ABSTRACT Regression analysis of interval-censored failure time data with noninformative censoring has been widely investigated and many methods have been proposed. Sometimes the mechanism behind the interval censoring may be informative and several approaches have also been developed for this latter situation. However, all of these existing methods are for single models and it is well known that in many situations, one may prefer more flexible models. Corresponding to this, the linear transformation model is considered and a maximum likelihood estimation method is established. In the proposed method, the association between the failure time of interest and the censoring time is modelled by the copula model, and the involved nonparametric functions are approximated by spline functions. The large sample properties of the proposed estimators are derived. Numerical results show that the proposed method performs well in practical application. Besides, a real data example is presented for the illustration.

中文翻译：

---

具有半参数线性变换模型的信息间隔删失失效时间数据的回归分析

摘要 使用非信息删失的区间删失失效时间数据的回归分析已被广泛研究，并提出了许多方法。有时，区间删失背后的机制可能提供信息，并且针对后一种情况还开发了几种方法。然而，所有这些现有方法都是针对单一模型的，众所周知，在许多情况下，人们可能更喜欢更灵活的模型。与此相对应，考虑线性变换模型，建立最大似然估计方法。在所提出的方法中，感兴趣的失效时间和审查时间之间的关联通过copula模型建模，并且涉及的非参数函数通过样条函数近似。推导出了所提出的估计量的大样本特性。数值结果表明，该方法在实际应用中表现良好。此外，还提供了一个真实的数据示例以供说明。