Estimation of The Unknown Parameter of A Uniform Distribution

The goal this paper is study of three Maximum likelihood estimators for the parameter θ in a population which is uniformly distributed over (θ-b/2, θ+b/2) where b is known and positive. For a sample of n observation, the maximum minus b/2, the maximum as corrected for bias, and the midrange, are compared as regard unbiasedness variance, symmetry, and several possible evaluations of relative error. Procedures for hypothesis testing and interval estimation of relative error. Procedures for hypothesis testing and interval estimation of θ are described and compared.

Keywords:

Estimator , maximum , midrange , relative accuracy , relative error , unbiasedness