3
$\begingroup$

Ask some statistical concepts of linear regression:

  1. Does the general linear regression specify that the independent variables are independent of each other?

    • I have seen a similar description that needs to match the characteristics of LINE

    • I also saw that some people say no, only when collinearity appears.

  2. Next, how do you prove the dependent/independent relationship between "independent variables"? Is it a method of using correlation coefficients or probability independent events?

Improve this question
$\endgroup$

2 Answers 2

Reset to default
6
$\begingroup$

The use of the name "independent variable" is unfortunately misleading. Personally I prefer to use other terms if I can. They are not assumed to be independent of each other.

In the regression model you condition on them so their dependence is generally not material to prediction, but there is a possible issue with estimation of individual effects/effect sizes (particularly in relation to deciding which variables are important, for example).

Improve this answer
$\endgroup$
1
$\begingroup$

It is not assumed that the "explicative variables" must be independent from each other. However, we must be aware of the problem of colinearity.

If your explicative variables are correlated, then the model may get confused about which one of them is having the effect on the "target". It's true that correlation and dependency are two different concepts, but this is still an idea to keep in mind

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.