Abstract

YANEZ, Sergio; ESCOBAR, Luis A.  and  GONZALEZ, Nelfi. Characteristics of two Competing Risks Models with Weibull Distributed Risks. Rev. acad. colomb. cienc. exact. fis. nat. [online]. 2014, vol.38, n.148, pp.298-311. ISSN 0370-3908.

Lifetime and survival data are usually the lives of subjects, units, or systems that has been exposed to multiple risks or modes of failure. In the analysis of the data, however, is common to ignore the modes of failure because they are unknown, they are not recorded, or because the complexity of including them in the modeling. It is of interest knowing when the conclusions might be robust to ignoring the failure modes in the analysis. In particular, it would be useful to characterize situations where it could be safely said that the modes of failure effect on the analysis would be negligible or that failing to include such information could completely invalidate the conclusions drawn from the study. As a first step in identifying when the failure modes have little or large influence on the competing risks model, this article studies two different competing risks models: (a) a model with independent risks; (b) a model derived from a multivariate Weibull with dependence.

Keywords : competing risks; probability plots; log-location-scale family; Weibull distribution; multivariate Weibull with dependence; lognormal distribution.

        · abstract in Spanish     · text in English     · English ( pdf )