Application Of Bayesian Model Average Method In Cox Proportional Hazard

In this paper,we apply Bayesian model averaging to the selection of variables in Cox proportional hazard models.We introduce a technique based on the leaps and bounds algorithm which efficiently locates and fits the best models in the very large model space,and saving the computer running time.For each variable,the Bayesian model averaging method provides a posteriori probability,which uses the size of the posteriori probability as an objective to objectively explain the significance of the influencing variable,and this posteriori probability is more important and direct than the P-values Variable evaluation criteria.P-values from models preferred by stepwise methods tend to overstate the evidence for the predictive value of a variable and do not account for model uncertainty.Finally,we introduce partial prediction scores to evaluate the predictive effects of two different methods.Based on the data of oropharyngeal cancer and veteran lung cancer data,the result show that the Bayesian model averaging predictively outperforms standard model selection and the Bayesian model averaging can better assess the significant factors that affect the survival time of patients,And has a better forecasting effect.