We develop a Bayesian local influence procedure for generalized failure time models with latent variables and multivariate censored data. We propose to use the penalized splines (P-splines) approach to formulate the unknown functions of the proposed models. We assess the effects of minor perturbations to individual observations, the prior distributions of parameters, and the sampling distribution on statistical inference through various perturbation schemes. The first-order local influence measure is used to quantify the degree of minor perturbations to different aspects of a statistical model with the use of Bayes factor as an objective function. Simulation studies show that the empirical performance of the Bayesian local influence procedure is satisfactory. An application to a study of renal disease for type 2 diabetes patients is presented.

中文翻译：

---

具有潜在变量和多元删失数据的广义失效时间模型的贝叶斯局部影响

我们为具有潜在变量和多元删失数据的广义故障时间模型开发了贝叶斯局部影响程序。我们建议使用惩罚样条（P-splines）方法来制定所提出模型的未知函数。我们通过各种扰动方案评估微小扰动对个体观察、参数的先验分布和抽样分布对统计推断的影响。一阶局部影响度量用于量化对统计模型不同方面的微小扰动的程度，使用贝叶斯因子作为目标函数。模拟研究表明，贝叶斯局部影响程序的经验性能是令人满意的。介绍了对 2 型糖尿病患者肾脏疾病研究的应用。