Bootstrap方法在Cox模型参数估计中的应用

Bootstrap Method on Parameter Estimation of Cox Model

摘要: 目的探讨自变量分布极偏时Cox模型的参数以及区间估计是否准确,引入Bootstrap法对其加以验证。方法拟合Cox模型研究痰检结果与肺癌发病风险的关系,使用非参数Bootstrap法验证Cox模型的参数估计值、可信区间是否有效。结果两者的参数估计值和大部分可信区间基本一致,个别变量的可信区间不一致,Bootstrap可信区间的下界更低,甚至变为负值。结论当自变量分布极偏时,Cox模型所估计的参数可信区间可能有偏,Bootstrap可信区间更准确。

Abstract: ObjectiveTo test the validity of parameter estimation of Cox model with the skewness of independent variables.MethodsTo introduce nonparametric bootstrap method to Cox model and check the difference of their results.ResultsTheir parameter estimation is almost same,but the confidence interval is not,some of the bootstrap down interval is much lower.ConclusionWhen the independent variables are significantly asymmetrical,the confidence interval of the Cox model is not as accuracy as that of bootstrap method.