The website is in Maintenance mode. We are in the process of adding more features.
Any new bookmarks, comments, or user profiles made during this time will not be saved.

AIML.com

Machine Learning Resources

Doesn’t polynomial regression violate the multicollinearity assumption for Linear Regression? 

Bookmark this question

While the original variable will likely be correlated with a higher order term constructed from the same variable, multicollinearity is not an overt cause of concern because the full effect of the variable consists of contributions from both the original and transformed component. In a model containing both X and X2, the effect of the variable X is the composite of the first and second order terms. Thus, the assumption of multicollinearity in linear regression is restricted to the independent, first-order effects. However, the inherent correlation introduced by including a polynomial term can be addressed by centering the original variable (subtracting each value by the overall mean of the variable) and then performing the transformation on the centered representation. 

Leave your Comments and Suggestions below:

Please Login or Sign Up to leave a comment

Partner Ad  

Find out all the ways
that you can

Explore Questions by Topics