Nonparametric maximum likelihood estimation of an increasing hazard rate for uncertain cause-of-death data
针对病理学家无法确定死因的数据，推导了疾病风险率的非参数最大似然估计，假设风险率递增且无病生存函数已知或可估计，并证明了估计的强相合性。
In Kaplan-Meier estimation of the survival function for diseased animals the cause of death has to be specified with certainty. When pathologists are unable to do so forced cause-of-death data can create substantial biases. For unidentifiable cause-of-death data we derive the nonparametric maximum likelihood estimator of the hazard rate due to the disease assuming it is increasing when the survival function without the disease is known or can be well-estimated. Strong consistency of the maximum likelihobd estimator is also obtained.