竞争风险下右删失数据的剩余寿命分位数回归预测  

Regression Inference on Quantile Residual Life of Right-Censored Data Under Competitive Risk

作　　者：高伟 冯海林 Gao Wei;Feng Hailin(School of Mathematics and Statistics,Xidian University,Xi＇an 710126,China)

机构地区：[1]西安电子科技大学数学与统计学院

出　　处：《统计与决策》2018年第21期19-22,共4页Statistics & Decision

基　　金：国家自然科学基金资助项目(71271165)

摘　　要：文章针对竞争风险下的右删失数据,建立了剩余寿命分位数回归模型,其中分位数与协变量呈对数线性关系,协变量随时间动态变化,并得到了动态协变量与剩余寿命分位数之间的函数关系。进一步得到模型参数的估计量,推导了估计量的渐进性和一致性。最后数值模拟验证了所提出的估计方法在有限样本时的性质,并给出了Channing House数据集的实例分析。Considering right-censored data under competitive risk, this paper establishes the quantile residual life regres- sion model, where the relationship between the qnantile and covariate is log-linear, and the covariate changes dynamically with time, and then obtains the functional relationship between the dynamic covariate and quantile residual life. Further, the estima- tors of the model parameters are obtained to derive asymptoties and consistency of the estimators. Finally, the properties of the proposed estimation method in finite sample are verified by numerical simulation, and a practical case analysis of Channing House dataset is given.

关 键 词：右删失数据 剩余寿命分位数 竞争风险 协变量 回归模型 

分 类 号：O212[理学—概率论与数理统计]