data estimator value even if the null hypothesis is not rejected. The improved estimator is compared with the regular pretest, shrinkage and shrinkage pretest estimators. Analytical and numerical results showed that the new version dominates all other estimators in a certain part of the parameter space and behaves like others in another part. A real data example to compare all estimators considered in this study and employed bootstrapping approach to compute the bootstrapping values of all estimators, and their mean squared errors.

 

Date: Tuesday, ​January 25 , 2022

Time: 4:00 pm

Join Zoom Meeting

Meeting ID: 977 6577 8824

Passcode: 676478




Back to list