Modelling accelerated life test data by using a Bayesian approach

Authors: Sinha D.¹; Patra K.²; Dey D.K.³

Source: Journal of the Royal Statistical Society: Series C (Applied Statistics), Volume 52, Number 2, May 2003 , pp. 249-259(11)

Publisher: Wiley-Blackwell

Buy & download fulltext article:

Price: $48.00 plus tax (Refund Policy)

Abstract:

Summary.

Because of the high reliability of many modern products, accelerated life tests are becoming widely used to obtain timely information about their time-to-failure distributions. We propose a general class of accelerated life testing models which are motivated by the actual failure process of units from a limited failure population with a positive probability of not failing during the technological lifetime. We demonstrate a Bayesian approach to this problem, using a new class of models with non-monotone hazard rates, the hazard model with potential scope for use far beyond accelerated life testing. Our methods are illustrated with the modelling and analysis of a data set on lifetimes of printed circuit boards under humidity accelerated life testing.

Keywords: Conditional predictive ordinate; Latent risk; Limited failure population; Markov chain Monte Carlo methods

Document Type: Research article

DOI: http://dx.doi.org/10.1111/1467-9876.00402

Affiliations: 1: Medical University of South Carolina, Charleston, USA 2: Harvard School of Public Health, Boston, USA 3: University of Connecticut, Storrs, USA

Publication date: 2003-05-01