I want to use R to select a variable. In particular, variables are selected considering the correlation, and variables with large multicollinearity, which are correlations between variables, should be removed. I have a question here.
I will consider the debt and assets as independent variables with the customer's credit (good and bad) as the dependent variable. If there are a lot of assets, it is good to customer, and if there is a lot of debt, it is bad.
Assuming here that assets and debt are highly correlated, we can conclude that the two variables are multi-collinear. In this multi-collinearity, it is correct to delete one of the two variables, but considering the degree of influence on the dependent variable, both variables are considered important.
So is it correct to delete one variable in this situation? Or is it right to use it?
When there is multicollinearity, some people said that it should consider VIF index. If so, what are the VIF index criteria for deleting variables?
I would appreciate it if you could give me the correct answer for this part.