l*********s
发帖数: 5409 | 1 我觉得好像先调整似乎比较合理，但是，这样计算次序可能就影响结果了，一般应该怎
么处理？（SS已经给定） |
D******n
发帖数: 2836 | 2 dont understand.... u mean estimation of component A needs the estimation of
component B which is negative? this situations exists?
【在 l*********s 的大作中提到】
: 我觉得好像先调整似乎比较合理，但是，这样计算次序可能就影响结果了，一般应该怎
: 么处理？（SS已经给定）
|
l*********s
发帖数: 5409 | 3 Yes, say in multi-factorial mixed models,several variance components need to
be estimated. If some terms are slightly negative, do you retro-correct
other positive estimates as well?
of
【在 D******n 的大作中提到】
: dont understand.... u mean estimation of component A needs the estimation of
: component B which is negative? this situations exists?
|
D******n
发帖数: 2836 | 4 I guess, for each component, u can have one final formula which involves no
estimation of other component. Can you? I am not very sure. I guess one solu
tion for negative value is to use ML.
to
【在 l*********s 的大作中提到】
: Yes, say in multi-factorial mixed models,several variance components need to
: be estimated. If some terms are slightly negative, do you retro-correct
: other positive estimates as well?
:
: of
|
l*********s
发帖数: 5409 | 5 No. I cannot use ML since the SS is given. Anyway, it probably is a moot
question. After all it is just an estimate.
no
solu
【在 D******n 的大作中提到】
: I guess, for each component, u can have one final formula which involves no
: estimation of other component. Can you? I am not very sure. I guess one solu
: tion for negative value is to use ML.
:
: to
|
