Abstract
The paper deals with the problem of estimating the mean life θ1 in exponential model under LINEX loss function. It is assumed that in addition to the current sample, a sample collected some time in past when the mean life might have been θ2 is also available. A some times pool estimator is proposed which estimates θ1 by ar always pool estimator if Ho :θ1 = θ2 is true and by a never pool estimator otherwise. Both the never pool and always pool estimators are derived so that they possess optimum risk properties under LINEX loss function. It is shown that usual estimators of the same class are inadmissible under LINEX loss function. The level of significance of the proposed some times pool estimator is decided using the AIC (Akaike's Information.
1Earlier version of this paper was presented in 81st Congress of ISCA held at Jaipur, India, Jan. 1994.
1Earlier version of this paper was presented in 81st Congress of ISCA held at Jaipur, India, Jan. 1994.
Notes
1Earlier version of this paper was presented in 81st Congress of ISCA held at Jaipur, India, Jan. 1994.