Abstract
We consider a novel family of cumulative hazard functions (CHFs) controlled by a single shape parameter, which corresponds to proportionality on a certain scale, in such a way that the family is closed under inversion of the CHF and under frailty mixing using an appropriate mixing distribution. The latter leads to natural shared frailty models in the bivariate case. We also suggest how best to incorporate a second, complementary, shape parameter in order to obtain especially useful parametric models for survival and reliability analysis.
Original language | English |
---|---|
Article number | 109059 |
Journal | Statistics and Probability Letters |
Volume | 172 |
Early online date | 18 Feb 2021 |
DOIs | |
Publication status | Published - 31 May 2021 |
Bibliographical note
Funding Information:Alice Davis was supported by a UK EPSRC DTA doctoral studentship at the University of Bath.
Publisher Copyright:
© 2021 Elsevier B.V.
Keywords
- Archimedean survival copula
- Parametric survival analysis
- Proportionality parameter
- Reliability analysis
- Shared frailty
ASJC Scopus subject areas
- Statistics and Probability
- Statistics, Probability and Uncertainty