聚类失效时间带混合效应的加性乘积模型的统计推断

Inference on the mixed effect additive-multiplicative hazard model for clustered failure time data

  • 摘要: 在生存分析中,会收集到来自不同医院或治疗中心的聚类数据,每个医院或每个治疗中心称为一个类,在同一个类中的个体具有相似性. 本文对这种聚类失效时间建立了加性乘积风险率函数模型,对同一类的个体用一个共同的混合效应项表示类内部的相关性. 在统计推断中,利用估计方程的方法对参数进行估计, 给出了估计量的大样本理论结果和证明,然后进行数值模拟, 验证有限样本下的估计量的表现, 并将模型和方法应用到实际数据中进行分析.

     

    Abstract: In survival analysis, data from different hospital or treatment center may be collected. The individuals from the same hospital are grouped into a cluster, and they have similarity. An additive-multiplicative hazard model was developed for the clustered failure time. A mixed effect was added to the model denoting the correlation among the cluster. The individuals in the same cluster share the same effect. Estimating equation method was applied to obtainthe estimator. Asymptotic properties were given and proved. Some simulation was conducted to verify the performance of the estimate under the finite sample. Real data was illustrated using the model and the proposed method.

     

/

返回文章
返回