A comparison of estimation methods for reliability function of inverse generalized Weibull distribution under new loss function

Abstract

In this paper, we focussed on the scale parameter and reliability estimations of the inverse generalized Weibull distribution. Both classical and Bayesian approaches are considered with various loss functions as general entropy, squared log error and weight squared error. For the Bayesian method, both informative and non-informative priors are applied for the reliability and scale parameter estimation. Furthermore, we introduce a new loss function that exhibits some attractive performances. The reliability function and scale parameter of the inverse generalized Weibull distribution are estimated based on the new loss function. By the Monte Carlo simulation procedure, we demonstrate the efficiency of the new proposed loss function among some competitors in estimating the reliability function . Finally, the analysis of two real data set has also been represented for illustration purposes. Some goodness of fit measures affirmed the adequacy of the inverse generalized Weibull distribution in modelling real data sets.

Keywords:

• Integrated mean squared error
• uniformly minimum variance unbiased estimator
• Bayes estimators
• non-informative prior
• reliability function

2010 Mathematics Subject Classifications:

• 62F15
• 62P10
• 62P30
• 62N02

Disclosure statement

No potential conflict of interest was reported by the author(s).