h******3
发帖数: 190 | 1 刚进program, 没学过survival analysis, 基础也奇差，但是为了一个project只能自
学。
Does cox proportional hazard model control for type 1 error?
Just wondering if there are many covarites, the likelihood that one of
coefficients is significant should be quite high.
I realize score test is also called log rank test. Is the log rank test for
coxph in some way related to the log rank test for kaplan meier?
If data is not complete for all covariates, can the subject still be used
and in what way?
多谢帮忙！ | f***a
发帖数: 329 | | Y******Y
发帖数: 8753 | 3 仍然可以用subjects with incomplete covariate. 有好多作者都提出过如何处理。
可以参考Prentice(1982)，Zhou and Pepe(1995)，Lin and Ying (1993)，Paik and
Tsai (1997)，Chen and Little(1999)，Herring and Ibrahim (2001)
for
【在 h******3 的大作中提到】
: 刚进program, 没学过survival analysis, 基础也奇差，但是为了一个project只能自
: 学。
: Does cox proportional hazard model control for type 1 error?
: Just wondering if there are many covarites, the likelihood that one of
: coefficients is significant should be quite high.
: I realize score test is also called log rank test. Is the log rank test for
: coxph in some way related to the log rank test for kaplan meier?
: If data is not complete for all covariates, can the subject still be used
: and in what way?
: 多谢帮忙！
| h******3
发帖数: 190 | 4 Thanks a lot!
For the log rank test, if the covariate in cox model is continuous, do they
always find an optimal cut point and then do the log rank test just as
categorical covariates? I cannot link log rank test with continuous covariates. How does the cox model do that? |
|
