R 언어를 이용한 프레일티를 포함한 비례위험모형에 대한 베이지안 추론

Abstract

Survival analysis is, statistic methods analyzing data about lifetime which gotten in animal experiment or clinical demonstration, or survival time. Statistic analysis of survival time is applied widely in not only medical and physiology but also technology and sociology. Generally when analyzing survival data, we use Cox(1972)'s proportional hazards model which doesn't need a assumption about distribution of survival time when there are covariates which have effect on survival time. Cox's proportional hazards model assumed that every subject's survival time is independent each other and relation along covariates is linear. but in survival data which doesn't satisfy the independent assumption about survival time, use of cox's proportional hazards model isn't proper. Many researchers suggest the frailty model to solve this. In this paper, we compare and analyze the Cox's proportional hazards model and frailty model using Cox's partial likelihood by using real data.