I have a set of 10 variables all based around inventory optimization. Many of the variables are highly correlated. The ask is to determine the magnitude of the increase in one variable based on the change in several other variables. Is there an appropriate technique to estimate that?

I had thought about creating multiple regression models, however as the variables are highly correlated (and the variance inflation factors [VIFs] are HUGE) and I'm concerned that the model(s) would not be robust. Any suggestions here?

  • $\begingroup$ how many samples do you have? what's the goal of your model? make the best predictions, or describe the relationship between variables or something else? $\endgroup$ – rep_ho Sep 22 '20 at 10:08
  • $\begingroup$ I have 6 variables that are the weekly level. There are 148 observations in total. I want to try to determine the magnitude of the change in one variable based on changing values in other variables. The variables are highly correlated so regression modeling makes me nervous. $\endgroup$ – Patrick Sep 23 '20 at 12:10
  • $\begingroup$ would it be enough to do it separately for each variable? And maybe just describe that the correlations are high. $\endgroup$ – rep_ho Sep 23 '20 at 13:44

Your Answer

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

Browse other questions tagged or ask your own question.