Project Fact File

Title: Estimation of Mean Under Uncertain Prior Information
Category: statistics
Area: Statistical Inference
No. of units: 1 or 2 or 4
Supervisor: Shahjahan Khan (Staff Profile)
Description:

The most commonly used estimator for the mean vector is the sample mean. This estimator is unbiased and depends only on the sample data. At times, the researchers may have additional information regarding the value of the mean vector from the past experience or other sources of knowledge. Such a prior information is often uncertain in nature. Nevertheless, it can be used in the estimation of the mean vector to improve the quality of the estimator. However, the classical maximum likelihood or least square estimators (e.g., sample mean) is incapable of incorporating such additional information in to the estimation of the unknown parameter. Unlike the sample mean, the preliminary test and shrinkage estimators can include both the sample data and the uncertain prior information for the estimation of the unknown parameters, like the population mean vector. I have investigated different improved estimators for various models and studied their properties under various conditions.

Student: NULL