光滑误差项分布的加速失效时间模型及其医学应用

Accelerated Failure Time Model with Smoothed Error Distribution and its Application in Medicine

目的介绍光滑误差项分布的加速失效时间模型并探讨其在医学中的应用。方法通过惩罚最大似然的方法对模型中的参数做出估计,以"伪方差估计"的方法做进一步推断,以实例说明模型的实际应用并与Cox模型结果相比较。结果资料不满足Cox模型比例风险假定时,加速失效时间模型的结果优于Cox模型。结论光滑误差项分布的加速失效时间模型不需要知道误差项的基准分布,也不需要满足比例风险假定,是Cox模型很好的替代模型。

Objective To introduce the accelerated failure time model with smoothed error distribution and its application.Methods Parameters of model are quickly and accurately estimated via penalized maximum likelihood.The Pseudo-variance has been used for drawing accurate inferences.An example has been illustrated for its application.The results are compared with the results of Cox model.Results The model is better than Cox model when the proportional hazards assumption is not satisfied.Conclusion The accelerated fa...