SOMETIMES POOL TESTIMATION OF SCALE PARAMETER FOR NEGATIVE EXPONENTIAL MODEL UNDER GENERAL ENTROPY LOSS FUNCTION

  • Rakesh Srivastava Department of Statistics, The M.S. University of Baroda, Vadodara, India
  • Vilpa Tanna Department of P.S.M., P.D.U. Govt. Medical College, Rajkot, Gujarat, India
Keywords: Negative exponential distribution, Scale Parameter, Type – II Censored Samples, Preliminary Test, Level of Significance, General Entropy Loss Function, Life Ratio, Euler’s Psi Function, Relative Risk

Abstract

The present paper proposes a sometimes pool testimator of scale parameter (mean life) of negative exponential distribution under general entropy loss function, when it is assume that both the guarantees are known If a real life situation is modeled by this distribution having mean life time of certain items as θ1 and now it is suspected that the mean life may change due to some technological advances and assumes the value θ2 then we may have conditional information on θ1 as θ1≥θ2(however θ1 may be less than or equal to θ2).This uncertainty can be resolved by using preliminary testing and then a sometimes pool testimator is proposed for θ1The risk properties of this estimator have been studied under General entropy loss function and it is claimed that the estimator dominates the never pool estimator (in terms of having smaller risk) in certain range of life ratio. Use of a general entropy loss function facilitates to control the risk of proposed estimators for various directions and degrees of asymmetry.

 

Downloads

Download data is not yet available.
Published
2015-08-25
Section
Articles