Estimation of the Base Hazard Function by Bootstrapping

This thesis examines the techniques in estimating the base hazard function by bootstrapping. The base hazard function is a crucial part of survival analysis. It is used to construct an estimate of the proportional hazard model for every individual. As in many methods for analysing survival data, this thesis utilizes the nonparametric model of Kaplan Meier, the Cox proportional hazard regression of the parametric model and the data validation by bootstrapping. The Cox proportional hazard regression is used to model failure time data in censored data. Bootstrapping schemes validate the models based on Efron’s technique and the data samples are generated using S-Plus programme randomizer. v Assessment of this method is investigated by performing simulation study on generated data. Two simulation studies are carried out to confirm the suitability of the models. Graph obtained from the results indicated that bootstrapping provides an alternative method in constructing estimation for base hazard function. This method is good alternative for a distribution-free approach with a minimal set of data.

Preview PDF 1000548962_t_FS_2004_2.pdf Download (174kB)

Actions (login required)

View Item