Statistical inference of constant-stress partially accelerated life tests under type II generalized hybrid censored data from Burr III distribution

Statistical inference of constant-stress partially accelerated life tests under type II generalized hybrid censored data from Burr III distribution

Accelerated life tests (ALTs) involve subjecting units to more extreme conditions than normal to reduce the duration of testing. These tests, whether fully accelerated or partially accelerated, are crucial in life testing research as they help save both time and money. When results from ALT cannot b...

Full description

Saved in:
Bibliographic Details
Main Authors: Hassan Amal S., Alsadat Najwan, Kayid Mohamed, Elgarhy Mohammed, Balogun Oluwafemi Samson, Almetwally Ehab M.
Format: Article
Language:English
Published: De Gruyter 2025-07-01
Series:Open Physics
Subjects:
constant-stress partially accelerated life tests
burr type iii distribution, maximum-likelihood estimation
asymptotic confidence intervals
bayesian estimation
markov chain monte carlo
Online Access:https://doi.org/10.1515/phys-2025-0161
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Accelerated life tests (ALTs) involve subjecting units to more extreme conditions than normal to reduce the duration of testing. These tests, whether fully accelerated or partially accelerated, are crucial in life testing research as they help save both time and money. When results from ALT cannot be extrapolated to normal circumstances, partial ALTs are performed. This study introduces a constant-stress partial ALT, which relies on a generalized Type-II hybrid censoring scheme and assumes that the lifetimes of units under the specified conditions follow the Burr III distribution. The estimates of parameters and accelerated factor of the Burr III distribution are obtained under normal use settings by applying the maximum likelihood and Bayesian techniques. Bayesian estimators are generated using symmetric and asymmetric loss functions through the Monte Carlo Markov Chain approach. Additionally, credible intervals and asymptotic confidence intervals are constructed. Simulation research is carried out using different censoring techniques and sample sizes in order to compare the suggested methodologies. Next, two data sets are analyzed to show the value of the proposed approaches. The study’s conclusion contains a summary of its main conclusions.
ISSN:2391-5471

Similar Items