The robust estimators of reliability function for some compound distribution

Abstract

This paper investigates the use of the POT technique for natural phenomena weather. Their data contain extreme values (high or rare values) by applying two methods. The first method is based on the estimate using the order statistical, and the second method is based on likelihood ratio. The comparative model is based on MSE of the estimated parameters, the probability density function for each distribution of the mixed distributions, and their reliability function. The POT sample has shown significant improvement in data and its superiority, especially in simple exponential distribution.