s*****p
发帖数: 299 | 1 如果两个dependant with each other的变量被用在MODEL里了，会怎么样呢？ |
l***a
发帖数: 12410 | 2 会挺好
【在 s*****p 的大作中提到】
: 如果两个dependant with each other的变量被用在MODEL里了，会怎么样呢？
|
d*******y
发帖数: 1154 | 3 木有什么关系。只要不是完全的线性相关，可能会出现矩阵无法求逆
【在 s*****p 的大作中提到】
: 如果两个dependant with each other的变量被用在MODEL里了，会怎么样呢？
|
s*****p
发帖数: 299 | |
n*****n
发帖数: 3123 | 5 如果线性相关，去掉一个没有影响
【在 s*****p 的大作中提到】
: 如果是线性相关的话会怎么样呢？
|
s*****p
发帖数: 299 | |
t*****2
发帖数: 94 | 7 none of the t-ratios of the coefficient are statistically significant, but
the F-test for the equation as a whole is significant, which is a typical
sign of multicollinearity. The reason why this happens is that we probably
put too many variables that measure the same thing. |
h**t
发帖数: 1678 | 8 nod
【在 t*****2 的大作中提到】
: none of the t-ratios of the coefficient are statistically significant, but
: the F-test for the equation as a whole is significant, which is a typical
: sign of multicollinearity. The reason why this happens is that we probably
: put too many variables that measure the same thing.
|
w******4
发帖数: 488 | 9 very true!
【在 t*****2 的大作中提到】
: none of the t-ratios of the coefficient are statistically significant, but
: the F-test for the equation as a whole is significant, which is a typical
: sign of multicollinearity. The reason why this happens is that we probably
: put too many variables that measure the same thing.
|
s*****p
发帖数: 299 | |
