What is it about?

Multicollinearity is the phenomenon in which two or more identified predictor variables in a multiple regression model are highly correlated. The presence of this phenomenon can have a negative impact on the analysis as a whole and can severely limit the conclusions of the research study. This paper reviews and provides examples of the different ways in which multicollinearity can affect a research project, how to detect multicollinearity and how one can reduce it through Ridge Regression applications. This paper is intended for any level of SAS® user.

Featured Image

Why is it important?

Multicollinearity is an assumption violation that if left unchecked, can have a detrimental effect on a model. Ridge Regression is one way that multicollinearity can be addressed, while maintaining the structure of the model.

Perspectives

I wrote this paper after coming face to face with a significant issue of multicollinearity in my personal work. Through extensive research and hands-on problem solving, I found that I was mainly needing to rely on the results of blog searches and one-on-one conversations with other statisticians. Given how individualized these experiences were, I wanted to fill the gap in available literature and make sure that individuals who may not have access to the same blogs and statisticians that I did, could still learn from my experience and findings.

Deanna N Schreiber-Gregory
Henry M Jackson Foundation for the Advancement of Military Medicine Inc

Read the Original

This page is a summary of: Ridge Regression and multicollinearity: An in-depth review, Model Assisted Statistics and Applications, October 2018, IOS Press,
DOI: 10.3233/mas-180446.
You can read the full text:

Read

Resources

Contributors

The following have contributed to this page