1Universidad Nacional de Colombia, Facultad de Ciencias, Escuela de Estadística, Medellín, Colombia. Profesor asociado. Email: syanez@unal.edu.co
2Universidad de Córdoba, Departamento de Matemáticas y Estadística, Montería, Colombia. Profesor asistente. Email: habrango@hotmail.com
3Universidad Nacional de Colombia, Facultad de Ciencias, Escuela de Estadística, Medellín, Colombia. Profesor asociado. Email: mcjarami@unal.edu.co
]]>
En riesgos competitivos, el problema de identificabilidad asociado a la dependencia entre los modos de falla, se puede resolver utilizando el estimador cópula-gráfico que asume la forma de la cópula conocida. En este trabajo se compara, mediante simulación, este estimador para un modelo de riesgos competitivos dependientes con el estimador tradicional que asume independencia. Se muestra que ignorar la dependencia puede causar subestimación significativa de la función de confiabilidad. Se ilustra la temática con un ejemplo y se visualiza a futuro la posible utilidad del estimador cópula-gráfico en confiabilidad industrial para predecir en experimentos acelerados.
Palabras clave: confiabilidad, cópula, identificabilidad, riesgos competitivos.
The identifiability problem in competing risks associated to the dependence measure between the failure modes, can be solved by means of the copula-graphic estimator which assumes the copula function known. This paper compares, through a simulation study, the copula-graphic estimator for a dependent competing risks model with the traditional one under the assumption of independence. It is shown that a significant subestimation of the reliability function can be made when the dependence is not taken into account. An example ilustrates this point and as a future work it is seen the possibility of using copula-graphic in industrial reliability to predict in accelerated life tests.
Key words: Competing risks, Copula, Identifiability, Reliability.
Referencias
1. Barlow, R. & Proschan, F. (1975), Statistical Theory of Reliability and Life Testing, Holt, Rinehart and Winston, Inc, New York. [ Links ]
2. Basu, A. & Klein, P. (1982), `Some recent results in competing risks theory´, Proceedings on Survival Analysis. IMS monograph series 2, 216-229. [ Links ]
3. Bedford, T. (2005), Competing risk modelling in reliability, `Modern Statistical and Mathematical Methods in Reliability´, World Scientific Books, United Kingdom. [ Links ]
4. Bedford, T. & Lindqvist, B. (2004), `The identifiability problem for repairable systems subject to competing risks´, Advances in Applied Probability 36(3), 774-790. [ Links ]
5. Crowder, M. (2001), Classical Competing Risks, first edn, Chapman & Hall, London. [ Links ]
6. Frees, E. & Valdez, E. (1998), `Understanding relationships using copulas´, North American Actuarial Journal 2(1), 1-25. [ Links ]
7. Jaramillo, M. C., Lopera, C. M., Manotas, E. C. & Y\'a\~nez, S. (2008), `Generaci\'on de Tiempos de Falla Dependientes Weibull Bivariados Usando C\'opulas´, Revista Colombiana de Estad\'\istica 31(2), 169-181. [ Links ]
8. Johnson, M. (1987), Multivariate Statistical Simulation, John Wiley & Sons, New York. [ Links ]
9. Lu, J. & Bhattacharyya, G. (1990), `Some new constructions of bivariate weibull models´, Annals of the Institute of Statistical Mathematics 42(3), 543-559. [ Links ]
10. Manotas, E. C., Y\'a\~nez, S., Lopera, C. M. & Jaramillo, M. C. (2008), `Estudio del efecto de la dependencia en la estimaci\'on de la confiabilidad de un sistema con dos modos de falla concurrentes´, DYNA 75(154), 5-21. [ Links ]
11. Meeker, W., Escobar, L. A. & Hong, Y. (2009), `Using accelerated life tests results to predict product field reliability´, Technometrics 51(2), 146-161. [ Links ]
12. Meeker, W. & Escobar, L. (1998), Statistical Methods for Reliability Data, John Wiley and Sons, Inc, New York. [ Links ]
13. Nelsen, R. B. (2006), An Introduction to Copulas, Second edn, Springer, New York. [ Links ]
14. Nelson, W. (1990), Accelerated Testing: Statistical Models Test Plans and Data Analyses, John Wiley & Sons, New York. [ Links ]
15. Pintilie, M. (2006), Competing Risks: A Practical Perspective, John Wiley & Sons, United Kingdom. [ Links ]
16. R Development Core Team, (2010), R: A Language and Environment for Statistical Computing, R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0. *http://www.R-project.org/, \text{accedido en 01/08/2010} [ Links ]
17. Tsiatis, A. (1975), `A nonidentifiability aspect of the problem of competing risks´, Proceedings of the National Academy of Sciences 72, 20-22. [ Links ]
18. Y\'a\~nez, S., Manotas, E. C. & Lopera, C. M. (2006), `Estudio de la dependencia en la estimaci\'o de la confiabilidad de un sistema con dos modos de falla´, Memorias del XV Simposio de Estad\'\istica. [ Links ]
19. Zheng, M. & Klein, J. P. (1995), `Estimates of marginal survival for dependent competing risks based on an assumed copula´, Biometrika 82(1), 12-38. [ Links ]
Este artículo se puede citar en LaTeX utilizando la siguiente referencia bibliográfica de BibTeX:
@ARTICLE{RCEv34n2a02,
AUTHOR = {Yáñez, Sergio and Brango, Hugo and Jaramillo, Mario C. and Lopera, Carlos M.},
TITLE = {{Comparación entre riesgos competitivos vía el estimador cópula-gráfico}},
JOURNAL = {Revista Colombiana de Estadística},
YEAR = {2011},
volume = {34},
number = {2},
pages = {231-248} ]]>
}