Robustness Studies for Bayesian Developments in Reliability.
Abstract
The Weibull probability density function is considered as a failure model under the influence of a stochastic scale parameter. Bayesian estimates of the scale parameter, reliability function and hazard rate are given for the conjugate prior distribution. A method for evaluating the robustness of the conjugate prior distribution which characterizes the behavior of the parameter is presented. A computer simulation to investigate the robustness of the conjugate prior distribution with respect to six other prior probability distributions, namely, beta, Poisson, inverted gamma, truncated normal, log-normal and extreme value, were employed. These results indicate that there is a significant variation in the average mean squared error even when the priors were chosen so that their first two moments approximately agreed with that of the conjugate.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 1976
- Accession Number
- ADA020673
Entities
People
- A. N. V. Rao
- C. P. Tsokos
Organizations
- University of South Florida