考虑带有截尾数据的线性回归模型,在较为一般的截尾数据情形下,本文给出了其参数极大似然估计(MLE)存在且唯一的充分必要条件,且对MLE不存在的样本数据集Q的描述定义更为准确.
Considering the linear regression models with censored data, a necessary and sufficient condition is obtained for the existence and uniqueness of maximum likelihood estimates of parameters under the more general situation of censored data, and the set of data Q in which MLE does not exist is more accurately described.