1
$\begingroup$

What does the sign of the estimated co-efficient mean in a Regression Model? (Like some are positive and some are negative)

My model when I use all independent variables

Also I understand the concept of correlation. But I fail to understand that correlation and multi-collinearity decrease the strength of the variable. Please explain intuitively.

enter image description here

$\endgroup$
1
$\begingroup$

Multicollinearity, intuitively: The way I explain multicollinearity to my clients is, coefficients determine how much of the change to ascribe to driver A and how much to B. But, if A and B move so closely together (are correlated), the model doesn't know whether it's A or B. This isn't really a problem for prediction, since the model still gets the same estimate (choosing one or the other), but it's a problem if you want to infer anything meaningful about the strength or impact of A or B as drivers in themselves.

$\endgroup$
  • $\begingroup$ together they ll have impact on the model, that is why R^2 remains same. I get it, Beautifully explained! And also what does the Estimate Co-efficient determine explain with respect to magnitude and sign. $\endgroup$ – Sivagami Nambi Jun 24 '17 at 11:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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