Artikel
Copulas in Survival Analysis
Suche in Medline nach
Autoren
Veröffentlicht: | 13. September 2012 |
---|
Gliederung
Text
Copulas are useful to model correlated data. Especially in the case of continuous data it is always possible (based on the theorem of Sklar) to construct a multivariate distribution with given marginal distributions. Most applications of copulas are devoted to the field of finance. However, also in the modelling of multivariate event time distributions with applications in medicine, epidemiology, and demography copulas are of interest. We focus on the case of bivariate survival data and illustrate the advantages and limitations of copula based survival models illustrated with real data examples.