Prediction for Coherent System Lifetime Based on Type-II Censored Data from Half Logistic Distribution
Year: 2024
Type: article
Abstract: Results and DiscussionComparing the Bayes estimators obtained using the Metropolis-Hastings algorithm and importance sampling methods, we observe that the Metropolis-Hastings algorithm method is better than the importance sampling method.By comparing the classical predictors and the Bayesian predictor, it can be seen that in all coherent systems, the Bayesian predictor performs better than the classical predictors.Also, among the methods for constructing prediction intervals considered here, the pivotal quantity method can provide prediction intervals for all the censored system failures.It gives the best performance in terms of coverage probabilities and average widths. ConclusionUsing Monte Carlo simulation for several coherent systems, we recommend using the Bayesian predictor for point prediction unless reliable prior information on the unknown parameter is unavailable.For interval prediction, we suggest using the pivotal quantity method to construct prediction intervals for the censored system failures.
Source: Journal of Statistical Sciences
DOI: 10.61186/jss.18.2.9
Authors: Adeleh Fallah Taloki, Roshanak Zaman
Institution: Payame Noor University
Language: English
Subfield: Statistics and Probability
Field: Mathematics
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W4405736192