A new weighted exponential distribution as an alternative to the Weibull distribution and its fit to reliability data. (22nd February 2022)

• Record Type: Journal Article
• Authors: Bakouch, Hassan S.; Chesneau, Christophe; Enany, Mai G.
• Is Part Of: International journal of data science. Volume 6:Number 3(2021)
• Page Start: 223
• Page End: 240
• Publication Date: 2022-02-22
• Subjects: weighted distributions -- statistical and reliability properties -- reliability data -- goodness-of-fit statistics; Big data -- Periodicals; Data mining -- Periodicals; Information storage and retrieval systems -- Periodicals; Decision making -- Data processing -- Periodicals; 005.7
• Languages: English
• ISSNs: 2053-0811

Abstract

In this paper, we introduce and study an unexplored two-parameter weighted exponential distribution, having deep connections with the so-called exponentiated exponential and exponential-logarithmic distributions. Among its advantages, the corresponding probability density and hazard rate functions display quite attractive shapes for various modelling aims. Our theoretical contributions on the new distribution include some results on first-order stochastic dominance, the expression of the quantile function, expansion series for the moments, with discussions on the incomplete moment and moment generating function. Entropy and extropy are also investigated. Inferential work is performed on the related model; the estimation of parameters is justified by the method of maximum likelihood. The potential and elasticity of the model is illustrated by fitting it to two reliability datasets.