Accelerated Failure Time models (AFT) are useful alternatives to the Cox model in survival analysis. Recently, AFT models for multivariate data have been considered with exible distributions of the error term. In this paper, we focus on AFT models with exible distributions of random eects. In particular, we consider multivariate skew-normally distributed random effects. When the sample size is large, exible distribution of the random effects provides a better description of the dependence structure on the data. The performance of the model is evaluated by simulations. Further, the proposed log-skew-normal AFT model is illustrated with data on multiple myeloma patients with autologous transplantation from the European Bone Marrow Transplantation Registry.
Log-skew-normal accelerated failure time models
Callegaro, Andrea
2012
Abstract
Accelerated Failure Time models (AFT) are useful alternatives to the Cox model in survival analysis. Recently, AFT models for multivariate data have been considered with exible distributions of the error term. In this paper, we focus on AFT models with exible distributions of random eects. In particular, we consider multivariate skew-normally distributed random effects. When the sample size is large, exible distribution of the random effects provides a better description of the dependence structure on the data. The performance of the model is evaluated by simulations. Further, the proposed log-skew-normal AFT model is illustrated with data on multiple myeloma patients with autologous transplantation from the European Bone Marrow Transplantation Registry.File | Dimensione | Formato | |
---|---|---|---|
2012_3_20120329084929.pdf
accesso aperto
Licenza:
Non specificato
Dimensione
1.68 MB
Formato
Adobe PDF
|
1.68 MB | Adobe PDF | Visualizza/Apri |
Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.