A General Class of Multicomponent System Models with Limited Failure

• Activity Number: 298
• Type: Topic Contributed
• Date/Time: Tuesday, August 11, 2015 : 8:30 AM to 10:20 AM
• Sponsor: Section on Physical and Engineering Sciences
• Author(s): Sanjib Basu* and Qi Jiang
• Companies: Northern Illinois University and Northern Illinois University
• Keywords: Bayesian ; Competing risks ; Cumulative Incidence ; Cure rate

Abstract

A multi-component series system is often modeled in a competing risks framework. A limited-failure model postulates a fraction of the systems to be failure-free. Popular limited failure formulations include a mixture model and a bounded cumulative intensity model. Modeling of competing risks data based on the cumulative incidence functions approach has the advantage of providing direct inference on the failure probabilities from each risk. We propose a unified competing risks cure rate model that directly models the cumulative incidence functions. The proposed model further accounts for the possibility of limited failure from one or more of the competing risks. We propose a Bayesian approach for the analysis of this model and provide numerical examples to illustrate the performance of the proposed model.