Estimation in the General Multiplicative Model for Survival.
NORTH CAROLINA UNIV AT CHAPEL HILL DEPT OF STATISTICS
Pagination or Media Count:
The general multiplicative model represents the hazard function as the product of an underlying hazard rate, gamma t, of unspecified form and a certain function of known form, gz beta, where z is a vector of concomitant variables, and Beta is a vector of unknown parameters. Assuming that gamma t can be approximated by a constant between any two consecutive failures, the general forms of likelihood function are derived. The likelihood utilizes the available information on the time of exposure to risk of each individual until failure or withdrawal. Special cases, when the zs do not depend on t are discussed in some detail. Multiple failures are handled in a simple manner - no ordering of failures is required. Estimation of empirical survival function when there are no covariates is discussed. An example using heart transplant data, is given for illustrative purpose only.
- Anatomy and Physiology
- Theoretical Mathematics