两样本Logrank检验的确切概率计算法  

CALCULATION OF EXACT PROBABILITY OF TWO SAMPLES LOGRAND TEST

作　　者：苏炳华[1] 杨琦[1] 

机构地区：[1]上海第二医科大学生物统计教研室

出　　处：《现代预防医学》1991年第4期205-208,250-252,共6页Modern Preventive Medicine

摘　　要：Logrank检验用于两个总体生存期的比较,一般用于有截尾数据的样本,当生存期为非截尾数据时,就是Savage检验法。本文对于秩变换作出一种第二次变换。称为Logrank变换,并将原始资料化成2×C有序列联表,导得其确切概率计算法。Logrank test is used to compare two populations with censored survival data. When the data are uncensored, the test specializes to the Savage test.We define the Logrand transform, one of the second transform to rank transform, and get an 2 * C ordered contingency tables from the raw data. Then the calculation method of the exact probability of the Logrank test can be gotten. This paper presents the method.

关 键 词：2×C有序列联表 Logrank检验 确切概率 

分 类 号：R1[医药卫生—公共卫生与预防医学]