摘要
在观察性试验研究中,某些受试者的诊断检验值可能缺失,若只用完全观测的数据可能会得到有偏的估计.考虑在诊断检验值随机缺失的情形下,基于灵敏性和特异性相等的对称点准则将逆概率加权和多重插补相结合建立起包含待估参数的非光滑估计方程,采用非参数两样本经验似然方法给出阈值和灵敏性的联合估计和置信域.在一定的正则条件下,建立了阈值和灵敏性联合估计的渐近性质.模拟研究证实所提方法的估计要优于其他方法的估计.
In observational studies,diagnostic test values for some subjects may be missing.Only if completely observed data are used,biased estimates may be obtained.In the presence of missing at random diagnostic test values,non-smooth estimation quations with the unknown parameters are established by combining inverse probability weighting and multiple imputation based on the symmetry point criterion with equal sensitivity and specificity.The joint nonparametric confidence regions for the optimal cutoff value and sensitivity are obtained by the two-sample empirical likelihood method.Under certain regular conditions,the asymptotic properties of the maximum empirical likelihood estimators of cutoff value and sensitivity are established.Simulation studies show that the proposed approach is better than other approaches.
作者
程伟丽
吴莹
左卫兵
Cheng Weili;Wu Ying;Zuo Weibing(School of Mathematics and Statistics,North China University of Water Resources and Electric Power,Zhengzhou 450046,China;School of Mathematics and Statistics,Yunnan University,Kunming 650500,China)
出处
《河南师范大学学报（自然科学版）》
CAS
北大核心
2023年第3期97-104,共8页
Journal of Henan Normal University(Natural Science Edition)
基金
国家自然科学基金(11871419,12201550)
河南省高等学校重点科研项目(22A560003).
关键词
随机缺失
阈值
灵敏性
对称点准则
经验似然
missing at random
cutoff value
sensitivity
symmetry point criterion
empirical likelihood
