Hello,

I don't know if I am writing in the right group, but I have a question.

I used a linear regression to evaluate the impact of some variables on a dependent variable. The correlation of each variable with the dependent one was tested and proven to be linear. Now, I want to compute an importance score of each independent variable in determining the value of the dependent one. It would be wrong if I would use a non-linear model like Random Forest to compute them or I should also use a linear method to do it? Maybe the Random Forest and the linear regression could be complementary? (The idea is that I have already computed the importance scores using the Random Forest method and after that I realized this is a non-linear method).

I tend to believe that I should also use a linear method to compute these importance scores, but I want to make sure, as I would have to do the work again. :)) 

Any suggestion is welcome!

Thank you!