考虑在一阶线性自回归模型下不会导致风险显著增加的最高剂量的估计问题,基于赤池信息准则的思想提出了一个新的估计方法,证明了该估计的强相合性,并通过模拟说明了其对于中小样本的优良性.
This paper is focused on the estimation of the highest dose level for no increased toxicity response in longitudinal data. A new estimator is proposed based on the idea of AIC. The estimator is strongly consistent when the number of subjects in trial goes to infinity, and is proved good by simulations with respect to population distribution for small and moderate sample sizes.