骨髓瘤数据中贝叶斯估计研究及SAS实现

The study of Bayesian estimation with Myeloma data and its SAS realization

基于右删失数据,借助于SAS软件,运用贝叶斯估计方法分析骨髓瘤患者的生存时间、死亡风险及影响因素。首先运用贝叶斯框架下的Cox比例风险模型对个体的死亡风险进行评估,分析死亡风险的影响因素,然后,通过极大似然估计和贝叶斯估计拟合生存数据服从不同参数分布时的回归模型,分析生存时间的影响因素。在骨髓瘤生存数据下,贝叶斯估计适用于Cox模型参数估计及参数模型的参数估计,而且比极大似然估计利用信息更多,在小样本情况下有更优良的性质,所得结果对骨髓瘤的防治具有重大实际意义。

In this paper,we aim at right censored data,by the means of SAS software,we use Bayesian Estimation Method to analyze the survival time of Myeloma patients,the risk of death and its factors.Firstly,we will use the Cox proportional hazards model under Bayesian conditions to assess the risk of death of individuals.Then,fit survival data Regression model by likelihood estimation and Bayesian estimation when the data subject to different parameter distribution,analyze the factors of the survival time.In Myeloma data,Bayesian estimation apply to the parameter estimation of Cox model and the estimation of the parameter model,and use more information than maximum likelihood estimation,in the case of small sample,Bayesian estimation has better properties,and the results are of great practical significance to the treatment of Myeloma cancer.