Análisis de Supervivencia. Enfoque Bayesiano
Seminari GRASS - 18 d'Octubre 2006
DATE: 18 d'Octubre 2006 a les 11 hores
SPEAKER: Lucas López Segovia
IDIOM: Castellano
PLACE: Aula C5016, Edifici C5, Campus Nord, UPC
SUMMARY:
This presentation is intended to give a big picture of the Bayesian modeling techniques with cure models. Bayesian methods for right-censored survival data for population with a surviving (cure) fraction are considered. The models studied are standard and nonstandard cure rate models. Also most recent results in Bayesian inference about nonstandard cure rate models are presented: parametric and semiparametric cases. A detailed description of the noninformative and informative prior distributions are discussed.
Several theoretical properties of the priors and resulting posteriors are derived, and comparisons are made to the standard cure rate model. Although these models were first presented in 1999, a large body of theorical results and applications has been reported to date. Theses issues are discussed in the context of survival data in cancer.
SPEAKER: Lucas López Segovia
IDIOM: Castellano
PLACE: Aula C5016, Edifici C5, Campus Nord, UPC
SUMMARY:
This presentation is intended to give a big picture of the Bayesian modeling techniques with cure models. Bayesian methods for right-censored survival data for population with a surviving (cure) fraction are considered. The models studied are standard and nonstandard cure rate models. Also most recent results in Bayesian inference about nonstandard cure rate models are presented: parametric and semiparametric cases. A detailed description of the noninformative and informative prior distributions are discussed.
Several theoretical properties of the priors and resulting posteriors are derived, and comparisons are made to the standard cure rate model. Although these models were first presented in 1999, a large body of theorical results and applications has been reported to date. Theses issues are discussed in the context of survival data in cancer.
Share: