Exact Kolmogorov-Type Tests for Left-Truncated and/or Right-Censored Data
本文提出基于乘积限估计量的Kolmogorov-Smirnov单样本检验的自然推广，适用于左截断和/或右删失数据，且对任意样本量和零假设下的任何分布均有效。
Abstract The tests proposed in this article are natural generalizations of the ordinary Kolmogorov—Smirnov one-sample tests based on the product-limit estimator. They can be used with left-truncated and/or right-censored data and, as the ordinary tests, they have the nice properties of being valid for any sample size and any underlying distribution specified by the null hypothesis. The acceptance region for the product-limit estimator consists of a Kolmogorov—Smirnov-type region that is suitably reduced or enlarged to the right of each point at which there is a change in the number “at risk” due to entries and/or censorship.