Related Experiment Videos
On Bayesian inference for proportional hazards models using noninformative priors.
1Department of Biostatistics, Harvard School of Public Health and Dana Farber Cancer Institute, 44 Binney St., Boston, MA 02115, USA.
Lifetime Data Analysis
|February 24, 2001
Summary
This study examines the posterior distribution for Weibull and extreme value regression models with right-censored data. We identified conditions ensuring the posterior moment generating function exists for regression coefficients.
Area of Science:
- Statistics
- Survival Analysis
- Biostatistics
Background:
- Proportional hazards models are crucial in survival analysis.
- Weibull and extreme value models are widely applied.
- Handling right-censored data is a common challenge.
Purpose of the Study:
- To investigate the posterior distribution properties.
- To analyze the Weibull and extreme value regression models.
- To consider the impact of uniform improper priors and right censoring.
Main Methods:
- Utilized a uniform improper prior.
- Assessed properties of the posterior distribution.
- Derived conditions for the existence of the posterior moment generating function.
Main Results:
- Established sufficient conditions for the existence of the posterior moment generating function.
- Demonstrated the applicability of the findings through a lung cancer clinical trial dataset.
- Validated results using simulation.
Conclusions:
- The study provides theoretical insights into Bayesian inference for proportional hazards models.
- The findings are relevant for analyzing censored survival data.
- The methods are illustrated with practical examples.